[PATCH] D135392: Use PoisonValue in vector BIs [NFC]

2022-10-06 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito created this revision.
ManuelJBrito added a reviewer: nlopes.
ManuelJBrito added a project: clang.
Herald added subscribers: dmgreen, arphaman, kbarton, nemanjai.
Herald added a project: All.
ManuelJBrito requested review of this revision.
Herald added a subscriber: cfe-commits.

Replacing the following instances of UndefValue with PoisonValue in 
clang/lib/CodeGen/CGBuiltin.cpp where the UndefValue is used as an arbitrary 
value.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D135392

Files:
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/test/CodeGen/PowerPC/builtins-ppc-int128.c
  clang/test/CodeGen/aarch64-bf16-ldst-intrinsics.c
  clang/test/CodeGen/aarch64-neon-intrinsics.c
  clang/test/CodeGen/aarch64-neon-ldst-one.c
  clang/test/CodeGen/aarch64-neon-scalar-x-indexed-elem.c
  clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq-bfloat.c
  clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq.c
  clang/test/CodeGen/arm-mve-intrinsics/vld24.c
  clang/test/CodeGen/arm_neon_intrinsics.c

Index: clang/test/CodeGen/arm_neon_intrinsics.c
===
--- clang/test/CodeGen/arm_neon_intrinsics.c
+++ clang/test/CodeGen/arm_neon_intrinsics.c
@@ -4077,7 +4077,7 @@
 
 // CHECK-LABEL: @test_vld1q_dup_u8(
 // CHECK:   [[TMP0:%.*]] = load i8, i8* %a, align 1
-// CHECK:   [[TMP1:%.*]] = insertelement <16 x i8> undef, i8 [[TMP0]], i32 0
+// CHECK:   [[TMP1:%.*]] = insertelement <16 x i8> poison, i8 [[TMP0]], i32 0
 // CHECK:   [[LANE:%.*]] = shufflevector <16 x i8> [[TMP1]], <16 x i8> [[TMP1]], <16 x i32> zeroinitializer
 // CHECK:   ret <16 x i8> [[LANE]]
 uint8x16_t test_vld1q_dup_u8(uint8_t const * a) {
@@ -4088,7 +4088,7 @@
 // CHECK:   [[TMP0:%.*]] = bitcast i16* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to i16*
 // CHECK:   [[TMP2:%.*]] = load i16, i16* [[TMP1]], align 2
-// CHECK:   [[TMP3:%.*]] = insertelement <8 x i16> undef, i16 [[TMP2]], i32 0
+// CHECK:   [[TMP3:%.*]] = insertelement <8 x i16> poison, i16 [[TMP2]], i32 0
 // CHECK:   [[LANE:%.*]] = shufflevector <8 x i16> [[TMP3]], <8 x i16> [[TMP3]], <8 x i32> zeroinitializer
 // CHECK:   ret <8 x i16> [[LANE]]
 uint16x8_t test_vld1q_dup_u16(uint16_t const * a) {
@@ -4099,7 +4099,7 @@
 // CHECK:   [[TMP0:%.*]] = bitcast i32* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to i32*
 // CHECK:   [[TMP2:%.*]] = load i32, i32* [[TMP1]], align 4
-// CHECK:   [[TMP3:%.*]] = insertelement <4 x i32> undef, i32 [[TMP2]], i32 0
+// CHECK:   [[TMP3:%.*]] = insertelement <4 x i32> poison, i32 [[TMP2]], i32 0
 // CHECK:   [[LANE:%.*]] = shufflevector <4 x i32> [[TMP3]], <4 x i32> [[TMP3]], <4 x i32> zeroinitializer
 // CHECK:   ret <4 x i32> [[LANE]]
 uint32x4_t test_vld1q_dup_u32(uint32_t const * a) {
@@ -4110,7 +4110,7 @@
 // CHECK:   [[TMP0:%.*]] = bitcast i64* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to i64*
 // CHECK:   [[TMP2:%.*]] = load i64, i64* [[TMP1]], align 4
-// CHECK:   [[TMP3:%.*]] = insertelement <2 x i64> undef, i64 [[TMP2]], i32 0
+// CHECK:   [[TMP3:%.*]] = insertelement <2 x i64> poison, i64 [[TMP2]], i32 0
 // CHECK:   [[LANE:%.*]] = shufflevector <2 x i64> [[TMP3]], <2 x i64> [[TMP3]], <2 x i32> zeroinitializer
 // CHECK:   ret <2 x i64> [[LANE]]
 uint64x2_t test_vld1q_dup_u64(uint64_t const * a) {
@@ -4119,7 +4119,7 @@
 
 // CHECK-LABEL: @test_vld1q_dup_s8(
 // CHECK:   [[TMP0:%.*]] = load i8, i8* %a, align 1
-// CHECK:   [[TMP1:%.*]] = insertelement <16 x i8> undef, i8 [[TMP0]], i32 0
+// CHECK:   [[TMP1:%.*]] = insertelement <16 x i8> poison, i8 [[TMP0]], i32 0
 // CHECK:   [[LANE:%.*]] = shufflevector <16 x i8> [[TMP1]], <16 x i8> [[TMP1]], <16 x i32> zeroinitializer
 // CHECK:   ret <16 x i8> [[LANE]]
 int8x16_t test_vld1q_dup_s8(int8_t const * a) {
@@ -4130,7 +4130,7 @@
 // CHECK:   [[TMP0:%.*]] = bitcast i16* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to i16*
 // CHECK:   [[TMP2:%.*]] = load i16, i16* [[TMP1]], align 2
-// CHECK:   [[TMP3:%.*]] = insertelement <8 x i16> undef, i16 [[TMP2]], i32 0
+// CHECK:   [[TMP3:%.*]] = insertelement <8 x i16> poison, i16 [[TMP2]], i32 0
 // CHECK:   [[LANE:%.*]] = shufflevector <8 x i16> [[TMP3]], <8 x i16> [[TMP3]], <8 x i32> zeroinitializer
 // CHECK:   ret <8 x i16> [[LANE]]
 int16x8_t test_vld1q_dup_s16(int16_t const * a) {
@@ -4141,7 +4141,7 @@
 // CHECK:   [[TMP0:%.*]] = bitcast i32* %a to i8*
 // CHECK:   [[TMP1:%.*]] = bitcast i8* [[TMP0]] to i32*
 // CHECK:   [[TMP2:%.*]] = load i32, i32* [[TMP1]], align 4
-// CHECK:   [[TMP3:%.*]] = insertelement <4 x i32> undef, i32 [[TMP2]], i32 0
+// CHECK:   [[TMP3:%.*]] = insertelement <4 x i32> poison, i32 [[TMP2]], i32 0
 // CHECK:   [[LANE:%.*]] = shufflevector <4 x i32> [[TMP3]], <4 x i32> [[TMP3]], <4 x i32> zeroinitializer
 // CHECK:   ret <4 x i32> [[LANE]]
 int32x4_t test_vld1q_dup_s32(int32_t const * a) {
@@ -4152,7 +4152,7 @@
 // CHECK:   [[TMP0:%.*]] = bitcast i64* %a to i8*
 // CHECK:   [[TMP1:%

[PATCH] D140538: [Clang][CodeGen] Use poison instead of undef for dummy values in CGExpr [NFC]

2022-12-22 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito created this revision.
ManuelJBrito added a reviewer: nlopes.
ManuelJBrito added a project: clang.
Herald added a project: All.
ManuelJBrito requested review of this revision.
Herald added subscribers: libcxx-commits, cfe-commits.
Herald added a project: libc++abi.
Herald added a reviewer: libc++abi.

Replace the following uses of poison with undef where undef is used as a dummy 
value.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D140538

Files:
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/lib/CodeGen/CGCall.cpp
  clang/lib/CodeGen/CGExpr.cpp
  clang/lib/CodeGen/CGExprComplex.cpp
  clang/lib/CodeGen/CGExprScalar.cpp
  clang/lib/CodeGen/CodeGenFunction.h
  libcxxabi/test/test_demangle.pass.cpp

Index: libcxxabi/test/test_demangle.pass.cpp
===
--- libcxxabi/test/test_demangle.pass.cpp
+++ libcxxabi/test/test_demangle.pass.cpp
@@ -3169,7 +3169,7 @@
 {"_ZN5clang7CodeGen15CodeGenFunction9EmitCheckEPN4llvm5ValueEj", "clang::CodeGen::CodeGenFunction::EmitCheck(llvm::Value*, unsigned int)"},
 {"_ZN5clang7CodeGen15CodeGenFunction9getTrapBBEv", "clang::CodeGen::CodeGenFunction::getTrapBB()"},
 {"_ZN5clang7CodeGen15CodeGenFunction24EmitComplexPrePostIncDecEPKNS_13UnaryOperatorENS0_6LValueEbb", "clang::CodeGen::CodeGenFunction::EmitComplexPrePostIncDec(clang::UnaryOperator const*, clang::CodeGen::LValue, bool, bool)"},
-{"_ZN5clang7CodeGen15CodeGenFunction14GetUndefRValueENS_8QualTypeE", "clang::CodeGen::CodeGenFunction::GetUndefRValue(clang::QualType)"},
+{"_ZN5clang7CodeGen15CodeGenFunction14GetPoisonRValueENS_8QualTypeE", "clang::CodeGen::CodeGenFunction::GetPoisonRValue(clang::QualType)"},
 {"_ZN5clang7CodeGen15CodeGenFunction21EmitUnsupportedRValueEPKNS_4ExprEPKc", "clang::CodeGen::CodeGenFunction::EmitUnsupportedRValue(clang::Expr const*, char const*)"},
 {"_ZN5clang7CodeGen15CodeGenFunction21EmitUnsupportedLValueEPKNS_4ExprEPKc", "clang::CodeGen::CodeGenFunction::EmitUnsupportedLValue(clang::Expr const*, char const*)"},
 {"_ZN5clang7CodeGen15CodeGenFunction17EmitCheckedLValueEPKNS_4ExprE", "clang::CodeGen::CodeGenFunction::EmitCheckedLValue(clang::Expr const*)"},
Index: clang/lib/CodeGen/CodeGenFunction.h
===
--- clang/lib/CodeGen/CodeGenFunction.h
+++ clang/lib/CodeGen/CodeGenFunction.h
@@ -3740,8 +3740,8 @@
   /// Create a check that a scalar RValue is non-null.
   llvm::Value *EmitNonNullRValueCheck(RValue RV, QualType T);
 
-  /// GetUndefRValue - Get an appropriate 'undef' rvalue for the given type.
-  RValue GetUndefRValue(QualType Ty);
+  /// GetPoisonRValue - Get an appropriate 'poison' rvalue for the given type.
+  RValue GetPoisonRValue(QualType Ty);
 
   /// EmitUnsupportedRValue - Emit a dummy r-value using the type of E
   /// and issue an ErrorUnsupported style diagnostic (using the
Index: clang/lib/CodeGen/CGExprScalar.cpp
===
--- clang/lib/CodeGen/CGExprScalar.cpp
+++ clang/lib/CodeGen/CGExprScalar.cpp
@@ -1624,7 +1624,7 @@
   CGF.ErrorUnsupported(E, "scalar expression");
   if (E->getType()->isVoidType())
 return nullptr;
-  return llvm::UndefValue::get(CGF.ConvertType(E->getType()));
+  return llvm::PoisonValue::get(CGF.ConvertType(E->getType()));
 }
 
 Value *
@@ -4906,7 +4906,7 @@
   // If EmitVAArg fails, emit an error.
   if (!ArgPtr.isValid()) {
 CGF.ErrorUnsupported(VE, "va_arg expression");
-return llvm::UndefValue::get(ArgTy);
+return llvm::PoisonValue::get(ArgTy);
   }
 
   // FIXME Volatility.
Index: clang/lib/CodeGen/CGExprComplex.cpp
===
--- clang/lib/CodeGen/CGExprComplex.cpp
+++ clang/lib/CodeGen/CGExprComplex.cpp
@@ -419,7 +419,7 @@
   CGF.ErrorUnsupported(E, "complex expression");
   llvm::Type *EltTy =
 CGF.ConvertType(getComplexType(E->getType())->getElementType());
-  llvm::Value *U = llvm::UndefValue::get(EltTy);
+  llvm::Value *U = llvm::PoisonValue::get(EltTy);
   return ComplexPairTy(U, U);
 }
 
@@ -1274,7 +1274,7 @@
 CGF.ErrorUnsupported(E, "complex va_arg expression");
 llvm::Type *EltTy =
   CGF.ConvertType(E->getType()->castAs()->getElementType());
-llvm::Value *U = llvm::UndefValue::get(EltTy);
+llvm::Value *U = llvm::PoisonValue::get(EltTy);
 return ComplexPairTy(U, U);
   }
 
Index: clang/lib/CodeGen/CGExpr.cpp
===
--- clang/lib/CodeGen/CGExpr.cpp
+++ clang/lib/CodeGen/CGExpr.cpp
@@ -1169,7 +1169,7 @@
   return Builder.CreateICmpNE(V, llvm::Constant::getNullValue(V->getType()));
 }
 
-RValue CodeGenFunction::GetUndefRValue(QualType Ty) {
+RValue CodeGenFunction::GetPoisonRValue(QualType Ty) {
   if (Ty->isVoidType())
 return RValue::get(nullptr);
 
@@ -1177,7 +1177,7 @@
   case TEK_Complex: {
 llvm::Type *EltTy =
   ConvertType(Ty->

[PATCH] D140538: [Clang][CodeGen] Use poison instead of undef for dummy values in CGExpr [NFC]

2022-12-22 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

A 2022-12-22 10:51, Nuno Lopes via Phabricator escreveu:

> nlopes accepted this revision.
> nlopes added a comment.
>
> LGTM, thank you!
>
> Repository:
>
>   rG LLVM Github Monorepo
>
> CHANGES SINCE LAST ACTION
>
>   https://reviews.llvm.org/D140538/new/
>
> https://reviews.llvm.org/D140538

Está a falhar no test_demangle.pass.cpp, espero que alguém do libcxx 
diga alguma coisa ou faço ping a alguém diretamente?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D140538/new/

https://reviews.llvm.org/D140538

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D140538: [Clang][CodeGen] Use poison instead of undef for dummy values in CGExpr [NFC]

2022-12-23 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 485072.
ManuelJBrito added a comment.

Fix demangle test


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D140538/new/

https://reviews.llvm.org/D140538

Files:
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/lib/CodeGen/CGCall.cpp
  clang/lib/CodeGen/CGExpr.cpp
  clang/lib/CodeGen/CGExprComplex.cpp
  clang/lib/CodeGen/CGExprScalar.cpp
  clang/lib/CodeGen/CodeGenFunction.h
  libcxxabi/test/test_demangle.pass.cpp

Index: libcxxabi/test/test_demangle.pass.cpp
===
--- libcxxabi/test/test_demangle.pass.cpp
+++ libcxxabi/test/test_demangle.pass.cpp
@@ -3169,7 +3169,7 @@
 {"_ZN5clang7CodeGen15CodeGenFunction9EmitCheckEPN4llvm5ValueEj", "clang::CodeGen::CodeGenFunction::EmitCheck(llvm::Value*, unsigned int)"},
 {"_ZN5clang7CodeGen15CodeGenFunction9getTrapBBEv", "clang::CodeGen::CodeGenFunction::getTrapBB()"},
 {"_ZN5clang7CodeGen15CodeGenFunction24EmitComplexPrePostIncDecEPKNS_13UnaryOperatorENS0_6LValueEbb", "clang::CodeGen::CodeGenFunction::EmitComplexPrePostIncDec(clang::UnaryOperator const*, clang::CodeGen::LValue, bool, bool)"},
-{"_ZN5clang7CodeGen15CodeGenFunction14GetUndefRValueENS_8QualTypeE", "clang::CodeGen::CodeGenFunction::GetUndefRValue(clang::QualType)"},
+{"_ZN5clang7CodeGen15CodeGenFunction15GetPoisonRValueENS_8QualTypeE", "clang::CodeGen::CodeGenFunction::GetPoisonRValue(clang::QualType)"},
 {"_ZN5clang7CodeGen15CodeGenFunction21EmitUnsupportedRValueEPKNS_4ExprEPKc", "clang::CodeGen::CodeGenFunction::EmitUnsupportedRValue(clang::Expr const*, char const*)"},
 {"_ZN5clang7CodeGen15CodeGenFunction21EmitUnsupportedLValueEPKNS_4ExprEPKc", "clang::CodeGen::CodeGenFunction::EmitUnsupportedLValue(clang::Expr const*, char const*)"},
 {"_ZN5clang7CodeGen15CodeGenFunction17EmitCheckedLValueEPKNS_4ExprE", "clang::CodeGen::CodeGenFunction::EmitCheckedLValue(clang::Expr const*)"},
Index: clang/lib/CodeGen/CodeGenFunction.h
===
--- clang/lib/CodeGen/CodeGenFunction.h
+++ clang/lib/CodeGen/CodeGenFunction.h
@@ -3740,8 +3740,8 @@
   /// Create a check that a scalar RValue is non-null.
   llvm::Value *EmitNonNullRValueCheck(RValue RV, QualType T);
 
-  /// GetUndefRValue - Get an appropriate 'undef' rvalue for the given type.
-  RValue GetUndefRValue(QualType Ty);
+  /// GetPoisonRValue - Get an appropriate 'poison' rvalue for the given type.
+  RValue GetPoisonRValue(QualType Ty);
 
   /// EmitUnsupportedRValue - Emit a dummy r-value using the type of E
   /// and issue an ErrorUnsupported style diagnostic (using the
Index: clang/lib/CodeGen/CGExprScalar.cpp
===
--- clang/lib/CodeGen/CGExprScalar.cpp
+++ clang/lib/CodeGen/CGExprScalar.cpp
@@ -1624,7 +1624,7 @@
   CGF.ErrorUnsupported(E, "scalar expression");
   if (E->getType()->isVoidType())
 return nullptr;
-  return llvm::UndefValue::get(CGF.ConvertType(E->getType()));
+  return llvm::PoisonValue::get(CGF.ConvertType(E->getType()));
 }
 
 Value *
@@ -4906,7 +4906,7 @@
   // If EmitVAArg fails, emit an error.
   if (!ArgPtr.isValid()) {
 CGF.ErrorUnsupported(VE, "va_arg expression");
-return llvm::UndefValue::get(ArgTy);
+return llvm::PoisonValue::get(ArgTy);
   }
 
   // FIXME Volatility.
Index: clang/lib/CodeGen/CGExprComplex.cpp
===
--- clang/lib/CodeGen/CGExprComplex.cpp
+++ clang/lib/CodeGen/CGExprComplex.cpp
@@ -419,7 +419,7 @@
   CGF.ErrorUnsupported(E, "complex expression");
   llvm::Type *EltTy =
 CGF.ConvertType(getComplexType(E->getType())->getElementType());
-  llvm::Value *U = llvm::UndefValue::get(EltTy);
+  llvm::Value *U = llvm::PoisonValue::get(EltTy);
   return ComplexPairTy(U, U);
 }
 
@@ -1274,7 +1274,7 @@
 CGF.ErrorUnsupported(E, "complex va_arg expression");
 llvm::Type *EltTy =
   CGF.ConvertType(E->getType()->castAs()->getElementType());
-llvm::Value *U = llvm::UndefValue::get(EltTy);
+llvm::Value *U = llvm::PoisonValue::get(EltTy);
 return ComplexPairTy(U, U);
   }
 
Index: clang/lib/CodeGen/CGExpr.cpp
===
--- clang/lib/CodeGen/CGExpr.cpp
+++ clang/lib/CodeGen/CGExpr.cpp
@@ -1169,7 +1169,7 @@
   return Builder.CreateICmpNE(V, llvm::Constant::getNullValue(V->getType()));
 }
 
-RValue CodeGenFunction::GetUndefRValue(QualType Ty) {
+RValue CodeGenFunction::GetPoisonRValue(QualType Ty) {
   if (Ty->isVoidType())
 return RValue::get(nullptr);
 
@@ -1177,7 +1177,7 @@
   case TEK_Complex: {
 llvm::Type *EltTy =
   ConvertType(Ty->castAs()->getElementType());
-llvm::Value *U = llvm::UndefValue::get(EltTy);
+llvm::Value *U = llvm::PoisonValue::get(EltTy);
 return RValue::getComplex(std::make_pair(U, U));
   }
 
@@ -1190,7 +1190,7 @@
   }
 
   case TEK_Scalar:
-return RValue::get(llvm::U

[PATCH] D140538: [Clang][CodeGen] Use poison instead of undef for dummy values in CGExpr [NFC]

2023-01-05 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

Ping


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D140538/new/

https://reviews.llvm.org/D140538

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D142388: [clang] Add builtin_nondet

2023-01-24 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added inline comments.



Comment at: clang/include/clang/Basic/Builtins.def:658
 BUILTIN(__builtin_call_with_static_chain, "v.", "nt")
+BUILTIN(__builtin_nondet, "v.", "nt")
 

erichkeane wrote:
> aaron.ballman wrote:
> > aaron.ballman wrote:
> > > shafik wrote:
> > > > erichkeane wrote:
> > > > > Not a fan of the name in general, it doesn't really explain what is 
> > > > > happening.  Perhaps `__builtin_nondeterministic_value` or something?  
> > > > > 
> > > > > I also wonder if others might have a better name :) 
> > > > `__builtin_indeterminate_value` wdyt?
> > > I think `__builtin_unspecified_value()` makes sense, as that's what 
> > > you're getting back (there is a valid value, but you may get a different 
> > > value on every call). I don't think `indeterminate` makes sense because 
> > > that implies it may return trap values, and I'm guessing that's not 
> > > behavior we want here (or am I wrong about that)?
> > > 
> > > Also, should this be usable within a constant expression? I would assume 
> > > not; we should have a test case showing you can't use it in such a 
> > > context.
> > Why does this require custom typechecking?
> Our issue last time with "unspecified" value was that this has a distinct 
> meaning in C/C++, so we don't want to use that word.  
> I don't think `indeterminate` makes sense because that implies it may return 
> trap values, and I'm guessing that's not behavior we want here (or am I wrong 
> about that)?

The idea is for it to return some correct value of the same type as the 
argument, so no trap values.

Can we settle on `__builtin_nondeterministic_value`? 



Comment at: clang/include/clang/Basic/Builtins.def:658
 BUILTIN(__builtin_call_with_static_chain, "v.", "nt")
+BUILTIN(__builtin_nondet, "v.", "nt")
 

ManuelJBrito wrote:
> erichkeane wrote:
> > aaron.ballman wrote:
> > > aaron.ballman wrote:
> > > > shafik wrote:
> > > > > erichkeane wrote:
> > > > > > Not a fan of the name in general, it doesn't really explain what is 
> > > > > > happening.  Perhaps `__builtin_nondeterministic_value` or 
> > > > > > something?  
> > > > > > 
> > > > > > I also wonder if others might have a better name :) 
> > > > > `__builtin_indeterminate_value` wdyt?
> > > > I think `__builtin_unspecified_value()` makes sense, as that's what 
> > > > you're getting back (there is a valid value, but you may get a 
> > > > different value on every call). I don't think `indeterminate` makes 
> > > > sense because that implies it may return trap values, and I'm guessing 
> > > > that's not behavior we want here (or am I wrong about that)?
> > > > 
> > > > Also, should this be usable within a constant expression? I would 
> > > > assume not; we should have a test case showing you can't use it in such 
> > > > a context.
> > > Why does this require custom typechecking?
> > Our issue last time with "unspecified" value was that this has a distinct 
> > meaning in C/C++, so we don't want to use that word.  
> > I don't think `indeterminate` makes sense because that implies it may 
> > return trap values, and I'm guessing that's not behavior we want here (or 
> > am I wrong about that)?
> 
> The idea is for it to return some correct value of the same type as the 
> argument, so no trap values.
> 
> Can we settle on `__builtin_nondeterministic_value`? 
> Why does this require custom typechecking?

I wasn't sure how to have the builtin take one argument of any type. 



Comment at: clang/include/clang/Basic/Builtins.def:658
 BUILTIN(__builtin_call_with_static_chain, "v.", "nt")
+BUILTIN(__builtin_nondet, "v.", "nt")
 

ManuelJBrito wrote:
> ManuelJBrito wrote:
> > erichkeane wrote:
> > > aaron.ballman wrote:
> > > > aaron.ballman wrote:
> > > > > shafik wrote:
> > > > > > erichkeane wrote:
> > > > > > > Not a fan of the name in general, it doesn't really explain what 
> > > > > > > is happening.  Perhaps `__builtin_nondeterministic_value` or 
> > > > > > > something?  
> > > > > > > 
> > > > > > > I also wonder if others might have a better name :) 
> > > > > > `__builtin_indeterminate_value` wdyt?
> > > > > I think `__builtin_unspecified_value()` makes sense, as that's what 
> > > > > you're getting back (there is a valid value, but you may get a 
> > > > > different value on every call). I don't think `indeterminate` makes 
> > > > > sense because that implies it may return trap values, and I'm 
> > > > > guessing that's not behavior we want here (or am I wrong about that)?
> > > > > 
> > > > > Also, should this be usable within a constant expression? I would 
> > > > > assume not; we should have a test case showing you can't use it in 
> > > > > such a context.
> > > > Why does this require custom typechecking?
> > > Our issue last time with "unspecified" value was that this has a distinct 
> > > meaning in C/C++, so we don't want to use that word.  
> > > I don't think `indeterminate` ma

[PATCH] D142388: [clang] Add builtin_nondet

2023-01-24 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 491767.
ManuelJBrito added a comment.

Add tests with bool and bool-vector


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142388/new/

https://reviews.llvm.org/D142388

Files:
  clang/test/CodeGen/builtins-nondet.c


Index: clang/test/CodeGen/builtins-nondet.c
===
--- clang/test/CodeGen/builtins-nondet.c
+++ clang/test/CodeGen/builtins-nondet.c
@@ -1,6 +1,7 @@
 // RUN: %clang_cc1 %s -emit-llvm -o - | FileCheck %s
 
 typedef float float4 __attribute__((ext_vector_type(4)));
+typedef _Bool bool4 __attribute__((ext_vector_type(4)));
 
 int clang_nondet_i( int x ) {
 // CHECK-LABEL: entry
@@ -29,7 +30,17 @@
   return __builtin_nondet(x);
 } 
 
-float4 clang_nondet_v( float4 x ) {
+_Bool clang_nondet_b( _Bool x) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[B:%.*]] = zext i1 %x to i8
+// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: [[R:%.*]] = freeze i1 poison
+// CHECK: ret i1 [[R]]
+  return __builtin_nondet(x);
+} 
+
+float4 clang_nondet_fv( float4 x ) {
 // CHECK-LABEL: entry
 // CHECK: [[A:%.*]] = alloca <4 x float>, align 16
 // CHECK: store <4 x float> [[X:%.*]], ptr [[A]], align 16
@@ -38,3 +49,14 @@
   return __builtin_nondet(x);
 } 
 
+bool4 clang_nondet_bv( bool4 x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[V:%.*]] = shufflevector <4 x i1> [[X:%.*]], <4 x i1> poison, <8 x 
i32> 
+// CHECK: [[B:%.*]] = bitcast <8 x i1> [[V]] to i8
+// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: [[R:%.*]] = freeze <4 x i1> poison
+// CHECK: ret <4 x i1> [[R]]
+  return __builtin_nondet(x);
+} 
+


Index: clang/test/CodeGen/builtins-nondet.c
===
--- clang/test/CodeGen/builtins-nondet.c
+++ clang/test/CodeGen/builtins-nondet.c
@@ -1,6 +1,7 @@
 // RUN: %clang_cc1 %s -emit-llvm -o - | FileCheck %s
 
 typedef float float4 __attribute__((ext_vector_type(4)));
+typedef _Bool bool4 __attribute__((ext_vector_type(4)));
 
 int clang_nondet_i( int x ) {
 // CHECK-LABEL: entry
@@ -29,7 +30,17 @@
   return __builtin_nondet(x);
 } 
 
-float4 clang_nondet_v( float4 x ) {
+_Bool clang_nondet_b( _Bool x) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[B:%.*]] = zext i1 %x to i8
+// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: [[R:%.*]] = freeze i1 poison
+// CHECK: ret i1 [[R]]
+  return __builtin_nondet(x);
+} 
+
+float4 clang_nondet_fv( float4 x ) {
 // CHECK-LABEL: entry
 // CHECK: [[A:%.*]] = alloca <4 x float>, align 16
 // CHECK: store <4 x float> [[X:%.*]], ptr [[A]], align 16
@@ -38,3 +49,14 @@
   return __builtin_nondet(x);
 } 
 
+bool4 clang_nondet_bv( bool4 x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[V:%.*]] = shufflevector <4 x i1> [[X:%.*]], <4 x i1> poison, <8 x i32> 
+// CHECK: [[B:%.*]] = bitcast <8 x i1> [[V]] to i8
+// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: [[R:%.*]] = freeze <4 x i1> poison
+// CHECK: ret <4 x i1> [[R]]
+  return __builtin_nondet(x);
+} 
+
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D142388: [clang] Add builtin_nondet

2023-01-24 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

Oops wrong diff, i'll update it.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142388/new/

https://reviews.llvm.org/D142388

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D142388: [clang] Add builtin_nondet

2023-01-24 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 491783.
ManuelJBrito added a comment.

fixed diff


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142388/new/

https://reviews.llvm.org/D142388

Files:
  clang/include/clang/Basic/Builtins.def
  clang/include/clang/Sema/Sema.h
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/lib/Sema/SemaChecking.cpp
  clang/test/CodeGen/builtins-nondet.c

Index: clang/test/CodeGen/builtins-nondet.c
===
--- /dev/null
+++ clang/test/CodeGen/builtins-nondet.c
@@ -0,0 +1,62 @@
+// RUN: %clang_cc1 %s -emit-llvm -o - | FileCheck %s
+
+typedef float float4 __attribute__((ext_vector_type(4)));
+typedef _Bool bool4 __attribute__((ext_vector_type(4)));
+
+int clang_nondet_i( int x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i32, align 4
+// CHECK: store i32 [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[R:%.*]] = freeze i32 poison
+// CHECK: ret i32 [[R]]
+  return __builtin_nondet(x);
+} 
+
+float clang_nondet_f( float x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca float, align 4
+// CHECK: store float [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[R:%.*]] = freeze float poison
+// CHECK: ret float [[R]]
+  return __builtin_nondet(x);
+}
+
+double clang_nondet_d( double x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca double, align 8
+// CHECK: store double [[X:%.*]], ptr [[A]], align 8
+// CHECK: [[R:%.*]] = freeze double poison
+// CHECK: ret double [[R]]
+  return __builtin_nondet(x);
+} 
+
+_Bool clang_nondet_b( _Bool x) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[B:%.*]] = zext i1 %x to i8
+// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: [[R:%.*]] = freeze i1 poison
+// CHECK: ret i1 [[R]]
+  return __builtin_nondet(x);
+} 
+
+float4 clang_nondet_fv( float4 x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca <4 x float>, align 16
+// CHECK: store <4 x float> [[X:%.*]], ptr [[A]], align 16
+// CHECK: [[R:%.*]] = freeze <4 x float> poison
+// CHECK: ret <4 x float> [[R]]
+  return __builtin_nondet(x);
+} 
+
+bool4 clang_nondet_bv( bool4 x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[V:%.*]] = shufflevector <4 x i1> [[X:%.*]], <4 x i1> poison, <8 x i32> 
+// CHECK: [[B:%.*]] = bitcast <8 x i1> [[V]] to i8
+// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: [[R:%.*]] = freeze <4 x i1> poison
+// CHECK: ret <4 x i1> [[R]]
+  return __builtin_nondet(x);
+} 
+
Index: clang/lib/Sema/SemaChecking.cpp
===
--- clang/lib/Sema/SemaChecking.cpp
+++ clang/lib/Sema/SemaChecking.cpp
@@ -2580,6 +2580,12 @@
 break;
   }
 
+  case Builtin::BI__builtin_nondet: {
+if (SemaBuiltinNonDet(TheCall))
+  return ExprError();
+break;
+  }
+
   // __builtin_elementwise_abs restricts the element type to signed integers or
   // floating point types only.
   case Builtin::BI__builtin_elementwise_abs: {
@@ -17799,6 +17805,16 @@
   return false;
 }
 
+bool Sema::SemaBuiltinNonDet(CallExpr *TheCall) {
+  if (checkArgCount(*this, TheCall, 1))
+return true;
+
+  ExprResult Arg = TheCall->getArg(0);
+  QualType TyArg = Arg.get()->getType();
+  TheCall->setType(TyArg);
+  return false;
+}
+
 ExprResult Sema::SemaBuiltinMatrixTranspose(CallExpr *TheCall,
 ExprResult CallResult) {
   if (checkArgCount(*this, TheCall, 1))
Index: clang/lib/CodeGen/CGBuiltin.cpp
===
--- clang/lib/CodeGen/CGBuiltin.cpp
+++ clang/lib/CodeGen/CGBuiltin.cpp
@@ -3060,6 +3060,12 @@
 return RValue::get(V);
   }
 
+  case Builtin::BI__builtin_nondet: {
+Value *Result = PoisonValue::get(ConvertType(E->getArg(0)->getType()));
+Result = Builder.CreateFreeze(Result);
+return RValue::get(Result);
+  }
+
   case Builtin::BI__builtin_elementwise_abs: {
 Value *Result;
 QualType QT = E->getArg(0)->getType();
Index: clang/include/clang/Sema/Sema.h
===
--- clang/include/clang/Sema/Sema.h
+++ clang/include/clang/Sema/Sema.h
@@ -13529,6 +13529,8 @@
   bool PrepareBuiltinElementwiseMathOneArgCall(CallExpr *TheCall);
   bool PrepareBuiltinReduceMathOneArgCall(CallExpr *TheCall);
 
+  bool SemaBuiltinNonDet(CallExpr *TheCall);
+
   // Matrix builtin handling.
   ExprResult SemaBuiltinMatrixTranspose(CallExpr *TheCall,
 ExprResult CallResult);
Index: clang/include/clang/Basic/Builtins.def
===
--- clang/include/clang/Basic/Builtins.def
+++ clang/include/clang/Basic/Builtins.def
@@ -655,6 +655,7 @@
 BUILTIN(__builtin_alloca_with_align, "v*zIz", "Fn")
 BUILTIN(__builtin_alloca_with_align_uninitialized, "v*zIz", "Fn")
 BUILTIN(__builtin_call_with_static_chain, "v.", "nt")
+BU

[PATCH] D142388: [clang] Add builtin_nondeterministic_value

2023-02-01 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 493904.
ManuelJBrito retitled this revision from "[clang] Add builtin_nondet" to 
"[clang] Add builtin_nondeterministic_value".
ManuelJBrito edited the summary of this revision.
ManuelJBrito added a comment.
Herald added a subscriber: mgrang.

- Change name to `__builtin_nondeterministic_value`
- Add struct support




Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142388/new/

https://reviews.llvm.org/D142388

Files:
  clang/include/clang/Basic/Builtins.def
  clang/include/clang/Sema/Sema.h
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/lib/Sema/SemaChecking.cpp
  clang/test/CodeGen/builtins-nondeterministic-value.c

Index: clang/test/CodeGen/builtins-nondeterministic-value.c
===
--- /dev/null
+++ clang/test/CodeGen/builtins-nondeterministic-value.c
@@ -0,0 +1,75 @@
+// RUN: %clang_cc1 %s -emit-llvm -o - | FileCheck %s
+
+typedef float float4 __attribute__((ext_vector_type(4)));
+typedef _Bool bool4 __attribute__((ext_vector_type(4)));
+struct testStruct{
+  int x;
+  int y;
+};
+
+int clang_nondet_i( int x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i32, align 4
+// CHECK: store i32 [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[R:%.*]] = freeze i32 poison
+// CHECK: ret i32 [[R]]
+  return __builtin_nondeterministic_value(x);
+} 
+
+float clang_nondet_f( float x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca float, align 4
+// CHECK: store float [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[R:%.*]] = freeze float poison
+// CHECK: ret float [[R]]
+  return __builtin_nondeterministic_value(x);
+}
+
+double clang_nondet_d( double x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca double, align 8
+// CHECK: store double [[X:%.*]], ptr [[A]], align 8
+// CHECK: [[R:%.*]] = freeze double poison
+// CHECK: ret double [[R]]
+  return __builtin_nondeterministic_value(x);
+} 
+
+_Bool clang_nondet_b( _Bool x) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[B:%.*]] = zext i1 %x to i8
+// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: [[R:%.*]] = freeze i1 poison
+// CHECK: ret i1 [[R]]
+  return __builtin_nondeterministic_value(x);
+} 
+
+float4 clang_nondet_fv( float4 x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca <4 x float>, align 16
+// CHECK: store <4 x float> [[X:%.*]], ptr [[A]], align 16
+// CHECK: [[R:%.*]] = freeze <4 x float> poison
+// CHECK: ret <4 x float> [[R]]
+  return __builtin_nondeterministic_value(x);
+} 
+
+bool4 clang_nondet_bv( bool4 x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[V:%.*]] = shufflevector <4 x i1> [[X:%.*]], <4 x i1> poison, <8 x i32> 
+// CHECK: [[B:%.*]] = bitcast <8 x i1> [[V]] to i8
+// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: [[R:%.*]] = freeze <4 x i1> poison
+// CHECK: ret <4 x i1> [[R]]
+  return __builtin_nondeterministic_value(x);
+} 
+
+void clang_nondet_struct(struct testStruct x) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca [[S:%.*]], align 4
+// CHECK: [[B:%.*]] = alloca [[S]], align 4
+// CHECK: [[R:%.*]] = freeze [[S]] poison
+// CHECK: store [[S]] [[R]], ptr [[B]], align 4
+// CHECK: ret void
+  x = __builtin_nondeterministic_value(x);
+} 
Index: clang/lib/Sema/SemaChecking.cpp
===
--- clang/lib/Sema/SemaChecking.cpp
+++ clang/lib/Sema/SemaChecking.cpp
@@ -2580,6 +2580,12 @@
 break;
   }
 
+  case Builtin::BI__builtin_nondeterministic_value: {
+if (SemaBuiltinNonDeterministicValue(TheCall))
+  return ExprError();
+break;
+  }
+
   // __builtin_elementwise_abs restricts the element type to signed integers or
   // floating point types only.
   case Builtin::BI__builtin_elementwise_abs: {
@@ -17799,6 +17805,16 @@
   return false;
 }
 
+bool Sema::SemaBuiltinNonDeterministicValue(CallExpr *TheCall) {
+  if (checkArgCount(*this, TheCall, 1))
+return true;
+
+  ExprResult Arg = TheCall->getArg(0);
+  QualType TyArg = Arg.get()->getType();
+  TheCall->setType(TyArg);
+  return false;
+}
+
 ExprResult Sema::SemaBuiltinMatrixTranspose(CallExpr *TheCall,
 ExprResult CallResult) {
   if (checkArgCount(*this, TheCall, 1))
Index: clang/lib/CodeGen/CGBuiltin.cpp
===
--- clang/lib/CodeGen/CGBuiltin.cpp
+++ clang/lib/CodeGen/CGBuiltin.cpp
@@ -3060,6 +3060,22 @@
 return RValue::get(V);
   }
 
+  case Builtin::BI__builtin_nondeterministic_value: {
+QualType QT = E->getArg(0)->getType();
+llvm::Type * Ty = ConvertType(QT); 
+
+Value *Result = PoisonValue::get(Ty);
+Result = Builder.CreateFreeze(Result);
+
+if(Ty->isStructTy()){
+  Address StructAddr = ReturnValue.getValue();
+  Builder.CreateStore(Result, StructAddr);
+  return RValue::getAggregate(StructAddr);
+}
+
+r

[PATCH] D142388: [clang] Add builtin_nondeterministic_value

2023-02-01 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added inline comments.



Comment at: clang/lib/CodeGen/CGBuiltin.cpp:3070
+
+if(Ty->isStructTy()){
+  Address StructAddr = ReturnValue.getValue();

erichkeane wrote:
> This gets REALLY complicated, you can't just create a store, this might end 
> up hitting conversion operators/etc, and is subject to triviality/etc, and 
> also probably needs to go through a constructor.  I suspect you're going to 
> prefer to just decide this isn't a valid builtin for structs instead of 
> getting bogged down in that mess.
The motivation for this builtin was to match intel's definition of undefined in 
the lowering of intrinsics such as  
https://www.intel.com/content/www/us/en/docs/intrinsics-guide/index.html#text=_mm256_castsi128_si256&expand=628&ig_expand=755,
  so struct support isn't critical.
So if everyone else agrees i'll drop struct support, do as you suggest and if 
there are use cases for nondeterministic values of other types add support for 
them later?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142388/new/

https://reviews.llvm.org/D142388

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D142388: [clang] Add builtin_nondeterministic_value

2023-02-01 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 493984.
ManuelJBrito added a comment.

- Restrict builtin to base types and vectors


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142388/new/

https://reviews.llvm.org/D142388

Files:
  clang/include/clang/Basic/Builtins.def
  clang/include/clang/Sema/Sema.h
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/lib/Sema/SemaChecking.cpp
  clang/test/CodeGen/builtins-nondet.c
  clang/test/CodeGen/builtins-nondeterministic-value.c

Index: clang/test/CodeGen/builtins-nondeterministic-value.c
===
--- clang/test/CodeGen/builtins-nondeterministic-value.c
+++ clang/test/CodeGen/builtins-nondeterministic-value.c
@@ -9,7 +9,7 @@
 // CHECK: store i32 [[X:%.*]], ptr [[A]], align 4
 // CHECK: [[R:%.*]] = freeze i32 poison
 // CHECK: ret i32 [[R]]
-  return __builtin_nondet(x);
+  return __builtin_nondeterministic_value(x);
 } 
 
 float clang_nondet_f( float x ) {
@@ -18,7 +18,7 @@
 // CHECK: store float [[X:%.*]], ptr [[A]], align 4
 // CHECK: [[R:%.*]] = freeze float poison
 // CHECK: ret float [[R]]
-  return __builtin_nondet(x);
+  return __builtin_nondeterministic_value(x);
 }
 
 double clang_nondet_d( double x ) {
@@ -27,7 +27,7 @@
 // CHECK: store double [[X:%.*]], ptr [[A]], align 8
 // CHECK: [[R:%.*]] = freeze double poison
 // CHECK: ret double [[R]]
-  return __builtin_nondet(x);
+  return __builtin_nondeterministic_value(x);
 } 
 
 _Bool clang_nondet_b( _Bool x) {
@@ -37,7 +37,7 @@
 // CHECK: store i8 [[B]], ptr [[A]], align 1
 // CHECK: [[R:%.*]] = freeze i1 poison
 // CHECK: ret i1 [[R]]
-  return __builtin_nondet(x);
+  return __builtin_nondeterministic_value(x);
 } 
 
 float4 clang_nondet_fv( float4 x ) {
@@ -46,7 +46,7 @@
 // CHECK: store <4 x float> [[X:%.*]], ptr [[A]], align 16
 // CHECK: [[R:%.*]] = freeze <4 x float> poison
 // CHECK: ret <4 x float> [[R]]
-  return __builtin_nondet(x);
+  return __builtin_nondeterministic_value(x);
 } 
 
 bool4 clang_nondet_bv( bool4 x ) {
@@ -57,6 +57,5 @@
 // CHECK: store i8 [[B]], ptr [[A]], align 1
 // CHECK: [[R:%.*]] = freeze <4 x i1> poison
 // CHECK: ret <4 x i1> [[R]]
-  return __builtin_nondet(x);
+  return __builtin_nondeterministic_value(x);
 } 
-
Index: clang/lib/Sema/SemaChecking.cpp
===
--- clang/lib/Sema/SemaChecking.cpp
+++ clang/lib/Sema/SemaChecking.cpp
@@ -2580,8 +2580,8 @@
 break;
   }
 
-  case Builtin::BI__builtin_nondet: {
-if (SemaBuiltinNonDet(TheCall))
+  case Builtin::BI__builtin_nondeterministic_value: {
+if (SemaBuiltinNonDeterministicValue(TheCall))
   return ExprError();
 break;
   }
@@ -17805,12 +17805,17 @@
   return false;
 }
 
-bool Sema::SemaBuiltinNonDet(CallExpr *TheCall) {
+bool Sema::SemaBuiltinNonDeterministicValue(CallExpr *TheCall) {
   if (checkArgCount(*this, TheCall, 1))
 return true;
 
   ExprResult Arg = TheCall->getArg(0);
   QualType TyArg = Arg.get()->getType();
+
+  if (!TyArg->isBuiltinType() && !TyArg->isVectorType()) 
+Diag(TheCall->getArg(0)->getBeginLoc(), diag::err_builtin_invalid_arg_type)
+   << 1 << /*vector, integer or floating point ty*/ 0 << TyArg;
+
   TheCall->setType(TyArg);
   return false;
 }
Index: clang/lib/CodeGen/CGBuiltin.cpp
===
--- clang/lib/CodeGen/CGBuiltin.cpp
+++ clang/lib/CodeGen/CGBuiltin.cpp
@@ -3060,9 +3060,12 @@
 return RValue::get(V);
   }
 
-  case Builtin::BI__builtin_nondet: {
-Value *Result = PoisonValue::get(ConvertType(E->getArg(0)->getType()));
+  case Builtin::BI__builtin_nondeterministic_value: {
+llvm::Type * Ty = ConvertType(E->getArg(0)->getType()); 
+
+Value *Result = PoisonValue::get(Ty);
 Result = Builder.CreateFreeze(Result);
+
 return RValue::get(Result);
   }
 
Index: clang/include/clang/Sema/Sema.h
===
--- clang/include/clang/Sema/Sema.h
+++ clang/include/clang/Sema/Sema.h
@@ -13529,7 +13529,7 @@
   bool PrepareBuiltinElementwiseMathOneArgCall(CallExpr *TheCall);
   bool PrepareBuiltinReduceMathOneArgCall(CallExpr *TheCall);
 
-  bool SemaBuiltinNonDet(CallExpr *TheCall);
+  bool SemaBuiltinNonDeterministicValue(CallExpr *TheCall);
 
   // Matrix builtin handling.
   ExprResult SemaBuiltinMatrixTranspose(CallExpr *TheCall,
Index: clang/include/clang/Basic/Builtins.def
===
--- clang/include/clang/Basic/Builtins.def
+++ clang/include/clang/Basic/Builtins.def
@@ -655,7 +655,7 @@
 BUILTIN(__builtin_alloca_with_align, "v*zIz", "Fn")
 BUILTIN(__builtin_alloca_with_align_uninitialized, "v*zIz", "Fn")
 BUILTIN(__builtin_call_with_static_chain, "v.", "nt")
-BUILTIN(__builtin_nondet, "v.", "nt")
+BUILTIN(__builtin_nondeterministic_value, "v.", "nt")
 
 BUILTIN(__builtin_elementwise_abs, "v.", "nct")
 BUILTIN(__bu

[PATCH] D142388: [clang] Add builtin_nondeterministic_value

2023-02-01 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

Can the release note and documentation update be part of this patch or should i 
create a new one ?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142388/new/

https://reviews.llvm.org/D142388

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D142388: [clang] Add builtin_nondeterministic_value

2023-02-01 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 494024.
ManuelJBrito added a comment.

- set Invalid in type checking
- add builtin documentation
- add release note




Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142388/new/

https://reviews.llvm.org/D142388

Files:
  clang/docs/LanguageExtensions.rst
  clang/docs/ReleaseNotes.rst
  clang/include/clang/Basic/Builtins.def
  clang/include/clang/Sema/Sema.h
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/lib/Sema/SemaChecking.cpp
  clang/test/CodeGen/builtins-nondet.c
  clang/test/CodeGen/builtins-nondeterministic-value.c

Index: clang/test/CodeGen/builtins-nondeterministic-value.c
===
--- clang/test/CodeGen/builtins-nondeterministic-value.c
+++ clang/test/CodeGen/builtins-nondeterministic-value.c
@@ -9,7 +9,7 @@
 // CHECK: store i32 [[X:%.*]], ptr [[A]], align 4
 // CHECK: [[R:%.*]] = freeze i32 poison
 // CHECK: ret i32 [[R]]
-  return __builtin_nondet(x);
+  return __builtin_nondeterministic_value(x);
 } 
 
 float clang_nondet_f( float x ) {
@@ -18,7 +18,7 @@
 // CHECK: store float [[X:%.*]], ptr [[A]], align 4
 // CHECK: [[R:%.*]] = freeze float poison
 // CHECK: ret float [[R]]
-  return __builtin_nondet(x);
+  return __builtin_nondeterministic_value(x);
 }
 
 double clang_nondet_d( double x ) {
@@ -27,7 +27,7 @@
 // CHECK: store double [[X:%.*]], ptr [[A]], align 8
 // CHECK: [[R:%.*]] = freeze double poison
 // CHECK: ret double [[R]]
-  return __builtin_nondet(x);
+  return __builtin_nondeterministic_value(x);
 } 
 
 _Bool clang_nondet_b( _Bool x) {
@@ -37,7 +37,7 @@
 // CHECK: store i8 [[B]], ptr [[A]], align 1
 // CHECK: [[R:%.*]] = freeze i1 poison
 // CHECK: ret i1 [[R]]
-  return __builtin_nondet(x);
+  return __builtin_nondeterministic_value(x);
 } 
 
 float4 clang_nondet_fv( float4 x ) {
@@ -46,7 +46,7 @@
 // CHECK: store <4 x float> [[X:%.*]], ptr [[A]], align 16
 // CHECK: [[R:%.*]] = freeze <4 x float> poison
 // CHECK: ret <4 x float> [[R]]
-  return __builtin_nondet(x);
+  return __builtin_nondeterministic_value(x);
 } 
 
 bool4 clang_nondet_bv( bool4 x ) {
@@ -57,6 +57,5 @@
 // CHECK: store i8 [[B]], ptr [[A]], align 1
 // CHECK: [[R:%.*]] = freeze <4 x i1> poison
 // CHECK: ret <4 x i1> [[R]]
-  return __builtin_nondet(x);
+  return __builtin_nondeterministic_value(x);
 } 
-
Index: clang/lib/Sema/SemaChecking.cpp
===
--- clang/lib/Sema/SemaChecking.cpp
+++ clang/lib/Sema/SemaChecking.cpp
@@ -2580,8 +2580,8 @@
 break;
   }
 
-  case Builtin::BI__builtin_nondet: {
-if (SemaBuiltinNonDet(TheCall))
+  case Builtin::BI__builtin_nondeterministic_value: {
+if (SemaBuiltinNonDeterministicValue(TheCall))
   return ExprError();
 break;
   }
@@ -17805,12 +17805,19 @@
   return false;
 }
 
-bool Sema::SemaBuiltinNonDet(CallExpr *TheCall) {
+bool Sema::SemaBuiltinNonDeterministicValue(CallExpr *TheCall) {
   if (checkArgCount(*this, TheCall, 1))
 return true;
 
   ExprResult Arg = TheCall->getArg(0);
   QualType TyArg = Arg.get()->getType();
+
+  if (!TyArg->isBuiltinType() && !TyArg->isVectorType()) {
+Diag(TheCall->getArg(0)->getBeginLoc(), diag::err_builtin_invalid_arg_type)
+   << 1 << /*vector, integer or floating point ty*/ 0 << TyArg;
+return true;
+  }
+
   TheCall->setType(TyArg);
   return false;
 }
Index: clang/lib/CodeGen/CGBuiltin.cpp
===
--- clang/lib/CodeGen/CGBuiltin.cpp
+++ clang/lib/CodeGen/CGBuiltin.cpp
@@ -3060,9 +3060,12 @@
 return RValue::get(V);
   }
 
-  case Builtin::BI__builtin_nondet: {
-Value *Result = PoisonValue::get(ConvertType(E->getArg(0)->getType()));
+  case Builtin::BI__builtin_nondeterministic_value: {
+llvm::Type *Ty = ConvertType(E->getArg(0)->getType()); 
+
+Value *Result = PoisonValue::get(Ty);
 Result = Builder.CreateFreeze(Result);
+
 return RValue::get(Result);
   }
 
Index: clang/include/clang/Sema/Sema.h
===
--- clang/include/clang/Sema/Sema.h
+++ clang/include/clang/Sema/Sema.h
@@ -13529,7 +13529,7 @@
   bool PrepareBuiltinElementwiseMathOneArgCall(CallExpr *TheCall);
   bool PrepareBuiltinReduceMathOneArgCall(CallExpr *TheCall);
 
-  bool SemaBuiltinNonDet(CallExpr *TheCall);
+  bool SemaBuiltinNonDeterministicValue(CallExpr *TheCall);
 
   // Matrix builtin handling.
   ExprResult SemaBuiltinMatrixTranspose(CallExpr *TheCall,
Index: clang/include/clang/Basic/Builtins.def
===
--- clang/include/clang/Basic/Builtins.def
+++ clang/include/clang/Basic/Builtins.def
@@ -655,7 +655,7 @@
 BUILTIN(__builtin_alloca_with_align, "v*zIz", "Fn")
 BUILTIN(__builtin_alloca_with_align_uninitialized, "v*zIz", "Fn")
 BUILTIN(__builtin_call_with_static_chain, "v.", "nt")
-BUILTIN(__builtin_nondet, "v.", "nt")

[PATCH] D142388: [clang] Add builtin_nondeterministic_value

2023-02-02 Thread Manuel Brito via Phabricator via cfe-commits 
This revision was landed with ongoing or failed builds.
This revision was automatically updated to reflect the committed changes.
Closed by commit rG4a1832a5c801: [Clang] Add builtin_nondeterministic_value 
(authored by ManuelJBrito).

Changed prior to commit:
  https://reviews.llvm.org/D142388?vs=494024&id=494237#toc

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142388/new/

https://reviews.llvm.org/D142388

Files:
  clang/docs/LanguageExtensions.rst
  clang/docs/ReleaseNotes.rst
  clang/include/clang/Basic/Builtins.def
  clang/include/clang/Sema/Sema.h
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/lib/Sema/SemaChecking.cpp
  clang/test/CodeGen/builtins-nondeterministic-value.c

Index: clang/test/CodeGen/builtins-nondeterministic-value.c
===
--- /dev/null
+++ clang/test/CodeGen/builtins-nondeterministic-value.c
@@ -0,0 +1,58 @@
+// RUN: %clang_cc1 %s -emit-llvm -o - | FileCheck %s
+
+typedef float float4 __attribute__((ext_vector_type(4)));
+typedef _Bool bool4 __attribute__((ext_vector_type(4)));
+
+int clang_nondet_i( int x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i32, align 4
+// CHECK: store i32 [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[R:%.*]] = freeze i32 poison
+// CHECK: ret i32 [[R]]
+  return __builtin_nondeterministic_value(x);
+}
+
+float clang_nondet_f( float x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca float, align 4
+// CHECK: store float [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[R:%.*]] = freeze float poison
+// CHECK: ret float [[R]]
+  return __builtin_nondeterministic_value(x);
+}
+
+double clang_nondet_d( double x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca double, align 8
+// CHECK: store double [[X:%.*]], ptr [[A]], align 8
+// CHECK: [[R:%.*]] = freeze double poison
+// CHECK: ret double [[R]]
+  return __builtin_nondeterministic_value(x);
+}
+
+_Bool clang_nondet_b( _Bool x) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[B:%.*]] = zext i1 %x to i8
+// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: [[R:%.*]] = freeze i1 poison
+// CHECK: ret i1 [[R]]
+  return __builtin_nondeterministic_value(x);
+}
+
+float4 clang_nondet_fv( float4 x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca <4 x float>, align 16
+// CHECK: store <4 x float> [[X:%.*]], ptr [[A]], align 16
+// CHECK: [[R:%.*]] = freeze <4 x float> poison
+// CHECK: ret <4 x float> [[R]]
+  return __builtin_nondeterministic_value(x);
+}
+
+bool4 clang_nondet_bv( bool4 x ) {
+// CHECK: [[V:%.*]] = freeze <4 x i1> poison
+// CHECK: store <4 x i1> [[V]], ptr [[P:%.*]], align 1
+// CHECK: [[R:%.*]] = load i8, ptr [[P]], align 1
+// CHECK: ret i8 [[R]]
+  return __builtin_nondeterministic_value(x);
+}
\ No newline at end of file
Index: clang/lib/Sema/SemaChecking.cpp
===
--- clang/lib/Sema/SemaChecking.cpp
+++ clang/lib/Sema/SemaChecking.cpp
@@ -2584,6 +2584,12 @@
 break;
   }
 
+  case Builtin::BI__builtin_nondeterministic_value: {
+if (SemaBuiltinNonDeterministicValue(TheCall))
+  return ExprError();
+break;
+  }
+
   // __builtin_elementwise_abs restricts the element type to signed integers or
   // floating point types only.
   case Builtin::BI__builtin_elementwise_abs: {
@@ -17857,6 +17863,21 @@
   return false;
 }
 
+bool Sema::SemaBuiltinNonDeterministicValue(CallExpr *TheCall) {
+  if (checkArgCount(*this, TheCall, 1))
+return true;
+
+  ExprResult Arg = TheCall->getArg(0);
+  QualType TyArg = Arg.get()->getType();
+
+  if (!TyArg->isBuiltinType() && !TyArg->isVectorType())
+return Diag(TheCall->getArg(0)->getBeginLoc(), diag::err_builtin_invalid_arg_type)
+   << 1 << /*vector, integer or floating point ty*/ 0 << TyArg;
+
+  TheCall->setType(TyArg);
+  return false;
+}
+
 ExprResult Sema::SemaBuiltinMatrixTranspose(CallExpr *TheCall,
 ExprResult CallResult) {
   if (checkArgCount(*this, TheCall, 1))
Index: clang/lib/CodeGen/CGBuiltin.cpp
===
--- clang/lib/CodeGen/CGBuiltin.cpp
+++ clang/lib/CodeGen/CGBuiltin.cpp
@@ -3060,6 +3060,15 @@
 return RValue::get(V);
   }
 
+  case Builtin::BI__builtin_nondeterministic_value: {
+llvm::Type *Ty = ConvertType(E->getArg(0)->getType());
+
+Value *Result = PoisonValue::get(Ty);
+Result = Builder.CreateFreeze(Result);
+
+return RValue::get(Result);
+  }
+
   case Builtin::BI__builtin_elementwise_abs: {
 Value *Result;
 QualType QT = E->getArg(0)->getType();
Index: clang/include/clang/Sema/Sema.h
===
--- clang/include/clang/Sema/Sema.h
+++ clang/include/clang/Sema/Sema.h
@@ -13571,6 +13571,8 @@
   bool PrepareBuiltinElementwiseMathOneArgCall(CallExpr *TheCall);
   bool PrepareBuiltinReduceMathOneArgCall

[PATCH] D142388: [clang] Add builtin_nondeterministic_value

2023-02-02 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

Thank you all for the reviews and helping me see this patch through.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142388/new/

https://reviews.llvm.org/D142388

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D142388: [clang] Add builtin_nondeterministic_value

2023-02-02 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

So this is getting some build failures:

- https://lab.llvm.org/buildbot#builders/38/builds/9446
- https://lab.llvm.org/buildbot#builders/245/builds/4189
- https://lab.llvm.org/buildbot#builders/65/builds/7949
- https://lab.llvm.org/buildbot#builders/188/builds/25538

I suspect it's because of the way different ABIs handle the vectors.  I was 
thinking that if we have the variable be a local variable and also not 
returning the value can fix the tests.

Does this make sense?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142388/new/

https://reviews.llvm.org/D142388

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D142388: [clang] Add builtin_nondeterministic_value

2023-02-02 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 494338.
ManuelJBrito added a comment.

Fix vector tests


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142388/new/

https://reviews.llvm.org/D142388

Files:
  clang/docs/LanguageExtensions.rst
  clang/docs/ReleaseNotes.rst
  clang/include/clang/Basic/Builtins.def
  clang/include/clang/Sema/Sema.h
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/lib/Sema/SemaChecking.cpp
  clang/test/CodeGen/builtins-nondeterministic-value.c

Index: clang/test/CodeGen/builtins-nondeterministic-value.c
===
--- /dev/null
+++ clang/test/CodeGen/builtins-nondeterministic-value.c
@@ -0,0 +1,60 @@
+// RUN: %clang_cc1 %s -emit-llvm -o - | FileCheck %s
+
+typedef float float4 __attribute__((ext_vector_type(4)));
+typedef _Bool bool4 __attribute__((ext_vector_type(4)));
+
+int clang_nondet_i( int x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i32, align 4
+// CHECK: store i32 [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[R:%.*]] = freeze i32 poison
+// CHECK: ret i32 [[R]]
+  return __builtin_nondeterministic_value(x);
+}
+
+float clang_nondet_f( float x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca float, align 4
+// CHECK: store float [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[R:%.*]] = freeze float poison
+// CHECK: ret float [[R]]
+  return __builtin_nondeterministic_value(x);
+}
+
+double clang_nondet_d( double x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca double, align 8
+// CHECK: store double [[X:%.*]], ptr [[A]], align 8
+// CHECK: [[R:%.*]] = freeze double poison
+// CHECK: ret double [[R]]
+  return __builtin_nondeterministic_value(x);
+}
+
+_Bool clang_nondet_b( _Bool x) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[B:%.*]] = zext i1 %x to i8
+// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: [[R:%.*]] = freeze i1 poison
+// CHECK: ret i1 [[R]]
+  return __builtin_nondeterministic_value(x);
+}
+
+void clang_nondet_fv( ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca <4 x float>, align 16
+// CHECK: [[R:%.*]] = freeze <4 x float> poison
+// CHECK: store <4 x float> [[R]], ptr [[A]], align 16
+// CHECK: ret void
+  float4 x = __builtin_nondeterministic_value(x);
+}
+
+void clang_nondet_bv( ) {
+// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[V:%.*]] = freeze <4 x i1> poison
+// CHECK: [[SV:%.*]] = shufflevector <4 x i1> [[V]], <4 x i1> poison, <8 x i32> 
+// CHECK: [[BC:%.*]] = bitcast <8 x i1> [[SV]] to i8
+// CHECK: store i8 [[BC]], ptr [[A]], align 1
+// CHECK: ret void
+  bool4 x = __builtin_nondeterministic_value(x);
+}
\ No newline at end of file
Index: clang/lib/Sema/SemaChecking.cpp
===
--- clang/lib/Sema/SemaChecking.cpp
+++ clang/lib/Sema/SemaChecking.cpp
@@ -2584,6 +2584,12 @@
 break;
   }
 
+  case Builtin::BI__builtin_nondeterministic_value: {
+if (SemaBuiltinNonDeterministicValue(TheCall))
+  return ExprError();
+break;
+  }
+
   // __builtin_elementwise_abs restricts the element type to signed integers or
   // floating point types only.
   case Builtin::BI__builtin_elementwise_abs: {
@@ -17857,6 +17863,21 @@
   return false;
 }
 
+bool Sema::SemaBuiltinNonDeterministicValue(CallExpr *TheCall) {
+  if (checkArgCount(*this, TheCall, 1))
+return true;
+
+  ExprResult Arg = TheCall->getArg(0);
+  QualType TyArg = Arg.get()->getType();
+
+  if (!TyArg->isBuiltinType() && !TyArg->isVectorType())
+return Diag(TheCall->getArg(0)->getBeginLoc(), diag::err_builtin_invalid_arg_type)
+   << 1 << /*vector, integer or floating point ty*/ 0 << TyArg;
+
+  TheCall->setType(TyArg);
+  return false;
+}
+
 ExprResult Sema::SemaBuiltinMatrixTranspose(CallExpr *TheCall,
 ExprResult CallResult) {
   if (checkArgCount(*this, TheCall, 1))
Index: clang/lib/CodeGen/CGBuiltin.cpp
===
--- clang/lib/CodeGen/CGBuiltin.cpp
+++ clang/lib/CodeGen/CGBuiltin.cpp
@@ -3060,6 +3060,15 @@
 return RValue::get(V);
   }
 
+  case Builtin::BI__builtin_nondeterministic_value: {
+llvm::Type *Ty = ConvertType(E->getArg(0)->getType());
+
+Value *Result = PoisonValue::get(Ty);
+Result = Builder.CreateFreeze(Result);
+
+return RValue::get(Result);
+  }
+
   case Builtin::BI__builtin_elementwise_abs: {
 Value *Result;
 QualType QT = E->getArg(0)->getType();
Index: clang/include/clang/Sema/Sema.h
===
--- clang/include/clang/Sema/Sema.h
+++ clang/include/clang/Sema/Sema.h
@@ -13571,6 +13571,8 @@
   bool PrepareBuiltinElementwiseMathOneArgCall(CallExpr *TheCall);
   bool PrepareBuiltinReduceMathOneArgCall(CallExpr *TheCall);
 
+  bool SemaBuiltinNonDeterministicValue(CallExpr *TheCall);
+
   // Matrix builtin handling.
   ExprResult SemaB

[PATCH] D142388: [clang] Add builtin_nondeterministic_value

2023-02-03 Thread Manuel Brito via Phabricator via cfe-commits 
This revision was landed with ongoing or failed builds.
This revision was automatically updated to reflect the committed changes.
Closed by commit rG450a4612c39c: [Clang] Add builtin_nondeterministic_value 
(authored by ManuelJBrito).

Changed prior to commit:
  https://reviews.llvm.org/D142388?vs=494338&id=494552#toc

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142388/new/

https://reviews.llvm.org/D142388

Files:
  clang/docs/LanguageExtensions.rst
  clang/docs/ReleaseNotes.rst
  clang/include/clang/Basic/Builtins.def
  clang/include/clang/Sema/Sema.h
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/lib/Sema/SemaChecking.cpp
  clang/test/CodeGen/builtins-nondeterministic-value.c

Index: clang/test/CodeGen/builtins-nondeterministic-value.c
===
--- /dev/null
+++ clang/test/CodeGen/builtins-nondeterministic-value.c
@@ -0,0 +1,60 @@
+// RUN: %clang_cc1 %s -emit-llvm -o - | FileCheck %s
+
+typedef float float4 __attribute__((ext_vector_type(4)));
+typedef _Bool bool4 __attribute__((ext_vector_type(4)));
+
+int clang_nondet_i( int x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i32, align 4
+// CHECK: store i32 [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[R:%.*]] = freeze i32 poison
+// CHECK: ret i32 [[R]]
+  return __builtin_nondeterministic_value(x);
+}
+
+float clang_nondet_f( float x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca float, align 4
+// CHECK: store float [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[R:%.*]] = freeze float poison
+// CHECK: ret float [[R]]
+  return __builtin_nondeterministic_value(x);
+}
+
+double clang_nondet_d( double x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca double, align 8
+// CHECK: store double [[X:%.*]], ptr [[A]], align 8
+// CHECK: [[R:%.*]] = freeze double poison
+// CHECK: ret double [[R]]
+  return __builtin_nondeterministic_value(x);
+}
+
+_Bool clang_nondet_b( _Bool x) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[B:%.*]] = zext i1 %x to i8
+// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: [[R:%.*]] = freeze i1 poison
+// CHECK: ret i1 [[R]]
+  return __builtin_nondeterministic_value(x);
+}
+
+void clang_nondet_fv( ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca <4 x float>, align 16
+// CHECK: [[R:%.*]] = freeze <4 x float> poison
+// CHECK: store <4 x float> [[R]], ptr [[A]], align 16
+// CHECK: ret void
+  float4 x = __builtin_nondeterministic_value(x);
+}
+
+void clang_nondet_bv( ) {
+// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[V:%.*]] = freeze <4 x i1> poison
+// CHECK: [[SV:%.*]] = shufflevector <4 x i1> [[V]], <4 x i1> poison, <8 x i32> 
+// CHECK: [[BC:%.*]] = bitcast <8 x i1> [[SV]] to i8
+// CHECK: store i8 [[BC]], ptr [[A]], align 1
+// CHECK: ret void
+  bool4 x = __builtin_nondeterministic_value(x);
+}
Index: clang/lib/Sema/SemaChecking.cpp
===
--- clang/lib/Sema/SemaChecking.cpp
+++ clang/lib/Sema/SemaChecking.cpp
@@ -2584,6 +2584,12 @@
 break;
   }
 
+  case Builtin::BI__builtin_nondeterministic_value: {
+if (SemaBuiltinNonDeterministicValue(TheCall))
+  return ExprError();
+break;
+  }
+
   // __builtin_elementwise_abs restricts the element type to signed integers or
   // floating point types only.
   case Builtin::BI__builtin_elementwise_abs: {
@@ -17858,6 +17864,21 @@
   return false;
 }
 
+bool Sema::SemaBuiltinNonDeterministicValue(CallExpr *TheCall) {
+  if (checkArgCount(*this, TheCall, 1))
+return true;
+
+  ExprResult Arg = TheCall->getArg(0);
+  QualType TyArg = Arg.get()->getType();
+
+  if (!TyArg->isBuiltinType() && !TyArg->isVectorType())
+return Diag(TheCall->getArg(0)->getBeginLoc(), diag::err_builtin_invalid_arg_type)
+   << 1 << /*vector, integer or floating point ty*/ 0 << TyArg;
+
+  TheCall->setType(TyArg);
+  return false;
+}
+
 ExprResult Sema::SemaBuiltinMatrixTranspose(CallExpr *TheCall,
 ExprResult CallResult) {
   if (checkArgCount(*this, TheCall, 1))
Index: clang/lib/CodeGen/CGBuiltin.cpp
===
--- clang/lib/CodeGen/CGBuiltin.cpp
+++ clang/lib/CodeGen/CGBuiltin.cpp
@@ -3060,6 +3060,15 @@
 return RValue::get(V);
   }
 
+  case Builtin::BI__builtin_nondeterministic_value: {
+llvm::Type *Ty = ConvertType(E->getArg(0)->getType());
+
+Value *Result = PoisonValue::get(Ty);
+Result = Builder.CreateFreeze(Result);
+
+return RValue::get(Result);
+  }
+
   case Builtin::BI__builtin_elementwise_abs: {
 Value *Result;
 QualType QT = E->getArg(0)->getType();
Index: clang/include/clang/Sema/Sema.h
===
--- clang/include/clang/Sema/Sema.h
+++ clang/include/clang/Sema/Sema.h
@@ -13571,6 +13571,8 @@
   bool PrepareBuiltinElementwiseMathOneArgCall(CallEx

[PATCH] D143287: [Clang][x86] Change x86 cast intrinsics to use __builtin_nondeterministic_value

2023-02-03 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito created this revision.
ManuelJBrito added reviewers: craig.topper, RKSimon.
Herald added subscribers: pengfei, mgrang.
Herald added a project: All.
ManuelJBrito requested review of this revision.
Herald added a project: clang.
Herald added a subscriber: cfe-commits.

The following intrinsics are currently implemented using a shufflevector with 
an undefined mask, this is however incorrect according to intel's semantics for 
undefined value which expect an unknown but consistent value.

With `__builtin_nondeterministic_value` we can now match intel's undefined 
value.

Related patch for more context : https://reviews.llvm.org/D103874


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D143287

Files:
  clang/lib/Headers/avx512fintrin.h
  clang/lib/Headers/avx512fp16intrin.h
  clang/lib/Headers/avxintrin.h
  clang/test/CodeGen/X86/avx-builtins.c
  clang/test/CodeGen/X86/avx512f-builtins.c
  clang/test/CodeGen/X86/avx512fp16-builtins.c

Index: clang/test/CodeGen/X86/avx512fp16-builtins.c
===
--- clang/test/CodeGen/X86/avx512fp16-builtins.c
+++ clang/test/CodeGen/X86/avx512fp16-builtins.c
@@ -326,19 +326,19 @@
 
 __m256h test_mm256_castph128_ph256(__m128h __a) {
   // CHECK-LABEL: test_mm256_castph128_ph256
-  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <16 x i32> 
+  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <16 x i32> 
   return _mm256_castph128_ph256(__a);
 }
 
 __m512h test_mm512_castph128_ph512(__m128h __a) {
   // CHECK-LABEL: test_mm512_castph128_ph512
-  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <32 x i32> 
+  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <32 x i32> 
   return _mm512_castph128_ph512(__a);
 }
 
 __m512h test_mm512_castph256_ph512(__m256h __a) {
   // CHECK-LABEL: test_mm512_castph256_ph512
-  // CHECK: shufflevector <16 x half> %{{.*}}, <16 x half> %{{.*}}, <32 x i32> 
+  // CHECK: shufflevector <16 x half> %{{.*}}, <16 x half> %{{.*}}, <32 x i32> 
   return _mm512_castph256_ph512(__a);
 }
 
Index: clang/test/CodeGen/X86/avx512f-builtins.c
===
--- clang/test/CodeGen/X86/avx512f-builtins.c
+++ clang/test/CodeGen/X86/avx512f-builtins.c
@@ -8987,13 +8987,13 @@
 
 __m512 test_mm512_castps128_ps512(__m128 __A) {
   // CHECK-LABEL: @test_mm512_castps128_ps512
-  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <16 x i32> 
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <16 x i32> 
   return _mm512_castps128_ps512(__A); 
 }
 
 __m512d test_mm512_castpd128_pd512(__m128d __A) {
   // CHECK-LABEL: @test_mm512_castpd128_pd512
-  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <8 x i32> 
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <8 x i32> 
   return _mm512_castpd128_pd512(__A); 
 }
 
@@ -9086,7 +9086,7 @@
 __m512d test_mm512_castpd256_pd512(__m256d a)
 {
   // CHECK-LABEL: @test_mm512_castpd256_pd512
-  // CHECK: shufflevector <4 x double> {{.*}} 
+  // CHECK: shufflevector <4 x double> {{.*}} 
   return _mm512_castpd256_pd512(a);
 }
 
@@ -9112,13 +9112,13 @@
 }
 __m512i test_mm512_castsi128_si512(__m128i __A) {
   // CHECK-LABEL: @test_mm512_castsi128_si512
-  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <8 x i32> 
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <8 x i32> 
   return _mm512_castsi128_si512(__A); 
 }
 
 __m512i test_mm512_castsi256_si512(__m256i __A) {
   // CHECK-LABEL: @test_mm512_castsi256_si512
-  // CHECK: shufflevector <4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <8 x i32> 
+  // CHECK: shufflevector <4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <8 x i32> 
   return _mm512_castsi256_si512(__A); 
 }
 
Index: clang/test/CodeGen/X86/avx-builtins.c
===
--- clang/test/CodeGen/X86/avx-builtins.c
+++ clang/test/CodeGen/X86/avx-builtins.c
@@ -143,7 +143,7 @@
 
 __m256d test_mm256_castpd128_pd256(__m128d A) {
   // CHECK-LABEL: test_mm256_castpd128_pd256
-  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <4 x i32> 
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <4 x i32> 
   return _mm256_castpd128_pd256(A);
 }
 
@@ -165,7 +165,7 @@
 
 __m256 test_mm256_castps128_ps256(__m128 A) {
   // CHECK-LABEL: test_mm256_castps128_ps256
-  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <8 x i32> 
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <8 x i32> 
   return _mm256_castps128_ps256(A);
 }
 
@@ -177,7 +177,7 @@
 
 __m256i test_mm256_castsi128_si256(__m128i A) {
   // CHECK-LABEL: test_mm256_castsi128_si256
-  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <4 x i32> 
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <4 x i32> 
   return _mm256_castsi128_si256(A);
 }
 
Index: clang/l

[PATCH] D143287: [Clang][x86] Change x86 cast intrinsics to use __builtin_nondeterministic_value

2023-02-06 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

In D143287#4103597 , @craig.topper 
wrote:

> We have a couple bugs that show (freeze (poison)) doesn't work past 
> SelectionDAG. Is that a concern here? The most recent 
> https://github.com/llvm/llvm-project/issues/60429

I don't think it's a concern . Here https://godbolt.org/z/1ecM8roYh the 
lowering seems ok.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D143287/new/

https://reviews.llvm.org/D143287

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D143287: [Clang][x86] Change x86 cast intrinsics to use __builtin_nondeterministic_value

2023-02-06 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

In D143287#4104056 , @RKSimon wrote:

> What do we gain from using __builtin_nondeterministic_value instead of just 
> setzero? https://godbolt.org/z/zrb6858Mr

__builtin_nondeterministic_value is lowered to freeze(poison)

These intrinsics are meant to be no-op and we can't to that with 
zeroinitializer, although as of right now instcombine seems to transform the 
freeze(poison) into zeroinitializer(as shown by your example), so i'll make 
another patch to fix that.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D143287/new/

https://reviews.llvm.org/D143287

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D142388: [clang] Add builtin_nondet

2023-01-23 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito created this revision.
ManuelJBrito added reviewers: aaron.ballman, erichkeane, shafik.
ManuelJBrito added a project: clang.
Herald added a project: All.
ManuelJBrito requested review of this revision.
Herald added a subscriber: cfe-commits.

This patch adds a builtin that returns a non deterministic value of the same 
type as the argument.

Example:
`int x = __builtin_nondet(x);`

Which is then lowered to freeze(poison).

This is a continuation of the work done in https://reviews.llvm.org/D136737, 
following the provided feedback.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D142388

Files:
  clang/include/clang/Basic/Builtins.def
  clang/include/clang/Sema/Sema.h
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/lib/Sema/SemaChecking.cpp
  clang/test/CodeGen/builtins-nondet.c

Index: clang/test/CodeGen/builtins-nondet.c
===
--- /dev/null
+++ clang/test/CodeGen/builtins-nondet.c
@@ -0,0 +1,40 @@
+// RUN: %clang_cc1 %s -emit-llvm -o - | FileCheck %s
+
+typedef float float4 __attribute__((ext_vector_type(4)));
+
+int clang_nondet_i( int x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca i32, align 4
+// CHECK: store i32 [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[R:%.*]] = freeze i32 poison
+// CHECK: ret i32 [[R]]
+  return __builtin_nondet(x);
+} 
+
+float clang_nondet_f( float x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca float, align 4
+// CHECK: store float [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[R:%.*]] = freeze float poison
+// CHECK: ret float [[R]]
+  return __builtin_nondet(x);
+}
+
+double clang_nondet_d( double x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca double, align 8
+// CHECK: store double [[X:%.*]], ptr [[A]], align 8
+// CHECK: [[R:%.*]] = freeze double poison
+// CHECK: ret double [[R]]
+  return __builtin_nondet(x);
+} 
+
+float4 clang_nondet_v( float4 x ) {
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca <4 x float>, align 16
+// CHECK: store <4 x float> [[X:%.*]], ptr [[A]], align 16
+// CHECK: [[R:%.*]] = freeze <4 x float> poison
+// CHECK: ret <4 x float> [[R]]
+  return __builtin_nondet(x);
+} 
+
Index: clang/lib/Sema/SemaChecking.cpp
===
--- clang/lib/Sema/SemaChecking.cpp
+++ clang/lib/Sema/SemaChecking.cpp
@@ -2580,6 +2580,12 @@
 break;
   }
 
+  case Builtin::BI__builtin_nondet: {
+if (SemaBuiltinNonDet(TheCall))
+  return ExprError();
+break;
+  }
+
   // __builtin_elementwise_abs restricts the element type to signed integers or
   // floating point types only.
   case Builtin::BI__builtin_elementwise_abs: {
@@ -17799,6 +17805,16 @@
   return false;
 }
 
+bool Sema::SemaBuiltinNonDet(CallExpr *TheCall) {
+  if (checkArgCount(*this, TheCall, 1))
+return true;
+
+  ExprResult Arg = TheCall->getArg(0);
+  QualType TyArg = Arg.get()->getType();
+  TheCall->setType(TyArg);
+  return false;
+}
+
 ExprResult Sema::SemaBuiltinMatrixTranspose(CallExpr *TheCall,
 ExprResult CallResult) {
   if (checkArgCount(*this, TheCall, 1))
Index: clang/lib/CodeGen/CGBuiltin.cpp
===
--- clang/lib/CodeGen/CGBuiltin.cpp
+++ clang/lib/CodeGen/CGBuiltin.cpp
@@ -3060,6 +3060,12 @@
 return RValue::get(V);
   }
 
+  case Builtin::BI__builtin_nondet: {
+Value *Result = PoisonValue::get(ConvertType(E->getArg(0)->getType()));
+Result = Builder.CreateFreeze(Result);
+return RValue::get(Result);
+  }
+
   case Builtin::BI__builtin_elementwise_abs: {
 Value *Result;
 QualType QT = E->getArg(0)->getType();
Index: clang/include/clang/Sema/Sema.h
===
--- clang/include/clang/Sema/Sema.h
+++ clang/include/clang/Sema/Sema.h
@@ -13529,6 +13529,8 @@
   bool PrepareBuiltinElementwiseMathOneArgCall(CallExpr *TheCall);
   bool PrepareBuiltinReduceMathOneArgCall(CallExpr *TheCall);
 
+  bool SemaBuiltinNonDet(CallExpr *TheCall);
+
   // Matrix builtin handling.
   ExprResult SemaBuiltinMatrixTranspose(CallExpr *TheCall,
 ExprResult CallResult);
Index: clang/include/clang/Basic/Builtins.def
===
--- clang/include/clang/Basic/Builtins.def
+++ clang/include/clang/Basic/Builtins.def
@@ -655,6 +655,7 @@
 BUILTIN(__builtin_alloca_with_align, "v*zIz", "Fn")
 BUILTIN(__builtin_alloca_with_align_uninitialized, "v*zIz", "Fn")
 BUILTIN(__builtin_call_with_static_chain, "v.", "nt")
+BUILTIN(__builtin_nondet, "v.", "nt")
 
 BUILTIN(__builtin_elementwise_abs, "v.", "nct")
 BUILTIN(__builtin_elementwise_max, "v.", "nct")
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D152658: [InstCombine] Change SimplifyDemandedVectorElts to use PoisonElts instead of UndefElts

2023-07-18 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito planned changes to this revision.
ManuelJBrito added a comment.

Hello @aqjune thanks for the comments !! Some users of 
SimplifyDemandedVectorElts might need some more tweaking to prevent regressions 
for the cases you noted. I'll work on this as soon as I can!


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D152658/new/

https://reviews.llvm.org/D152658

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D140538: [Clang][CodeGen] Use poison instead of undef for dummy values in CGExpr [NFC]

2023-09-11 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 556416.
ManuelJBrito removed a reviewer: libc++abi.
ManuelJBrito added a comment.
This revision is now accepted and ready to land.

Remove changes to demangle test


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D140538/new/

https://reviews.llvm.org/D140538

Files:
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/lib/CodeGen/CGCall.cpp
  clang/lib/CodeGen/CGExpr.cpp
  clang/lib/CodeGen/CGExprComplex.cpp
  clang/lib/CodeGen/CGExprScalar.cpp
  clang/lib/CodeGen/CodeGenFunction.h

Index: clang/lib/CodeGen/CodeGenFunction.h
===
--- clang/lib/CodeGen/CodeGenFunction.h
+++ clang/lib/CodeGen/CodeGenFunction.h
@@ -3774,8 +3774,8 @@
   /// Create a check that a scalar RValue is non-null.
   llvm::Value *EmitNonNullRValueCheck(RValue RV, QualType T);
 
-  /// GetUndefRValue - Get an appropriate 'undef' rvalue for the given type.
-  RValue GetUndefRValue(QualType Ty);
+  /// GetPoisonRValue - Get an appropriate 'poison' rvalue for the given type.
+  RValue GetPoisonRValue(QualType Ty);
 
   /// EmitUnsupportedRValue - Emit a dummy r-value using the type of E
   /// and issue an ErrorUnsupported style diagnostic (using the
Index: clang/lib/CodeGen/CGExprScalar.cpp
===
--- clang/lib/CodeGen/CGExprScalar.cpp
+++ clang/lib/CodeGen/CGExprScalar.cpp
@@ -1630,7 +1630,7 @@
   CGF.ErrorUnsupported(E, "scalar expression");
   if (E->getType()->isVoidType())
 return nullptr;
-  return llvm::UndefValue::get(CGF.ConvertType(E->getType()));
+  return llvm::PoisonValue::get(CGF.ConvertType(E->getType()));
 }
 
 Value *
@@ -4958,7 +4958,7 @@
   // If EmitVAArg fails, emit an error.
   if (!ArgPtr.isValid()) {
 CGF.ErrorUnsupported(VE, "va_arg expression");
-return llvm::UndefValue::get(ArgTy);
+return llvm::PoisonValue::get(ArgTy);
   }
 
   // FIXME Volatility.
Index: clang/lib/CodeGen/CGExprComplex.cpp
===
--- clang/lib/CodeGen/CGExprComplex.cpp
+++ clang/lib/CodeGen/CGExprComplex.cpp
@@ -415,7 +415,7 @@
   CGF.ErrorUnsupported(E, "complex expression");
   llvm::Type *EltTy =
 CGF.ConvertType(getComplexType(E->getType())->getElementType());
-  llvm::Value *U = llvm::UndefValue::get(EltTy);
+  llvm::Value *U = llvm::PoisonValue::get(EltTy);
   return ComplexPairTy(U, U);
 }
 
@@ -1275,7 +1275,7 @@
 CGF.ErrorUnsupported(E, "complex va_arg expression");
 llvm::Type *EltTy =
   CGF.ConvertType(E->getType()->castAs()->getElementType());
-llvm::Value *U = llvm::UndefValue::get(EltTy);
+llvm::Value *U = llvm::PoisonValue::get(EltTy);
 return ComplexPairTy(U, U);
   }
 
Index: clang/lib/CodeGen/CGExpr.cpp
===
--- clang/lib/CodeGen/CGExpr.cpp
+++ clang/lib/CodeGen/CGExpr.cpp
@@ -1176,7 +1176,7 @@
   return Builder.CreateICmpNE(V, llvm::Constant::getNullValue(V->getType()));
 }
 
-RValue CodeGenFunction::GetUndefRValue(QualType Ty) {
+RValue CodeGenFunction::GetPoisonRValue(QualType Ty) {
   if (Ty->isVoidType())
 return RValue::get(nullptr);
 
@@ -1184,7 +1184,7 @@
   case TEK_Complex: {
 llvm::Type *EltTy =
   ConvertType(Ty->castAs()->getElementType());
-llvm::Value *U = llvm::UndefValue::get(EltTy);
+llvm::Value *U = llvm::PoisonValue::get(EltTy);
 return RValue::getComplex(std::make_pair(U, U));
   }
 
@@ -1197,7 +1197,7 @@
   }
 
   case TEK_Scalar:
-return RValue::get(llvm::UndefValue::get(ConvertType(Ty)));
+return RValue::get(llvm::PoisonValue::get(ConvertType(Ty)));
   }
   llvm_unreachable("bad evaluation kind");
 }
@@ -1205,7 +1205,7 @@
 RValue CodeGenFunction::EmitUnsupportedRValue(const Expr *E,
   const char *Name) {
   ErrorUnsupported(E, Name);
-  return GetUndefRValue(E->getType());
+  return GetPoisonRValue(E->getType());
 }
 
 LValue CodeGenFunction::EmitUnsupportedLValue(const Expr *E,
@@ -1214,7 +1214,7 @@
   llvm::Type *ElTy = ConvertType(E->getType());
   llvm::Type *Ty = llvm::PointerType::getUnqual(ElTy);
   return MakeAddrLValue(
-  Address(llvm::UndefValue::get(Ty), ElTy, CharUnits::One()), E->getType());
+  Address(llvm::PoisonValue::get(Ty), ElTy, CharUnits::One()), E->getType());
 }
 
 bool CodeGenFunction::IsWrappedCXXThis(const Expr *Obj) {
Index: clang/lib/CodeGen/CGCall.cpp
===
--- clang/lib/CodeGen/CGCall.cpp
+++ clang/lib/CodeGen/CGCall.cpp
@@ -5712,7 +5712,7 @@
 EnsureInsertPoint();
 
 // Return a reasonable RValue.
-return GetUndefRValue(RetTy);
+return GetPoisonRValue(RetTy);
   }
 
   // If this is a musttail call, return immediately. We do not branch to the
@@ -5730,7 +5730,7 @@
   Builder.CreateRet(CI);
 Builder.ClearInsertionPoint();
 EnsureInsertPo

[PATCH] D152658: [InstCombine] Change SimplifyDemandedVectorElts to use PoisonElts instead of UndefElts

2023-07-11 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.
Herald added a subscriber: wangpc.

ping :)


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D152658/new/

https://reviews.llvm.org/D152658

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D138755: [Clang][CodeGen] Use poison instead of undef for extra argument in __builtin_amdgcn_mov_dpp [NFC]

2022-11-27 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito created this revision.
ManuelJBrito added a project: AMDGPU.
Herald added subscribers: kosarev, jvesely.
Herald added a project: All.
ManuelJBrito requested review of this revision.
Herald added a project: clang.
Herald added a subscriber: cfe-commits.

Use poison instead of undef when emitting llvm.amdgcn.update.dpp for 
__builtin_amdgcn_mov_dpp.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D138755

Files:
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/test/CodeGenOpenCL/builtins-amdgcn-vi.cl


Index: clang/test/CodeGenOpenCL/builtins-amdgcn-vi.cl
===
--- clang/test/CodeGenOpenCL/builtins-amdgcn-vi.cl
+++ clang/test/CodeGenOpenCL/builtins-amdgcn-vi.cl
@@ -101,7 +101,7 @@
 }
 
 // CHECK-LABEL: @test_mov_dpp
-// CHECK: call i32 @llvm.amdgcn.update.dpp.i32(i32 undef, i32 %src, i32 0, i32 
0, i32 0, i1 false)
+// CHECK: call i32 @llvm.amdgcn.update.dpp.i32(i32 poison, i32 %src, i32 0, 
i32 0, i32 0, i1 false)
 void test_mov_dpp(global int* out, int src)
 {
   *out = __builtin_amdgcn_mov_dpp(src, 0, 0, 0, false);
Index: clang/lib/CodeGen/CGBuiltin.cpp
===
--- clang/lib/CodeGen/CGBuiltin.cpp
+++ clang/lib/CodeGen/CGBuiltin.cpp
@@ -16893,7 +16893,7 @@
   Args.push_back(EmitScalarExpr(E->getArg(I)));
 assert(Args.size() == 5 || Args.size() == 6);
 if (Args.size() == 5)
-  Args.insert(Args.begin(), llvm::UndefValue::get(Args[0]->getType()));
+  Args.insert(Args.begin(), llvm::PoisonValue::get(Args[0]->getType()));
 Function *F =
 CGM.getIntrinsic(Intrinsic::amdgcn_update_dpp, Args[0]->getType());
 return Builder.CreateCall(F, Args);


Index: clang/test/CodeGenOpenCL/builtins-amdgcn-vi.cl
===
--- clang/test/CodeGenOpenCL/builtins-amdgcn-vi.cl
+++ clang/test/CodeGenOpenCL/builtins-amdgcn-vi.cl
@@ -101,7 +101,7 @@
 }
 
 // CHECK-LABEL: @test_mov_dpp
-// CHECK: call i32 @llvm.amdgcn.update.dpp.i32(i32 undef, i32 %src, i32 0, i32 0, i32 0, i1 false)
+// CHECK: call i32 @llvm.amdgcn.update.dpp.i32(i32 poison, i32 %src, i32 0, i32 0, i32 0, i1 false)
 void test_mov_dpp(global int* out, int src)
 {
   *out = __builtin_amdgcn_mov_dpp(src, 0, 0, 0, false);
Index: clang/lib/CodeGen/CGBuiltin.cpp
===
--- clang/lib/CodeGen/CGBuiltin.cpp
+++ clang/lib/CodeGen/CGBuiltin.cpp
@@ -16893,7 +16893,7 @@
   Args.push_back(EmitScalarExpr(E->getArg(I)));
 assert(Args.size() == 5 || Args.size() == 6);
 if (Args.size() == 5)
-  Args.insert(Args.begin(), llvm::UndefValue::get(Args[0]->getType()));
+  Args.insert(Args.begin(), llvm::PoisonValue::get(Args[0]->getType()));
 Function *F =
 CGM.getIntrinsic(Intrinsic::amdgcn_update_dpp, Args[0]->getType());
 return Builder.CreateCall(F, Args);
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D138755: [Clang][CodeGen] Use poison instead of undef for extra argument in __builtin_amdgcn_mov_dpp [NFC]

2022-11-27 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a reviewer: yaxunl.
ManuelJBrito added a comment.

Couldn't find documentation for this intrinsic, can you shed a light on whether 
or not this change is OK?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D138755/new/

https://reviews.llvm.org/D138755

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D143287: [Clang][X86] Change X86 cast intrinsics to use __builtin_nondeterministic_value

2023-03-17 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito planned changes to this revision.
ManuelJBrito added a comment.

Implementing the 128 to 512 casts by filling the rest of the vector with the 
same definition of a nondeterministic_value is not correct because  :

a = freeze poison
v = 

is not the same as

v = freeze poison

The only solution I'm seeing ,using the shufflevector, is doing the conversion 
in two steps:

- build a 256 vector with the upper half being undefined( freeze poison)
- build a 512 vector where the lower half is the previous 256 vector and the 
upper half being undefined

I think this would require two shuffles which is unfortunate.

This would ensure no miscompilations due to multiple uses of the same freeze 
undef/poison but would probably require some backend work to ensure the pattern 
is recognized to emit efficient assembly.

Would this work ? @RKSimon @craig.topper


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D143287/new/

https://reviews.llvm.org/D143287

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D144903: [X86] Drop single use check for freeze(undef) in LowerAVXCONCAT_VECTORS

2023-02-27 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito created this revision.
ManuelJBrito added a reviewer: RKSimon.
Herald added subscribers: pengfei, hiraditya.
Herald added a project: All.
ManuelJBrito requested review of this revision.
Herald added projects: clang, LLVM.
Herald added subscribers: llvm-commits, cfe-commits.

Ignoring freeze(undef) if it has multiple uses in LowerAVXCONCAT_VECTORS causes 
the custom INSERT_SUBVECTOR for vector widening to be ignored.

For example in https://godbolt.org/z/7hacPe1KM extra vinsertf128 instructions 
are introduced.

This is necessary to lower `mm512_cast*128` intel intrinsics as no-op 
intrinsics.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D144903

Files:
  clang/test/CodeGen/X86/avx-cast-builtins.c
  llvm/lib/Target/X86/X86ISelLowering.cpp
  llvm/test/CodeGen/X86/avx512-intrinsics.ll
  llvm/test/CodeGen/X86/avx512fp16-intrinsics.ll

Index: llvm/test/CodeGen/X86/avx512fp16-intrinsics.ll
===
--- llvm/test/CodeGen/X86/avx512fp16-intrinsics.ll
+++ llvm/test/CodeGen/X86/avx512fp16-intrinsics.ll
@@ -1231,10 +1231,7 @@
 define <32 x half> @test_mm512_castph128_ph512_freeze(<8 x half> %a0) nounwind {
 ; CHECK-LABEL: test_mm512_castph128_ph512_freeze:
 ; CHECK:   # %bb.0:
-; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm1
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; CHECK-NEXT:vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
+; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $zmm0
 ; CHECK-NEXT:retq
   %a1 = freeze <8 x half> poison
   %res = shufflevector <8 x half> %a0, <8 x half> %a1, <32 x i32> 
Index: llvm/test/CodeGen/X86/avx512-intrinsics.ll
===
--- llvm/test/CodeGen/X86/avx512-intrinsics.ll
+++ llvm/test/CodeGen/X86/avx512-intrinsics.ll
@@ -7495,10 +7495,7 @@
 define <8 x double> @test_mm256_castpd128_pd256_freeze(<2 x double> %a0) nounwind {
 ; CHECK-LABEL: test_mm256_castpd128_pd256_freeze:
 ; CHECK:   # %bb.0:
-; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm1
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; CHECK-NEXT:vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
+; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $zmm0
 ; CHECK-NEXT:ret{{[l|q]}}
   %a1 = freeze <2 x double> poison
   %res = shufflevector <2 x double> %a0, <2 x double> %a1, <8 x i32> 
@@ -7520,10 +7517,7 @@
 define <16 x float> @test_mm256_castps128_ps512_freeze(<4 x float> %a0) nounwind {
 ; CHECK-LABEL: test_mm256_castps128_ps512_freeze:
 ; CHECK:   # %bb.0:
-; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm1
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; CHECK-NEXT:vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
+; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $zmm0
 ; CHECK-NEXT:ret{{[l|q]}}
   %a1 = freeze <4 x float> poison
   %res = shufflevector <4 x float> %a0, <4 x float> %a1, <16x i32> 
@@ -7545,10 +7539,7 @@
 define <8 x i64> @test_mm512_castsi128_si512_freeze(<2 x i64> %a0) nounwind {
 ; CHECK-LABEL: test_mm512_castsi128_si512_freeze:
 ; CHECK:   # %bb.0:
-; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm1
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; CHECK-NEXT:vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
+; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $zmm0
 ; CHECK-NEXT:ret{{[l|q]}}
   %a1 = freeze <2 x i64> poison
   %res = shufflevector <2 x i64> %a0, <2 x i64> %a1, <8 x i32> 
Index: llvm/lib/Target/X86/X86ISelLowering.cpp
===
--- llvm/lib/Target/X86/X86ISelLowering.cpp
+++ llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -11656,7 +11656,7 @@
 SDValue SubVec = Op.getOperand(i);
 if (SubVec.isUndef())
   continue;
-if (ISD::isFreezeUndef(SubVec.getNode()) && SubVec.hasOneUse())
+if (ISD::isFreezeUndef(SubVec.getNode()))
   ++NumFreezeUndef;
 else if (ISD::isBuildVectorAllZeros(SubVec.getNode()))
   ++NumZero;
Index: clang/test/CodeGen/X86/avx-cast-builtins.c
===
--- clang/test/CodeGen/X86/avx-cast-builtins.c
+++ clang/test/CodeGen/X86/avx-cast-builtins.c
@@ -1,4 +1,4 @@
-// RUN: %clang_cc1 %s -O3 -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-unknown-unknown -target-feature +avx -target-feature +avx512f  -target-feature +avx512fp16 -S -o - | FileCheck %s
+// RUN: %clang_cc1 %s -O3  -ffreestanding %s -triple=x86_64-unknown-unknown -target-feature +avx -target-feature +avx512f  -target-feature +avx512fp16 -S -o - | FileCheck %s
 
 
 #include 
@@ -38,10 +38,7 @@
 __m512h test_mm512_castph128_ph512(__m128h A) {
   // CHECK-LABEL: test_mm512_castph128_ph512
   // CHECK: # %bb.0:
-  // CHECK-N

[PATCH] D144903: [X86] Drop single use check for freeze(undef) in LowerAVXCONCAT_VECTORS

2023-02-28 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 501079.
ManuelJBrito added a comment.

Recover mistakenly drop flag in avx-cast-builtins.
I was experimenting and forgot i had removed it.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D144903/new/

https://reviews.llvm.org/D144903

Files:
  clang/test/CodeGen/X86/avx-cast-builtins.c
  llvm/lib/Target/X86/X86ISelLowering.cpp
  llvm/test/CodeGen/X86/avx512-intrinsics.ll
  llvm/test/CodeGen/X86/avx512fp16-intrinsics.ll

Index: llvm/test/CodeGen/X86/avx512fp16-intrinsics.ll
===
--- llvm/test/CodeGen/X86/avx512fp16-intrinsics.ll
+++ llvm/test/CodeGen/X86/avx512fp16-intrinsics.ll
@@ -1231,10 +1231,7 @@
 define <32 x half> @test_mm512_castph128_ph512_freeze(<8 x half> %a0) nounwind {
 ; CHECK-LABEL: test_mm512_castph128_ph512_freeze:
 ; CHECK:   # %bb.0:
-; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm1
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; CHECK-NEXT:vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
+; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $zmm0
 ; CHECK-NEXT:retq
   %a1 = freeze <8 x half> poison
   %res = shufflevector <8 x half> %a0, <8 x half> %a1, <32 x i32> 
Index: llvm/test/CodeGen/X86/avx512-intrinsics.ll
===
--- llvm/test/CodeGen/X86/avx512-intrinsics.ll
+++ llvm/test/CodeGen/X86/avx512-intrinsics.ll
@@ -7495,10 +7495,7 @@
 define <8 x double> @test_mm256_castpd128_pd256_freeze(<2 x double> %a0) nounwind {
 ; CHECK-LABEL: test_mm256_castpd128_pd256_freeze:
 ; CHECK:   # %bb.0:
-; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm1
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; CHECK-NEXT:vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
+; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $zmm0
 ; CHECK-NEXT:ret{{[l|q]}}
   %a1 = freeze <2 x double> poison
   %res = shufflevector <2 x double> %a0, <2 x double> %a1, <8 x i32> 
@@ -7520,10 +7517,7 @@
 define <16 x float> @test_mm256_castps128_ps512_freeze(<4 x float> %a0) nounwind {
 ; CHECK-LABEL: test_mm256_castps128_ps512_freeze:
 ; CHECK:   # %bb.0:
-; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm1
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; CHECK-NEXT:vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
+; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $zmm0
 ; CHECK-NEXT:ret{{[l|q]}}
   %a1 = freeze <4 x float> poison
   %res = shufflevector <4 x float> %a0, <4 x float> %a1, <16x i32> 
@@ -7545,10 +7539,7 @@
 define <8 x i64> @test_mm512_castsi128_si512_freeze(<2 x i64> %a0) nounwind {
 ; CHECK-LABEL: test_mm512_castsi128_si512_freeze:
 ; CHECK:   # %bb.0:
-; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm1
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; CHECK-NEXT:vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
+; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $zmm0
 ; CHECK-NEXT:ret{{[l|q]}}
   %a1 = freeze <2 x i64> poison
   %res = shufflevector <2 x i64> %a0, <2 x i64> %a1, <8 x i32> 
Index: llvm/lib/Target/X86/X86ISelLowering.cpp
===
--- llvm/lib/Target/X86/X86ISelLowering.cpp
+++ llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -11656,7 +11656,7 @@
 SDValue SubVec = Op.getOperand(i);
 if (SubVec.isUndef())
   continue;
-if (ISD::isFreezeUndef(SubVec.getNode()) && SubVec.hasOneUse())
+if (ISD::isFreezeUndef(SubVec.getNode()))
   ++NumFreezeUndef;
 else if (ISD::isBuildVectorAllZeros(SubVec.getNode()))
   ++NumZero;
Index: clang/test/CodeGen/X86/avx-cast-builtins.c
===
--- clang/test/CodeGen/X86/avx-cast-builtins.c
+++ clang/test/CodeGen/X86/avx-cast-builtins.c
@@ -38,10 +38,7 @@
 __m512h test_mm512_castph128_ph512(__m128h A) {
   // CHECK-LABEL: test_mm512_castph128_ph512
   // CHECK: # %bb.0:
-  // CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
-  // CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm1
-  // CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm0
-  // CHECK-NEXT:vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
+  // CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $zmm0
   // CHECK-NEXT:ret{{[l|q]}}
   return _mm512_castph128_ph512(A);
 }
@@ -73,10 +70,7 @@
 __m512d test_mm512_castpd128_pd512(__m128d A){
   // CHECK-LABEL: test_mm512_castpd128_pd512
   // CHECK: # %bb.0:
-  // CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
-  // CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm1
-  // CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm0
-  // CHECK-NEXT:vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
+  // CHECK-NEXT:# kill: def $xmm0 k

[PATCH] D144903: [X86] Drop single use check for freeze(undef) in LowerAVXCONCAT_VECTORS

2023-03-02 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

I originally reverted it because of the wrong tag, but there were also some 
buildbot failures (see 
https://lab.llvm.org/buildbot/#/builders/139/builds/36736).
It appears to be failing an assert in DiagnosticsEngine::DiagStateMap::append, 
but i'm not very familiar with this part so i will have to investigate further.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D144903/new/

https://reviews.llvm.org/D144903

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D144903: [X86] Drop single use check for freeze(undef) in LowerAVXCONCAT_VECTORS

2023-03-05 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

It seems the build failure was caused by a known crash  
https://github.com/llvm/llvm-project/issues/55263. I tried to find some 
workaround but unsuccessfully.

So I'm thinking I can drop the end-to-end tests for now and commit them when 
the crash is fixed and for now just rely on the separate frontend and backend 
tests for regressions.

Is this OK with you @RKSimon ?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D144903/new/

https://reviews.llvm.org/D144903

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D143287: [Clang][X86] Change X86 cast intrinsics to use __builtin_nondeterministic_value

2023-03-08 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

In D143287#4179020 , @aqjune wrote:

> H, is D104790  superseded by this patch?

I don't think so we still need to fix the undefined intrinsics, right? Maybe 
I'm not understanding the question.

> I wonder what is the status of this patch as well.

We need to land D144903  first to have the 
correct assembly, but it's currently held up by a crash in one of the tests due 
to an unrelated issue.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D143287/new/

https://reviews.llvm.org/D143287

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D144903: [X86] Drop single use check for freeze(undef) in LowerAVXCONCAT_VECTORS

2023-03-09 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added inline comments.



Comment at: clang/test/CodeGen/X86/avx-cast-builtins.c:1
 // RUN: %clang_cc1 %s -O3 -flax-vector-conversions=none -ffreestanding %s 
-triple=x86_64-unknown-unknown -target-feature +avx -target-feature +avx512f  
-target-feature +avx512fp16 -S -o - | FileCheck %s
 

aqjune wrote:
> aqjune wrote:
> > This line contains `%s` twice, which seems to cause the crash.
> > 
> > Would removing one of `%s` resolve the crash issue? On my machine removing 
> > one of those worked well; with two `%s`; it crashed.
> Actually, with FileCheck's argument counted there are three `%s`, but it 
> isn't important; I meant two `%s` in clang_cc1's arguments.
Oops .. I should have noticed that...
I will try to commit again with this fix . Thanks for looking at this aqjune!!


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D144903/new/

https://reviews.llvm.org/D144903

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D144903: [X86] Drop single use check for freeze(undef) in LowerAVXCONCAT_VECTORS

2023-03-09 Thread Manuel Brito via Phabricator via cfe-commits 
This revision was landed with ongoing or failed builds.
This revision was automatically updated to reflect the committed changes.
Closed by commit rGe2817933fdd4: [X86] Drop single use check for freeze(undef) 
in LowerAVXCONCAT_VECTORS (authored by ManuelJBrito).

Changed prior to commit:
  https://reviews.llvm.org/D144903?vs=501079&id=503711#toc

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D144903/new/

https://reviews.llvm.org/D144903

Files:
  clang/test/CodeGen/X86/avx-cast-builtins.c
  llvm/lib/Target/X86/X86ISelLowering.cpp
  llvm/test/CodeGen/X86/avx512-intrinsics.ll
  llvm/test/CodeGen/X86/avx512fp16-intrinsics.ll

Index: llvm/test/CodeGen/X86/avx512fp16-intrinsics.ll
===
--- llvm/test/CodeGen/X86/avx512fp16-intrinsics.ll
+++ llvm/test/CodeGen/X86/avx512fp16-intrinsics.ll
@@ -1231,10 +1231,7 @@
 define <32 x half> @test_mm512_castph128_ph512_freeze(<8 x half> %a0) nounwind {
 ; CHECK-LABEL: test_mm512_castph128_ph512_freeze:
 ; CHECK:   # %bb.0:
-; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm1
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; CHECK-NEXT:vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
+; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $zmm0
 ; CHECK-NEXT:retq
   %a1 = freeze <8 x half> poison
   %res = shufflevector <8 x half> %a0, <8 x half> %a1, <32 x i32> 
Index: llvm/test/CodeGen/X86/avx512-intrinsics.ll
===
--- llvm/test/CodeGen/X86/avx512-intrinsics.ll
+++ llvm/test/CodeGen/X86/avx512-intrinsics.ll
@@ -7495,10 +7495,7 @@
 define <8 x double> @test_mm256_castpd128_pd256_freeze(<2 x double> %a0) nounwind {
 ; CHECK-LABEL: test_mm256_castpd128_pd256_freeze:
 ; CHECK:   # %bb.0:
-; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm1
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; CHECK-NEXT:vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
+; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $zmm0
 ; CHECK-NEXT:ret{{[l|q]}}
   %a1 = freeze <2 x double> poison
   %res = shufflevector <2 x double> %a0, <2 x double> %a1, <8 x i32> 
@@ -7520,10 +7517,7 @@
 define <16 x float> @test_mm256_castps128_ps512_freeze(<4 x float> %a0) nounwind {
 ; CHECK-LABEL: test_mm256_castps128_ps512_freeze:
 ; CHECK:   # %bb.0:
-; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm1
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; CHECK-NEXT:vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
+; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $zmm0
 ; CHECK-NEXT:ret{{[l|q]}}
   %a1 = freeze <4 x float> poison
   %res = shufflevector <4 x float> %a0, <4 x float> %a1, <16x i32> 
@@ -7545,10 +7539,7 @@
 define <8 x i64> @test_mm512_castsi128_si512_freeze(<2 x i64> %a0) nounwind {
 ; CHECK-LABEL: test_mm512_castsi128_si512_freeze:
 ; CHECK:   # %bb.0:
-; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm1
-; CHECK-NEXT:vinsertf128 $1, %xmm0, %ymm0, %ymm0
-; CHECK-NEXT:vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
+; CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $zmm0
 ; CHECK-NEXT:ret{{[l|q]}}
   %a1 = freeze <2 x i64> poison
   %res = shufflevector <2 x i64> %a0, <2 x i64> %a1, <8 x i32> 
Index: llvm/lib/Target/X86/X86ISelLowering.cpp
===
--- llvm/lib/Target/X86/X86ISelLowering.cpp
+++ llvm/lib/Target/X86/X86ISelLowering.cpp
@@ -11656,7 +11656,7 @@
 SDValue SubVec = Op.getOperand(i);
 if (SubVec.isUndef())
   continue;
-if (ISD::isFreezeUndef(SubVec.getNode()) && SubVec.hasOneUse())
+if (ISD::isFreezeUndef(SubVec.getNode()))
   ++NumFreezeUndef;
 else if (ISD::isBuildVectorAllZeros(SubVec.getNode()))
   ++NumZero;
Index: clang/test/CodeGen/X86/avx-cast-builtins.c
===
--- /dev/null
+++ clang/test/CodeGen/X86/avx-cast-builtins.c
@@ -0,0 +1,100 @@
+// RUN: %clang_cc1 -O3 -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-unknown-unknown -target-feature +avx -target-feature +avx512f  -target-feature +avx512fp16 -S -o - | FileCheck %s
+
+
+#include 
+
+__m256d test_mm256_castpd128_pd256(__m128d A) {
+  // CHECK-LABEL: test_mm256_castpd128_pd256
+  // CHECK: # %bb.0:
+  // CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
+  // CHECK-NEXT:ret{{[l|q]}}
+  return _mm256_castpd128_pd256(A);
+}
+
+__m256 test_mm256_castps128_ps256(__m128 A) {
+  // CHECK-LABEL: test_mm256_castps128_ps256
+  // CHECK: # %bb.0:
+  // CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
+  // CHECK-NEXT:ret{{[l|q]}}
+  return _mm256_castps128_ps256(A);
+}
+
+__m256i test_mm256_castsi128_si256(__

[PATCH] D143287: [Clang][X86] Change X86 cast intrinsics to use __builtin_nondeterministic_value

2023-03-09 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 503819.
ManuelJBrito added a comment.

Rebase

`avx-cast-builtins.c` was moved to D144903 


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D143287/new/

https://reviews.llvm.org/D143287

Files:
  clang/lib/Headers/avx512fintrin.h
  clang/lib/Headers/avx512fp16intrin.h
  clang/lib/Headers/avxintrin.h
  clang/test/CodeGen/X86/avx-builtins.c
  clang/test/CodeGen/X86/avx512f-builtins.c
  clang/test/CodeGen/X86/avx512fp16-builtins.c

Index: clang/test/CodeGen/X86/avx512fp16-builtins.c
===
--- clang/test/CodeGen/X86/avx512fp16-builtins.c
+++ clang/test/CodeGen/X86/avx512fp16-builtins.c
@@ -325,19 +325,22 @@
 
 __m256h test_mm256_castph128_ph256(__m128h __a) {
   // CHECK-LABEL: test_mm256_castph128_ph256
-  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <16 x i32> 
+  // CHECK: [[A:%.*]] = freeze <8 x half> poison 
+  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> [[A]], <16 x i32> 
   return _mm256_castph128_ph256(__a);
 }
 
 __m512h test_mm512_castph128_ph512(__m128h __a) {
   // CHECK-LABEL: test_mm512_castph128_ph512
-  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <32 x i32> 
+  // CHECK: [[A:%.*]] = freeze <8 x half> poison 
+  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> [[A]], <32 x i32> 
   return _mm512_castph128_ph512(__a);
 }
 
 __m512h test_mm512_castph256_ph512(__m256h __a) {
   // CHECK-LABEL: test_mm512_castph256_ph512
-  // CHECK: shufflevector <16 x half> %{{.*}}, <16 x half> %{{.*}}, <32 x i32> 
+  // CHECK: [[A:%.*]] = freeze <16 x half> poison 
+  // CHECK: shufflevector <16 x half> %{{.*}}, <16 x half> [[A]], <32 x i32> 
   return _mm512_castph256_ph512(__a);
 }
 
Index: clang/test/CodeGen/X86/avx512f-builtins.c
===
--- clang/test/CodeGen/X86/avx512f-builtins.c
+++ clang/test/CodeGen/X86/avx512f-builtins.c
@@ -8987,13 +8987,15 @@
 
 __m512 test_mm512_castps128_ps512(__m128 __A) {
   // CHECK-LABEL: @test_mm512_castps128_ps512
-  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <16 x i32> 
+  // CHECK: [[A:%.*]] = freeze <4 x float> poison 
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> [[A]], <16 x i32> 
   return _mm512_castps128_ps512(__A); 
 }
 
 __m512d test_mm512_castpd128_pd512(__m128d __A) {
   // CHECK-LABEL: @test_mm512_castpd128_pd512
-  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <8 x i32> 
+  // CHECK: [[A:%.*]] = freeze <2 x double> poison 
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> [[A]], <8 x i32> 
   return _mm512_castpd128_pd512(__A); 
 }
 
@@ -9086,7 +9088,8 @@
 __m512d test_mm512_castpd256_pd512(__m256d a)
 {
   // CHECK-LABEL: @test_mm512_castpd256_pd512
-  // CHECK: shufflevector <4 x double> {{.*}} 
+  // CHECK: [[A:%.*]] = freeze <4 x double> poison 
+  // CHECK: shufflevector <4 x double> {{.*}}, <4 x double> [[A]], 
   return _mm512_castpd256_pd512(a);
 }
 
@@ -9112,13 +9115,15 @@
 }
 __m512i test_mm512_castsi128_si512(__m128i __A) {
   // CHECK-LABEL: @test_mm512_castsi128_si512
-  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <8 x i32> 
+  // CHECK: [[A:%.*]] = freeze <2 x i64> poison 
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> [[A]], <8 x i32> 
   return _mm512_castsi128_si512(__A); 
 }
 
 __m512i test_mm512_castsi256_si512(__m256i __A) {
   // CHECK-LABEL: @test_mm512_castsi256_si512
-  // CHECK: shufflevector <4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <8 x i32> 
+  // CHECK: [[A:%.*]] = freeze <4 x i64> poison 
+  // CHECK: shufflevector <4 x i64> %{{.*}}, <4 x i64> [[A]], <8 x i32> 
   return _mm512_castsi256_si512(__A); 
 }
 
Index: clang/test/CodeGen/X86/avx-builtins.c
===
--- clang/test/CodeGen/X86/avx-builtins.c
+++ clang/test/CodeGen/X86/avx-builtins.c
@@ -143,7 +143,8 @@
 
 __m256d test_mm256_castpd128_pd256(__m128d A) {
   // CHECK-LABEL: test_mm256_castpd128_pd256
-  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <4 x i32> 
+  // CHECK: [[A:%.*]] = freeze <2 x double> poison
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> [[A]], <4 x i32> 
   return _mm256_castpd128_pd256(A);
 }
 
@@ -165,7 +166,8 @@
 
 __m256 test_mm256_castps128_ps256(__m128 A) {
   // CHECK-LABEL: test_mm256_castps128_ps256
-  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <8 x i32> 
+  // CHECK: [[A:%.*]] = freeze <4 x float> poison
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> [[A]], <8 x i32> 
   return _mm256_castps128_ps256(A);
 }
 
@@ -177,7 +179,8 @@
 
 __m256i test_mm256_castsi128_si256(__m128i A) {
   // CHECK-LABEL: test_mm256_castsi128_si256
-  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <4 x i32> 
+  // CHECK: [[A:%.*]] = freeze <

[PATCH] D144903: [X86] Drop single use check for freeze(undef) in LowerAVXCONCAT_VECTORS

2023-03-10 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

Reverted. I'll reassess what can be done here.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D144903/new/

https://reviews.llvm.org/D144903

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D152547: [clang][NFC] Drop alignment in builtin-nondeterministic-value test

2023-06-20 Thread Manuel Brito via Phabricator via cfe-commits 
This revision was automatically updated to reflect the committed changes.
Closed by commit rG36d50348a8b1: [clang][NFC] Drop alignment in 
builtin-nondeterministic-value test (authored by ManuelJBrito).

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D152547/new/

https://reviews.llvm.org/D152547

Files:
  clang/test/CodeGen/builtins-nondeterministic-value.c


Index: clang/test/CodeGen/builtins-nondeterministic-value.c
===
--- clang/test/CodeGen/builtins-nondeterministic-value.c
+++ clang/test/CodeGen/builtins-nondeterministic-value.c
@@ -5,8 +5,8 @@
 
 int clang_nondet_i( int x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca i32, align 4
-// CHECK: store i32 [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[A:%.*]] = alloca i32
+// CHECK: store i32 [[X:%.*]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze i32 poison
 // CHECK: ret i32 [[R]]
   return __builtin_nondeterministic_value(x);
@@ -14,8 +14,8 @@
 
 float clang_nondet_f( float x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca float, align 4
-// CHECK: store float [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[A:%.*]] = alloca float
+// CHECK: store float [[X:%.*]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze float poison
 // CHECK: ret float [[R]]
   return __builtin_nondeterministic_value(x);
@@ -23,8 +23,8 @@
 
 double clang_nondet_d( double x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca double, align 8
-// CHECK: store double [[X:%.*]], ptr [[A]], align 8
+// CHECK: [[A:%.*]] = alloca double
+// CHECK: store double [[X:%.*]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze double poison
 // CHECK: ret double [[R]]
   return __builtin_nondeterministic_value(x);
@@ -32,9 +32,9 @@
 
 _Bool clang_nondet_b( _Bool x) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[A:%.*]] = alloca i8
 // CHECK: [[B:%.*]] = zext i1 %x to i8
-// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: store i8 [[B]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze i1 poison
 // CHECK: ret i1 [[R]]
   return __builtin_nondeterministic_value(x);
@@ -42,19 +42,19 @@
 
 void clang_nondet_fv( ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca <4 x float>, align
+// CHECK: [[A:%.*]] = alloca <4 x float>
 // CHECK: [[R:%.*]] = freeze <4 x float> poison
-// CHECK: store <4 x float> [[R]], ptr [[A]], align
+// CHECK: store <4 x float> [[R]], ptr [[A]]
 // CHECK: ret void
   float4 x = __builtin_nondeterministic_value(x);
 }
 
 void clang_nondet_bv( ) {
-// CHECK: [[A:%.*]] = alloca i8, align
+// CHECK: [[A:%.*]] = alloca i8
 // CHECK: [[V:%.*]] = freeze <4 x i1> poison
 // CHECK: [[SV:%.*]] = shufflevector <4 x i1> [[V]], <4 x i1> poison, <8 x 
i32> 
 // CHECK: [[BC:%.*]] = bitcast <8 x i1> [[SV]] to i8
-// CHECK: store i8 [[BC]], ptr [[A]], align
+// CHECK: store i8 [[BC]], ptr [[A]]
 // CHECK: ret void
   bool4 x = __builtin_nondeterministic_value(x);
 }


Index: clang/test/CodeGen/builtins-nondeterministic-value.c
===
--- clang/test/CodeGen/builtins-nondeterministic-value.c
+++ clang/test/CodeGen/builtins-nondeterministic-value.c
@@ -5,8 +5,8 @@
 
 int clang_nondet_i( int x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca i32, align 4
-// CHECK: store i32 [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[A:%.*]] = alloca i32
+// CHECK: store i32 [[X:%.*]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze i32 poison
 // CHECK: ret i32 [[R]]
   return __builtin_nondeterministic_value(x);
@@ -14,8 +14,8 @@
 
 float clang_nondet_f( float x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca float, align 4
-// CHECK: store float [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[A:%.*]] = alloca float
+// CHECK: store float [[X:%.*]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze float poison
 // CHECK: ret float [[R]]
   return __builtin_nondeterministic_value(x);
@@ -23,8 +23,8 @@
 
 double clang_nondet_d( double x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca double, align 8
-// CHECK: store double [[X:%.*]], ptr [[A]], align 8
+// CHECK: [[A:%.*]] = alloca double
+// CHECK: store double [[X:%.*]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze double poison
 // CHECK: ret double [[R]]
   return __builtin_nondeterministic_value(x);
@@ -32,9 +32,9 @@
 
 _Bool clang_nondet_b( _Bool x) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[A:%.*]] = alloca i8
 // CHECK: [[B:%.*]] = zext i1 %x to i8
-// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: store i8 [[B]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze i1 poison
 // CHECK: ret i1 [[R]]
   return __builtin_nondeterministic_value(x);
@@ -42,19 +42,19 @@
 
 void clang_nondet_fv( ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca <4 x float>, align
+// CHECK: [[A:%.*]] = alloca <4 x float>
 // CHECK: [[R:%.*]] = freeze <4 x float> poison
-// CHECK: store <4 x float> [[R]], ptr [[A]], align
+// CHECK: store <4 x fl

[PATCH] D152658: [InstCombine] Change SimplifyDemandedVectorElts to use PoisonElts instead of UndefElts

2023-06-26 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

Ping


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D152658/new/

https://reviews.llvm.org/D152658

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D143287: [Clang][X86] Change X86 cast intrinsics to use __builtin_nondeterministic_value

2023-04-07 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 511663.
ManuelJBrito added a comment.

Update to remove multiple uses of freeze poison.
I am unsure about the code style used in the 128 to 512 casts. Any comments are 
appreciated.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D143287/new/

https://reviews.llvm.org/D143287

Files:
  clang/lib/Headers/avx512fintrin.h
  clang/lib/Headers/avx512fp16intrin.h
  clang/lib/Headers/avxintrin.h
  clang/test/CodeGen/X86/avx-builtins.c
  clang/test/CodeGen/X86/avx-cast-builtins.c
  clang/test/CodeGen/X86/avx512f-builtins.c
  clang/test/CodeGen/X86/avx512fp16-builtins.c

Index: clang/test/CodeGen/X86/avx512fp16-builtins.c
===
--- clang/test/CodeGen/X86/avx512fp16-builtins.c
+++ clang/test/CodeGen/X86/avx512fp16-builtins.c
@@ -325,19 +325,26 @@
 
 __m256h test_mm256_castph128_ph256(__m128h __a) {
   // CHECK-LABEL: test_mm256_castph128_ph256
-  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <16 x i32> 
+  // CHECK: [[A:%.*]] = freeze <8 x half> poison 
+  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> [[A]], <16 x i32> 
   return _mm256_castph128_ph256(__a);
 }
 
 __m512h test_mm512_castph128_ph512(__m128h __a) {
   // CHECK-LABEL: test_mm512_castph128_ph512
-  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <32 x i32> 
+  // CHECK: [[B:%.*]] = freeze <16 x half> poison
+  // CHECK: store <16 x half> [[B]], ptr [[BA:%.*]]
+  // CHECK: [[A:%.*]] = freeze <8 x half> poison
+  // CHECK: [[SV:%.*]] = shufflevector <8 x half> %{{.*}}, <8 x half> [[A]], <16 x i32> 
+  // CHECK: [[C:%.*]] = load <16 x half>, ptr [[BA]]
+  // CHECK: shufflevector <16 x half> [[SV]], <16 x half> [[C]], <32 x i32> 
   return _mm512_castph128_ph512(__a);
 }
 
 __m512h test_mm512_castph256_ph512(__m256h __a) {
   // CHECK-LABEL: test_mm512_castph256_ph512
-  // CHECK: shufflevector <16 x half> %{{.*}}, <16 x half> %{{.*}}, <32 x i32> 
+  // CHECK: [[A:%.*]] = freeze <16 x half> poison 
+  // CHECK: shufflevector <16 x half> %{{.*}}, <16 x half> [[A]], <32 x i32> 
   return _mm512_castph256_ph512(__a);
 }
 
Index: clang/test/CodeGen/X86/avx512f-builtins.c
===
--- clang/test/CodeGen/X86/avx512f-builtins.c
+++ clang/test/CodeGen/X86/avx512f-builtins.c
@@ -8987,13 +8987,23 @@
 
 __m512 test_mm512_castps128_ps512(__m128 __A) {
   // CHECK-LABEL: @test_mm512_castps128_ps512
-  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <16 x i32> 
+  // CHECK: [[B:%.*]] = freeze <8 x float> poison
+  // CHECK: store <8 x float> [[B]], ptr [[BA:%.*]]
+  // CHECK: [[A:%.*]] = freeze <4 x float> poison 
+  // CHECK: [[SV:%.*]] = shufflevector <4 x float> %{{.*}}, <4 x float> [[A]], <8 x i32> 
+  // CHECK: [[C:%.*]] = load <8 x float>, ptr [[BA]]
+  // CHECK: shufflevector <8 x float> [[SV]], <8 x float> [[C]], <16 x i32> 
   return _mm512_castps128_ps512(__A); 
 }
 
 __m512d test_mm512_castpd128_pd512(__m128d __A) {
   // CHECK-LABEL: @test_mm512_castpd128_pd512
-  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <8 x i32> 
+  // CHECK: [[B:%.*]] = freeze <4 x double> poison
+  // CHECK: store <4 x double> [[B]], ptr [[BA:%.*]]
+  // CHECK: [[A:%.*]] = freeze <2 x double> poison 
+  // CHECK: [[SV:%.*]] = shufflevector <2 x double> %{{.*}}, <2 x double> [[A]], <4 x i32> 
+  // CHECK: [[C:%.*]] = load <4 x double>, ptr [[BA]]
+  // CHECK: shufflevector <4 x double> [[SV]], <4 x double> [[C]], <8 x i32> 
   return _mm512_castpd128_pd512(__A); 
 }
 
@@ -9086,7 +9096,8 @@
 __m512d test_mm512_castpd256_pd512(__m256d a)
 {
   // CHECK-LABEL: @test_mm512_castpd256_pd512
-  // CHECK: shufflevector <4 x double> {{.*}} 
+  // CHECK: [[A:%.*]] = freeze <4 x double> poison 
+  // CHECK: shufflevector <4 x double> %{{.}}, <4 x double> [[A]], <8 x i32> 
   return _mm512_castpd256_pd512(a);
 }
 
@@ -9112,13 +9123,19 @@
 }
 __m512i test_mm512_castsi128_si512(__m128i __A) {
   // CHECK-LABEL: @test_mm512_castsi128_si512
-  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <8 x i32> 
+  // CHECK: [[B:%.*]] = freeze <4 x i64> poison
+  // CHECK: store <4 x i64> [[B]], ptr [[BA:%.*]]
+  // CHECK: [[A:%.*]] = freeze <2 x i64> poison 
+  // CHECK: [[SV:%.*]] = shufflevector <2 x i64> %{{.*}}, <2 x i64> [[A]], <4 x i32> 
+  // CHECK: [[C:%.*]] = load <4 x i64>, ptr [[BA]]
+  // CHECK: shufflevector <4 x i64> [[SV]], <4 x i64> [[C]], <8 x i32> 
   return _mm512_castsi128_si512(__A); 
 }
 
 __m512i test_mm512_castsi256_si512(__m256i __A) {
   // CHECK-LABEL: @test_mm512_castsi256_si512
-  // CHECK: shufflevector <4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <8 x i32> 
+  // CHECK: [[A:%.*]] = freeze <4 x i64> poison 
+  // CHECK: shufflevector <4 x i64> %{{.*}}, <4 x i64> [[A]], <8 x i32> 
   return _mm512_castsi256_si512(__A); 
 }
 
Index: clang/test/CodeGen/X86/avx-cast-builtins.c
=

[PATCH] D143287: [Clang][X86] Change X86 cast intrinsics to use __builtin_nondeterministic_value

2023-04-14 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

Any further comments @RKSimon ?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D143287/new/

https://reviews.llvm.org/D143287

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D143287: [Clang][X86] Change X86 cast intrinsics to use __builtin_nondeterministic_value

2023-04-17 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

Thanks for the review! I'll simplify the masks.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D143287/new/

https://reviews.llvm.org/D143287

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D143287: [Clang][X86] Change X86 cast intrinsics to use __builtin_nondeterministic_value

2023-04-17 Thread Manuel Brito via Phabricator via cfe-commits 
This revision was landed with ongoing or failed builds.
This revision was automatically updated to reflect the committed changes.
Closed by commit rG5184dc2d7cce: [Clang][X86] Change X86 cast intrinsics to use 
__builtin_nondeterministic_value (authored by ManuelJBrito).

Changed prior to commit:
  https://reviews.llvm.org/D143287?vs=511663&id=514165#toc

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D143287/new/

https://reviews.llvm.org/D143287

Files:
  clang/lib/Headers/avx512fintrin.h
  clang/lib/Headers/avx512fp16intrin.h
  clang/lib/Headers/avxintrin.h
  clang/test/CodeGen/X86/avx-builtins.c
  clang/test/CodeGen/X86/avx-cast-builtins.c
  clang/test/CodeGen/X86/avx512f-builtins.c
  clang/test/CodeGen/X86/avx512fp16-builtins.c

Index: clang/test/CodeGen/X86/avx512fp16-builtins.c
===
--- clang/test/CodeGen/X86/avx512fp16-builtins.c
+++ clang/test/CodeGen/X86/avx512fp16-builtins.c
@@ -325,19 +325,26 @@
 
 __m256h test_mm256_castph128_ph256(__m128h __a) {
   // CHECK-LABEL: test_mm256_castph128_ph256
-  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <16 x i32> 
+  // CHECK: [[A:%.*]] = freeze <8 x half> poison 
+  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> [[A]], <16 x i32> 
   return _mm256_castph128_ph256(__a);
 }
 
 __m512h test_mm512_castph128_ph512(__m128h __a) {
   // CHECK-LABEL: test_mm512_castph128_ph512
-  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <32 x i32> 
+  // CHECK: [[B:%.*]] = freeze <16 x half> poison
+  // CHECK: store <16 x half> [[B]], ptr [[BA:%.*]]
+  // CHECK: [[A:%.*]] = freeze <8 x half> poison
+  // CHECK: [[SV:%.*]] = shufflevector <8 x half> %{{.*}}, <8 x half> [[A]], <16 x i32> 
+  // CHECK: [[C:%.*]] = load <16 x half>, ptr [[BA]]
+  // CHECK: shufflevector <16 x half> [[SV]], <16 x half> [[C]], <32 x i32> 
   return _mm512_castph128_ph512(__a);
 }
 
 __m512h test_mm512_castph256_ph512(__m256h __a) {
   // CHECK-LABEL: test_mm512_castph256_ph512
-  // CHECK: shufflevector <16 x half> %{{.*}}, <16 x half> %{{.*}}, <32 x i32> 
+  // CHECK: [[A:%.*]] = freeze <16 x half> poison 
+  // CHECK: shufflevector <16 x half> %{{.*}}, <16 x half> [[A]], <32 x i32> 
   return _mm512_castph256_ph512(__a);
 }
 
Index: clang/test/CodeGen/X86/avx512f-builtins.c
===
--- clang/test/CodeGen/X86/avx512f-builtins.c
+++ clang/test/CodeGen/X86/avx512f-builtins.c
@@ -8987,13 +8987,23 @@
 
 __m512 test_mm512_castps128_ps512(__m128 __A) {
   // CHECK-LABEL: @test_mm512_castps128_ps512
-  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <16 x i32> 
+  // CHECK: [[B:%.*]] = freeze <8 x float> poison
+  // CHECK: store <8 x float> [[B]], ptr [[BA:%.*]]
+  // CHECK: [[A:%.*]] = freeze <4 x float> poison 
+  // CHECK: [[SV:%.*]] = shufflevector <4 x float> %{{.*}}, <4 x float> [[A]], <8 x i32> 
+  // CHECK: [[C:%.*]] = load <8 x float>, ptr [[BA]]
+  // CHECK: shufflevector <8 x float> [[SV]], <8 x float> [[C]], <16 x i32> 
   return _mm512_castps128_ps512(__A); 
 }
 
 __m512d test_mm512_castpd128_pd512(__m128d __A) {
   // CHECK-LABEL: @test_mm512_castpd128_pd512
-  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <8 x i32> 
+  // CHECK: [[B:%.*]] = freeze <4 x double> poison
+  // CHECK: store <4 x double> [[B]], ptr [[BA:%.*]]
+  // CHECK: [[A:%.*]] = freeze <2 x double> poison 
+  // CHECK: [[SV:%.*]] = shufflevector <2 x double> %{{.*}}, <2 x double> [[A]], <4 x i32> 
+  // CHECK: [[C:%.*]] = load <4 x double>, ptr [[BA]]
+  // CHECK: shufflevector <4 x double> [[SV]], <4 x double> [[C]], <8 x i32> 
   return _mm512_castpd128_pd512(__A); 
 }
 
@@ -9086,7 +9096,8 @@
 __m512d test_mm512_castpd256_pd512(__m256d a)
 {
   // CHECK-LABEL: @test_mm512_castpd256_pd512
-  // CHECK: shufflevector <4 x double> {{.*}} 
+  // CHECK: [[A:%.*]] = freeze <4 x double> poison 
+  // CHECK: shufflevector <4 x double> %{{.}}, <4 x double> [[A]], <8 x i32> 
   return _mm512_castpd256_pd512(a);
 }
 
@@ -9112,13 +9123,19 @@
 }
 __m512i test_mm512_castsi128_si512(__m128i __A) {
   // CHECK-LABEL: @test_mm512_castsi128_si512
-  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <8 x i32> 
+  // CHECK: [[B:%.*]] = freeze <4 x i64> poison
+  // CHECK: store <4 x i64> [[B]], ptr [[BA:%.*]]
+  // CHECK: [[A:%.*]] = freeze <2 x i64> poison 
+  // CHECK: [[SV:%.*]] = shufflevector <2 x i64> %{{.*}}, <2 x i64> [[A]], <4 x i32> 
+  // CHECK: [[C:%.*]] = load <4 x i64>, ptr [[BA]]
+  // CHECK: shufflevector <4 x i64> [[SV]], <4 x i64> [[C]], <8 x i32> 
   return _mm512_castsi128_si512(__A); 
 }
 
 __m512i test_mm512_castsi256_si512(__m256i __A) {
   // CHECK-LABEL: @test_mm512_castsi256_si512
-  // CHECK: shufflevector <4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <8 x i32> 
+  // CHECK: [[A:%.*]] = freeze <4 x i64> poison 
+  // CHECK: shufflevector <4 x i64> %{{.*}}, <4 x

[PATCH] D143287: [Clang][x86] Change x86 cast intrinsics to use __builtin_nondeterministic_value

2023-02-07 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

In D143287#4108165 , @craig.topper 
wrote:

> In D143287#4107439 , @ManuelJBrito 
> wrote:
>
>> In D143287#4103597 , @craig.topper 
>> wrote:
>>
>>> We have a couple bugs that show (freeze (poison)) doesn't work past 
>>> SelectionDAG. Is that a concern here? The most recent 
>>> https://github.com/llvm/llvm-project/issues/60429
>>
>> I don't think it's a concern . Here https://godbolt.org/z/1ecM8roYh the 
>> lowering seems ok.
>
> The bugs start occurring if you use the same nondeterministic value multiple 
> times and expect the value to be the same for all uses.

I understand now ... so if the bug is present it works as if it were an undef.  
For these intrinsics we would just regress to the current behavior.
So is it okay if we land this patch? Because it will still be an improvement.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D143287/new/

https://reviews.llvm.org/D143287

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D143287: [Clang][x86] Change x86 cast intrinsics to use __builtin_nondeterministic_value

2023-02-07 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 495611.
ManuelJBrito added a comment.

Add end to end tests

Currrently these expect a mov that will be removed when the instcombine bug is 
fixed


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D143287/new/

https://reviews.llvm.org/D143287

Files:
  clang/lib/Headers/avx512fintrin.h
  clang/lib/Headers/avx512fp16intrin.h
  clang/lib/Headers/avxintrin.h
  clang/test/CodeGen/X86/avx-builtins.c
  clang/test/CodeGen/X86/avx-cast-builtins.c
  clang/test/CodeGen/X86/avx512f-builtins.c
  clang/test/CodeGen/X86/avx512fp16-builtins.c

Index: clang/test/CodeGen/X86/avx512fp16-builtins.c
===
--- clang/test/CodeGen/X86/avx512fp16-builtins.c
+++ clang/test/CodeGen/X86/avx512fp16-builtins.c
@@ -326,19 +326,19 @@
 
 __m256h test_mm256_castph128_ph256(__m128h __a) {
   // CHECK-LABEL: test_mm256_castph128_ph256
-  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <16 x i32> 
+  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <16 x i32> 
   return _mm256_castph128_ph256(__a);
 }
 
 __m512h test_mm512_castph128_ph512(__m128h __a) {
   // CHECK-LABEL: test_mm512_castph128_ph512
-  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <32 x i32> 
+  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <32 x i32> 
   return _mm512_castph128_ph512(__a);
 }
 
 __m512h test_mm512_castph256_ph512(__m256h __a) {
   // CHECK-LABEL: test_mm512_castph256_ph512
-  // CHECK: shufflevector <16 x half> %{{.*}}, <16 x half> %{{.*}}, <32 x i32> 
+  // CHECK: shufflevector <16 x half> %{{.*}}, <16 x half> %{{.*}}, <32 x i32> 
   return _mm512_castph256_ph512(__a);
 }
 
Index: clang/test/CodeGen/X86/avx512f-builtins.c
===
--- clang/test/CodeGen/X86/avx512f-builtins.c
+++ clang/test/CodeGen/X86/avx512f-builtins.c
@@ -8987,13 +8987,13 @@
 
 __m512 test_mm512_castps128_ps512(__m128 __A) {
   // CHECK-LABEL: @test_mm512_castps128_ps512
-  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <16 x i32> 
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <16 x i32> 
   return _mm512_castps128_ps512(__A); 
 }
 
 __m512d test_mm512_castpd128_pd512(__m128d __A) {
   // CHECK-LABEL: @test_mm512_castpd128_pd512
-  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <8 x i32> 
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <8 x i32> 
   return _mm512_castpd128_pd512(__A); 
 }
 
@@ -9086,7 +9086,7 @@
 __m512d test_mm512_castpd256_pd512(__m256d a)
 {
   // CHECK-LABEL: @test_mm512_castpd256_pd512
-  // CHECK: shufflevector <4 x double> {{.*}} 
+  // CHECK: shufflevector <4 x double> {{.*}} 
   return _mm512_castpd256_pd512(a);
 }
 
@@ -9112,13 +9112,13 @@
 }
 __m512i test_mm512_castsi128_si512(__m128i __A) {
   // CHECK-LABEL: @test_mm512_castsi128_si512
-  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <8 x i32> 
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <8 x i32> 
   return _mm512_castsi128_si512(__A); 
 }
 
 __m512i test_mm512_castsi256_si512(__m256i __A) {
   // CHECK-LABEL: @test_mm512_castsi256_si512
-  // CHECK: shufflevector <4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <8 x i32> 
+  // CHECK: shufflevector <4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <8 x i32> 
   return _mm512_castsi256_si512(__A); 
 }
 
Index: clang/test/CodeGen/X86/avx-cast-builtins.c
===
--- /dev/null
+++ clang/test/CodeGen/X86/avx-cast-builtins.c
@@ -0,0 +1,100 @@
+// RUN: %clang_cc1 %s -O3 -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-unknown-unknown -target-feature +avx -target-feature +avx512f  -target-feature +avx512fp16 -S -o - | FileCheck %s
+
+
+#include 
+
+__m256d test_mm256_castpd128_pd256(__m128d A) {
+  // CHECK-LABEL: test_mm256_castpd128_pd256
+  // CHECK: # %bb.0:
+  // CHECK-NEXT:vmovaps %xmm0, %xmm0
+  // CHECK-NEXT:ret{{[l|q]}}
+  return _mm256_castpd128_pd256(A);
+}
+
+__m256 test_mm256_castps128_ps256(__m128 A) {
+  // CHECK-LABEL: test_mm256_castps128_ps256
+  // CHECK: # %bb.0:
+  // CHECK-NEXT:vmovaps %xmm0, %xmm0
+  // CHECK-NEXT:ret{{[l|q]}}
+  return _mm256_castps128_ps256(A);
+}
+
+__m256i test_mm256_castsi128_si256(__m128i A) {
+  // CHECK-LABEL: test_mm256_castsi128_si256
+  // CHECK: # %bb.0:
+  // CHECK-NEXT:vmovaps %xmm0, %xmm0
+  // CHECK-NEXT:ret{{[l|q]}}
+  return _mm256_castsi128_si256(A);
+}
+
+__m256h test_mm256_castph128_ph256(__m128h A) {
+  // CHECK-LABEL: test_mm256_castph128_ph256
+  // CHECK: # %bb.0:
+  // CHECK-NEXT:vmovaps %xmm0, %xmm0
+  // CHECK-NEXT:ret{{[l|q]}}
+  return _mm256_castph128_ph256(A);
+}
+
+__m512h test_mm512_castph128_ph512(__m128h A) {
+  // CHECK-LABEL: test_mm512_castph128_ph512
+  // CHECK: # %bb.0:
+

[PATCH] D143287: [Clang][X86] Change X86 cast intrinsics to use __builtin_nondeterministic_value

2023-02-09 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 496129.
ManuelJBrito retitled this revision from "[Clang][x86] Change x86 cast 
intrinsics to use __builtin_nondeterministic_value" to "[Clang][X86] Change X86 
cast intrinsics to use __builtin_nondeterministic_value".
ManuelJBrito added a comment.

Match freeze(poison) operand in tests


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D143287/new/

https://reviews.llvm.org/D143287

Files:
  clang/lib/Headers/avx512fintrin.h
  clang/lib/Headers/avx512fp16intrin.h
  clang/lib/Headers/avxintrin.h
  clang/test/CodeGen/X86/avx-builtins.c
  clang/test/CodeGen/X86/avx-cast-builtins.c
  clang/test/CodeGen/X86/avx512f-builtins.c
  clang/test/CodeGen/X86/avx512fp16-builtins.c

Index: clang/test/CodeGen/X86/avx512fp16-builtins.c
===
--- clang/test/CodeGen/X86/avx512fp16-builtins.c
+++ clang/test/CodeGen/X86/avx512fp16-builtins.c
@@ -326,19 +326,22 @@
 
 __m256h test_mm256_castph128_ph256(__m128h __a) {
   // CHECK-LABEL: test_mm256_castph128_ph256
-  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <16 x i32> 
+  // CHECK: [[A:%.*]] = freeze <8 x half> poison 
+  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> [[A]], <16 x i32> 
   return _mm256_castph128_ph256(__a);
 }
 
 __m512h test_mm512_castph128_ph512(__m128h __a) {
   // CHECK-LABEL: test_mm512_castph128_ph512
-  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <32 x i32> 
+  // CHECK: [[A:%.*]] = freeze <8 x half> poison 
+  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> [[A]], <32 x i32> 
   return _mm512_castph128_ph512(__a);
 }
 
 __m512h test_mm512_castph256_ph512(__m256h __a) {
   // CHECK-LABEL: test_mm512_castph256_ph512
-  // CHECK: shufflevector <16 x half> %{{.*}}, <16 x half> %{{.*}}, <32 x i32> 
+  // CHECK: [[A:%.*]] = freeze <16 x half> poison 
+  // CHECK: shufflevector <16 x half> %{{.*}}, <16 x half> [[A]], <32 x i32> 
   return _mm512_castph256_ph512(__a);
 }
 
Index: clang/test/CodeGen/X86/avx512f-builtins.c
===
--- clang/test/CodeGen/X86/avx512f-builtins.c
+++ clang/test/CodeGen/X86/avx512f-builtins.c
@@ -8987,13 +8987,15 @@
 
 __m512 test_mm512_castps128_ps512(__m128 __A) {
   // CHECK-LABEL: @test_mm512_castps128_ps512
-  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <16 x i32> 
+  // CHECK: [[A:%.*]] = freeze <4 x float> poison 
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> [[A]], <16 x i32> 
   return _mm512_castps128_ps512(__A); 
 }
 
 __m512d test_mm512_castpd128_pd512(__m128d __A) {
   // CHECK-LABEL: @test_mm512_castpd128_pd512
-  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <8 x i32> 
+  // CHECK: [[A:%.*]] = freeze <2 x double> poison 
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> [[A]], <8 x i32> 
   return _mm512_castpd128_pd512(__A); 
 }
 
@@ -9086,7 +9088,8 @@
 __m512d test_mm512_castpd256_pd512(__m256d a)
 {
   // CHECK-LABEL: @test_mm512_castpd256_pd512
-  // CHECK: shufflevector <4 x double> {{.*}} 
+  // CHECK: [[A:%.*]] = freeze <4 x double> poison 
+  // CHECK: shufflevector <4 x double> {{.*}}, <4 x double> [[A]], 
   return _mm512_castpd256_pd512(a);
 }
 
@@ -9112,13 +9115,15 @@
 }
 __m512i test_mm512_castsi128_si512(__m128i __A) {
   // CHECK-LABEL: @test_mm512_castsi128_si512
-  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <8 x i32> 
+  // CHECK: [[A:%.*]] = freeze <2 x i64> poison 
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> [[A]], <8 x i32> 
   return _mm512_castsi128_si512(__A); 
 }
 
 __m512i test_mm512_castsi256_si512(__m256i __A) {
   // CHECK-LABEL: @test_mm512_castsi256_si512
-  // CHECK: shufflevector <4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <8 x i32> 
+  // CHECK: [[A:%.*]] = freeze <4 x i64> poison 
+  // CHECK: shufflevector <4 x i64> %{{.*}}, <4 x i64> [[A]], <8 x i32> 
   return _mm512_castsi256_si512(__A); 
 }
 
Index: clang/test/CodeGen/X86/avx-cast-builtins.c
===
--- /dev/null
+++ clang/test/CodeGen/X86/avx-cast-builtins.c
@@ -0,0 +1,100 @@
+// RUN: %clang_cc1 %s -O3 -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-unknown-unknown -target-feature +avx -target-feature +avx512f  -target-feature +avx512fp16 -S -o - | FileCheck %s
+
+
+#include 
+
+__m256d test_mm256_castpd128_pd256(__m128d A) {
+  // CHECK-LABEL: test_mm256_castpd128_pd256
+  // CHECK: # %bb.0:
+  // CHECK-NEXT:vmovaps %xmm0, %xmm0
+  // CHECK-NEXT:ret{{[l|q]}}
+  return _mm256_castpd128_pd256(A);
+}
+
+__m256 test_mm256_castps128_ps256(__m128 A) {
+  // CHECK-LABEL: test_mm256_castps128_ps256
+  // CHECK: # %bb.0:
+  // CHECK-NEXT:vmovaps %xmm0, %xmm0
+  // CHECK-NEXT:ret{{[l|q]}}
+  return _mm256_castps128_ps256(A);
+}
+
+__m256i test_mm256_castsi128_si256(__m128i A) {
+

[PATCH] D143287: [Clang][X86] Change X86 cast intrinsics to use __builtin_nondeterministic_value

2023-02-24 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 500149.
ManuelJBrito added a comment.

Update tests after D143593 .

There are some performance regressions with casts from 128 to 512. The backend 
inserts vinsertf instructions. So that as to be fixed.
In D130339  @RKSimon  mentioned something 
about custom vector widening patterns that need to be adjusted to handle 
freeze(undef), any pointers for how a patch for that would look like?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D143287/new/

https://reviews.llvm.org/D143287

Files:
  clang/lib/Headers/avx512fintrin.h
  clang/lib/Headers/avx512fp16intrin.h
  clang/lib/Headers/avxintrin.h
  clang/test/CodeGen/X86/avx-builtins.c
  clang/test/CodeGen/X86/avx-cast-builtins.c
  clang/test/CodeGen/X86/avx512f-builtins.c
  clang/test/CodeGen/X86/avx512fp16-builtins.c

Index: clang/test/CodeGen/X86/avx512fp16-builtins.c
===
--- clang/test/CodeGen/X86/avx512fp16-builtins.c
+++ clang/test/CodeGen/X86/avx512fp16-builtins.c
@@ -325,19 +325,22 @@
 
 __m256h test_mm256_castph128_ph256(__m128h __a) {
   // CHECK-LABEL: test_mm256_castph128_ph256
-  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <16 x i32> 
+  // CHECK: [[A:%.*]] = freeze <8 x half> poison
+  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> [[A]], <16 x i32> 
   return _mm256_castph128_ph256(__a);
 }
 
 __m512h test_mm512_castph128_ph512(__m128h __a) {
   // CHECK-LABEL: test_mm512_castph128_ph512
-  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> %{{.*}}, <32 x i32> 
+  // CHECK: [[A:%.*]] = freeze <8 x half> poison
+  // CHECK: shufflevector <8 x half> %{{.*}}, <8 x half> [[A]], <32 x i32> 
   return _mm512_castph128_ph512(__a);
 }
 
 __m512h test_mm512_castph256_ph512(__m256h __a) {
   // CHECK-LABEL: test_mm512_castph256_ph512
-  // CHECK: shufflevector <16 x half> %{{.*}}, <16 x half> %{{.*}}, <32 x i32> 
+  // CHECK: [[A:%.*]] = freeze <16 x half> poison
+  // CHECK: shufflevector <16 x half> %{{.*}}, <16 x half> [[A]], <32 x i32> 
   return _mm512_castph256_ph512(__a);
 }
 
Index: clang/test/CodeGen/X86/avx512f-builtins.c
===
--- clang/test/CodeGen/X86/avx512f-builtins.c
+++ clang/test/CodeGen/X86/avx512f-builtins.c
@@ -8987,13 +8987,15 @@
 
 __m512 test_mm512_castps128_ps512(__m128 __A) {
   // CHECK-LABEL: @test_mm512_castps128_ps512
-  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <16 x i32> 
+  // CHECK: [[A:%.*]] = freeze <4 x float> poison
+  // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> [[A]], <16 x i32> 
   return _mm512_castps128_ps512(__A); 
 }
 
 __m512d test_mm512_castpd128_pd512(__m128d __A) {
   // CHECK-LABEL: @test_mm512_castpd128_pd512
-  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <8 x i32> 
+  // CHECK: [[A:%.*]] = freeze <2 x double> poison
+  // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> [[A]], <8 x i32> 
   return _mm512_castpd128_pd512(__A); 
 }
 
@@ -9086,7 +9088,8 @@
 __m512d test_mm512_castpd256_pd512(__m256d a)
 {
   // CHECK-LABEL: @test_mm512_castpd256_pd512
-  // CHECK: shufflevector <4 x double> {{.*}} 
+  // CHECK: [[A:%.*]] = freeze <4 x double> poison
+  // CHECK: shufflevector <4 x double> {{.*}}, <4 x double> [[A]], <8 x i32> 
   return _mm512_castpd256_pd512(a);
 }
 
@@ -9112,13 +9115,15 @@
 }
 __m512i test_mm512_castsi128_si512(__m128i __A) {
   // CHECK-LABEL: @test_mm512_castsi128_si512
-  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> %{{.*}}, <8 x i32> 
+  // CHECK: [[A:%.*]] = freeze <2 x i64> poison
+  // CHECK: shufflevector <2 x i64> %{{.*}}, <2 x i64> [[A]], <8 x i32> 
   return _mm512_castsi128_si512(__A); 
 }
 
 __m512i test_mm512_castsi256_si512(__m256i __A) {
   // CHECK-LABEL: @test_mm512_castsi256_si512
-  // CHECK: shufflevector <4 x i64> %{{.*}}, <4 x i64> %{{.*}}, <8 x i32> 
+  // CHECK: [[A:%.*]] = freeze <4 x i64> poison
+  // CHECK: shufflevector <4 x i64> %{{.*}}, <4 x i64> [[A]], <8 x i32> 
   return _mm512_castsi256_si512(__A); 
 }
 
Index: clang/test/CodeGen/X86/avx-cast-builtins.c
===
--- /dev/null
+++ clang/test/CodeGen/X86/avx-cast-builtins.c
@@ -0,0 +1,112 @@
+// RUN: %clang_cc1 %s -O3 -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-unknown-unknown -target-feature +avx -target-feature +avx512f  -target-feature +avx512fp16 -S -o - | FileCheck %s
+
+
+#include 
+
+__m256d test_mm256_castpd128_pd256(__m128d A) {
+  // CHECK-LABEL: test_mm256_castpd128_pd256
+  // CHECK: # %bb.0:
+  // CHECK-NEXT:# kill: def $xmm0 killed $xmm0 def $ymm0
+  // CHECK-NEXT:ret{{[l|q]}}
+  return _mm256_castpd128_pd256(A);
+}
+
+__m256 test_mm256_castps128_ps256(__m128 A) {
+  // CHECK-LABEL: test_mm256_castps128_ps256
+  // C

[PATCH] D149548: [IR] Update to use new shufflevector semantics

2023-05-06 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a reviewer: nemanjai.
ManuelJBrito added a comment.

Hello Nemanja is the change  codegen for vec_promote OK?


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D149548/new/

https://reviews.llvm.org/D149548

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D149548: [IR] Update to use new shufflevector semantics

2023-05-18 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added reviewers: qiucf, lkail, shchenz.
ManuelJBrito added a comment.

Adding some PowerPC folks to review the change in codegen for vec_promote.
AFAICT its OK to have the undefined values be poison.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D149548/new/

https://reviews.llvm.org/D149548

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D149210: [IR] Change shufflevector undef mask to poison

2023-04-25 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito created this revision.
ManuelJBrito added reviewers: aqjune, efriedma, RKSimon, nikic, nlopes, spatel.
Herald added a reviewer: deadalnix.
Herald added subscribers: luke, kosarev, StephenFan, frasercrmck, okura, kuter, 
lebedev.ri, kerbowa, luismarques, apazos, sameer.abuasal, pengfei, s.egerton, 
dmgreen, Jim, jocewei, PkmX, arphaman, the_o, brucehoult, MartinMosbeck, 
rogfer01, edward-jones, zzheng, jrtc27, niosHD, sabuasal, simoncook, johnrusso, 
rbar, asb, kbarton, hiraditya, jvesely, nemanjai.
Herald added a reviewer: lebedev.ri.
Herald added a project: All.
ManuelJBrito requested review of this revision.
Herald added subscribers: llvm-commits, cfe-commits, pcwang-thead, MaskRay.
Herald added a reviewer: jdoerfert.
Herald added a reviewer: sstefan1.
Herald added projects: clang, LLVM.

With this patch an undefined mask in a shufflevector will be printed as poison.
This change is done to support the new shufflevector semantics for undefined 
mask elements (see  D148637 ).


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D149210

Files:
  clang/test/CodeGen/PowerPC/builtins-ppc-p10vector.c
  clang/test/CodeGen/PowerPC/builtins-ppc-p9vector.c
  clang/test/CodeGen/builtins-nondeterministic-value.c
  clang/test/CodeGen/builtinshufflevector2.c
  clang/test/CodeGen/ext-vector.c
  clang/test/CodeGenOpenCL/as_type.cl
  clang/test/CodeGenOpenCL/partial_initializer.cl
  clang/test/CodeGenOpenCL/preserve_vec3.cl
  clang/test/CodeGenOpenCL/vector_literals.cl
  clang/test/CodeGenOpenCL/vector_shufflevector.cl
  llvm/include/llvm-c/Core.h
  llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
  llvm/include/llvm/IR/Instructions.h
  llvm/lib/Analysis/InstructionSimplify.cpp
  llvm/lib/Analysis/ValueTracking.cpp
  llvm/lib/Analysis/VectorUtils.cpp
  llvm/lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp
  llvm/lib/IR/AsmWriter.cpp
  llvm/lib/IR/ConstantFold.cpp
  llvm/lib/IR/Core.cpp
  llvm/lib/IR/Instructions.cpp
  llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
  llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
  llvm/lib/Target/X86/X86TargetTransformInfo.cpp
  llvm/lib/Transforms/InstCombine/InstCombineCalls.cpp
  llvm/lib/Transforms/InstCombine/InstCombineSelect.cpp
  llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp
  llvm/lib/Transforms/InstCombine/InstCombineVectorOps.cpp
  llvm/lib/Transforms/InstCombine/InstructionCombining.cpp
  llvm/lib/Transforms/Scalar/ScalarizeMaskedMemIntrin.cpp
  llvm/lib/Transforms/Vectorize/SLPVectorizer.cpp
  llvm/lib/Transforms/Vectorize/VectorCombine.cpp
  llvm/test/Analysis/CostModel/RISCV/shuffle-extract_subvector.ll
  llvm/test/Analysis/CostModel/RISCV/shuffle-insert_subvector.ll
  llvm/test/Analysis/CostModel/X86/reduction.ll
  llvm/test/Analysis/CostModel/X86/shuffle-extract_subvector-codesize.ll
  llvm/test/Analysis/CostModel/X86/shuffle-extract_subvector-latency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-extract_subvector-sizelatency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-extract_subvector.ll
  llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-codesize.ll
  llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-latency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-sizelatency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector.ll
  llvm/test/Analysis/CostModel/X86/shuffle-single-src-codesize.ll
  llvm/test/Analysis/CostModel/X86/shuffle-single-src-latency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-single-src-sizelatency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-single-src.ll
  llvm/test/CodeGen/AMDGPU/amdgpu-codegenprepare-break-large-phis-heuristics.ll
  llvm/test/CodeGen/AMDGPU/rewrite-out-arguments.ll
  llvm/test/CodeGen/Generic/expand-experimental-reductions.ll
  llvm/test/CodeGen/PowerPC/arg_promotion.ll
  llvm/test/Instrumentation/MemorySanitizer/avx-intrinsics-x86.ll
  llvm/test/Transforms/Attributor/nofpclass.ll
  llvm/test/Transforms/CodeGenPrepare/X86/x86-shuffle-sink-inseltpoison.ll
  llvm/test/Transforms/CodeGenPrepare/X86/x86-shuffle-sink.ll
  llvm/test/Transforms/DeadStoreElimination/masked-dead-store-inseltpoison.ll
  llvm/test/Transforms/DeadStoreElimination/masked-dead-store.ll
  llvm/test/Transforms/InstCombine/AArch64/demandelts.ll
  
llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-demanded-vector-elts-inseltpoison.ll
  llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-demanded-vector-elts.ll
  llvm/test/Transforms/InstCombine/X86/x86-addsub-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-addsub.ll
  llvm/test/Transforms/InstCombine/X86/x86-avx2-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-avx2.ll
  llvm/test/Transforms/InstCombine/X86/x86-avx512-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-avx512.ll
  llvm/test/Transforms/InstCombine/X86/x86-muldq-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-muldq.ll
  llvm/test/Transforms/InstCombine/X86/x86-pack-inseltpoison.ll
  llvm

[PATCH] D149210: [IR] Change shufflevector undef mask to poison

2023-04-26 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

In D149210#4298124 , @nikic wrote:

> Could you please split the change to printing + the test updates from all the 
> other changes? The code changes get lost in the large diff.

Yes, that makes sense. The internal representation can be addressed separately 
from the printing.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D149210/new/

https://reviews.llvm.org/D149210

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D149210: [IR] Change shufflevector undef mask to poison

2023-04-26 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 517150.
ManuelJBrito added a comment.

Split patch into print update and internal representation.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D149210/new/

https://reviews.llvm.org/D149210

Files:
  clang/test/CodeGen/PowerPC/builtins-ppc-p10vector.c
  clang/test/CodeGen/PowerPC/builtins-ppc-p9vector.c
  clang/test/CodeGen/builtins-nondeterministic-value.c
  clang/test/CodeGen/builtinshufflevector2.c
  clang/test/CodeGen/ext-vector.c
  clang/test/CodeGenOpenCL/as_type.cl
  clang/test/CodeGenOpenCL/partial_initializer.cl
  clang/test/CodeGenOpenCL/preserve_vec3.cl
  clang/test/CodeGenOpenCL/vector_literals.cl
  clang/test/CodeGenOpenCL/vector_shufflevector.cl
  llvm/lib/IR/AsmWriter.cpp
  llvm/test/Analysis/CostModel/RISCV/shuffle-extract_subvector.ll
  llvm/test/Analysis/CostModel/RISCV/shuffle-insert_subvector.ll
  llvm/test/Analysis/CostModel/X86/reduction.ll
  llvm/test/Analysis/CostModel/X86/shuffle-extract_subvector-codesize.ll
  llvm/test/Analysis/CostModel/X86/shuffle-extract_subvector-latency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-extract_subvector-sizelatency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-extract_subvector.ll
  llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-codesize.ll
  llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-latency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-sizelatency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector.ll
  llvm/test/Analysis/CostModel/X86/shuffle-single-src-codesize.ll
  llvm/test/Analysis/CostModel/X86/shuffle-single-src-latency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-single-src-sizelatency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-single-src.ll
  llvm/test/CodeGen/AMDGPU/amdgpu-codegenprepare-break-large-phis-heuristics.ll
  llvm/test/CodeGen/AMDGPU/rewrite-out-arguments.ll
  llvm/test/CodeGen/Generic/expand-experimental-reductions.ll
  llvm/test/CodeGen/PowerPC/arg_promotion.ll
  llvm/test/Instrumentation/MemorySanitizer/avx-intrinsics-x86.ll
  llvm/test/Transforms/Attributor/nofpclass.ll
  llvm/test/Transforms/CodeGenPrepare/X86/x86-shuffle-sink-inseltpoison.ll
  llvm/test/Transforms/CodeGenPrepare/X86/x86-shuffle-sink.ll
  llvm/test/Transforms/DeadStoreElimination/masked-dead-store-inseltpoison.ll
  llvm/test/Transforms/DeadStoreElimination/masked-dead-store.ll
  llvm/test/Transforms/InstCombine/AArch64/demandelts.ll
  
llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-demanded-vector-elts-inseltpoison.ll
  llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-demanded-vector-elts.ll
  llvm/test/Transforms/InstCombine/X86/x86-addsub-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-addsub.ll
  llvm/test/Transforms/InstCombine/X86/x86-avx2-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-avx2.ll
  llvm/test/Transforms/InstCombine/X86/x86-avx512-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-avx512.ll
  llvm/test/Transforms/InstCombine/X86/x86-muldq-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-muldq.ll
  llvm/test/Transforms/InstCombine/X86/x86-pack-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-pack.ll
  llvm/test/Transforms/InstCombine/X86/x86-pshufb-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-pshufb.ll
  llvm/test/Transforms/InstCombine/X86/x86-sse4a-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-sse4a.ll
  llvm/test/Transforms/InstCombine/X86/x86-vpermil-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-vpermil.ll
  llvm/test/Transforms/InstCombine/bitreverse.ll
  llvm/test/Transforms/InstCombine/broadcast-inseltpoison.ll
  llvm/test/Transforms/InstCombine/broadcast.ll
  llvm/test/Transforms/InstCombine/bswap-inseltpoison.ll
  llvm/test/Transforms/InstCombine/bswap.ll
  llvm/test/Transforms/InstCombine/canonicalize-vector-insert.ll
  llvm/test/Transforms/InstCombine/extractelement-inseltpoison.ll
  llvm/test/Transforms/InstCombine/extractelement.ll
  llvm/test/Transforms/InstCombine/insert-const-shuf-inseltpoison.ll
  llvm/test/Transforms/InstCombine/insert-const-shuf.ll
  llvm/test/Transforms/InstCombine/insert-extract-shuffle-inseltpoison.ll
  llvm/test/Transforms/InstCombine/insert-extract-shuffle.ll
  llvm/test/Transforms/InstCombine/masked_intrinsics-inseltpoison.ll
  llvm/test/Transforms/InstCombine/masked_intrinsics.ll
  llvm/test/Transforms/InstCombine/minmax-intrinsics.ll
  llvm/test/Transforms/InstCombine/nsw-inseltpoison.ll
  llvm/test/Transforms/InstCombine/nsw.ll
  llvm/test/Transforms/InstCombine/reduction-shufflevector.ll
  llvm/test/Transforms/InstCombine/select-extractelement-inseltpoison.ll
  llvm/test/Transforms/InstCombine/select-extractelement.ll
  llvm/test/Transforms/InstCombine/select-select.ll
  llvm/test/Transforms/InstCombine/shuffle-binop.ll
  llvm/test/Transforms/InstCombine/shuffle-select-narrow-inseltpoison.ll
  llvm/test/Transforms/InstCombine/shuffle-select-narrow.ll
  llv

[PATCH] D149210: [IR] Change shufflevector undef mask to poison

2023-04-27 Thread Manuel Brito via Phabricator via cfe-commits 
This revision was landed with ongoing or failed builds.
This revision was automatically updated to reflect the committed changes.
Closed by commit rG8b56da5e9f3b: [IR] Change shufflevector undef mask to poison 
(authored by ManuelJBrito).

Changed prior to commit:
  https://reviews.llvm.org/D149210?vs=517150&id=517542#toc

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D149210/new/

https://reviews.llvm.org/D149210

Files:
  clang/test/CodeGen/PowerPC/builtins-ppc-p10vector.c
  clang/test/CodeGen/PowerPC/builtins-ppc-p9vector.c
  clang/test/CodeGen/builtins-nondeterministic-value.c
  clang/test/CodeGen/builtinshufflevector2.c
  clang/test/CodeGen/ext-vector.c
  clang/test/CodeGenOpenCL/as_type.cl
  clang/test/CodeGenOpenCL/partial_initializer.cl
  clang/test/CodeGenOpenCL/preserve_vec3.cl
  clang/test/CodeGenOpenCL/vector_literals.cl
  clang/test/CodeGenOpenCL/vector_shufflevector.cl
  llvm/lib/IR/AsmWriter.cpp
  llvm/test/Analysis/CostModel/RISCV/shuffle-extract_subvector.ll
  llvm/test/Analysis/CostModel/RISCV/shuffle-insert_subvector.ll
  llvm/test/Analysis/CostModel/X86/reduction.ll
  llvm/test/Analysis/CostModel/X86/shuffle-extract_subvector-codesize.ll
  llvm/test/Analysis/CostModel/X86/shuffle-extract_subvector-latency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-extract_subvector-sizelatency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-extract_subvector.ll
  llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-codesize.ll
  llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-latency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector-sizelatency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-insert_subvector.ll
  llvm/test/Analysis/CostModel/X86/shuffle-single-src-codesize.ll
  llvm/test/Analysis/CostModel/X86/shuffle-single-src-latency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-single-src-sizelatency.ll
  llvm/test/Analysis/CostModel/X86/shuffle-single-src.ll
  llvm/test/CodeGen/AMDGPU/amdgpu-codegenprepare-break-large-phis-heuristics.ll
  llvm/test/CodeGen/AMDGPU/rewrite-out-arguments.ll
  llvm/test/CodeGen/Generic/expand-experimental-reductions.ll
  llvm/test/CodeGen/PowerPC/arg_promotion.ll
  llvm/test/Instrumentation/MemorySanitizer/avx-intrinsics-x86.ll
  llvm/test/Transforms/Attributor/nofpclass.ll
  llvm/test/Transforms/CodeGenPrepare/X86/x86-shuffle-sink-inseltpoison.ll
  llvm/test/Transforms/CodeGenPrepare/X86/x86-shuffle-sink.ll
  llvm/test/Transforms/DeadStoreElimination/masked-dead-store-inseltpoison.ll
  llvm/test/Transforms/DeadStoreElimination/masked-dead-store.ll
  llvm/test/Transforms/InstCombine/AArch64/demandelts.ll
  
llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-demanded-vector-elts-inseltpoison.ll
  llvm/test/Transforms/InstCombine/AMDGPU/amdgcn-demanded-vector-elts.ll
  llvm/test/Transforms/InstCombine/X86/x86-addsub-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-addsub.ll
  llvm/test/Transforms/InstCombine/X86/x86-avx2-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-avx2.ll
  llvm/test/Transforms/InstCombine/X86/x86-avx512-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-avx512.ll
  llvm/test/Transforms/InstCombine/X86/x86-muldq-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-muldq.ll
  llvm/test/Transforms/InstCombine/X86/x86-pack-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-pack.ll
  llvm/test/Transforms/InstCombine/X86/x86-pshufb-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-pshufb.ll
  llvm/test/Transforms/InstCombine/X86/x86-sse4a-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-sse4a.ll
  llvm/test/Transforms/InstCombine/X86/x86-vpermil-inseltpoison.ll
  llvm/test/Transforms/InstCombine/X86/x86-vpermil.ll
  llvm/test/Transforms/InstCombine/bitreverse.ll
  llvm/test/Transforms/InstCombine/broadcast-inseltpoison.ll
  llvm/test/Transforms/InstCombine/broadcast.ll
  llvm/test/Transforms/InstCombine/bswap-inseltpoison.ll
  llvm/test/Transforms/InstCombine/bswap.ll
  llvm/test/Transforms/InstCombine/canonicalize-vector-insert.ll
  llvm/test/Transforms/InstCombine/extractelement-inseltpoison.ll
  llvm/test/Transforms/InstCombine/extractelement.ll
  llvm/test/Transforms/InstCombine/insert-const-shuf-inseltpoison.ll
  llvm/test/Transforms/InstCombine/insert-const-shuf.ll
  llvm/test/Transforms/InstCombine/insert-extract-shuffle-inseltpoison.ll
  llvm/test/Transforms/InstCombine/insert-extract-shuffle.ll
  llvm/test/Transforms/InstCombine/masked_intrinsics-inseltpoison.ll
  llvm/test/Transforms/InstCombine/masked_intrinsics.ll
  llvm/test/Transforms/InstCombine/minmax-intrinsics.ll
  llvm/test/Transforms/InstCombine/nsw-inseltpoison.ll
  llvm/test/Transforms/InstCombine/nsw.ll
  llvm/test/Transforms/InstCombine/reduction-shufflevector.ll
  llvm/test/Transforms/InstCombine/select-extractelement-inseltpoison.ll
  llvm/test/Transforms/InstCombine/select-extractelement.ll
  llvm/test/Transforms/InstCombine/select-select.ll
  llvm/t

[PATCH] D149548: [IR] Update to use new shufflevector semantics

2023-05-01 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 518434.
ManuelJBrito added a comment.
Herald added a project: clang.
Herald added a subscriber: cfe-commits.

This change is visible in clang in particular in codegen for the vec_promote 
builtin.
It now produces poison instead of undef for the undefined elements.
I think this is OK according to IBM's 
documentation[https://www.ibm.com/docs/en/epfz/5.3?topic=book-semantics]:

> When a result or behavior is undefined, it is something you “must not” do.
> Use of an undefined feature is likely to produce different results on 
> different > implementations or releases of a PL/I product.
> The application program is considered to be in error.




Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D149548/new/

https://reviews.llvm.org/D149548

Files:
  clang/test/CodeGen/SystemZ/builtins-systemz-zvector-constrained.c
  clang/test/CodeGen/SystemZ/builtins-systemz-zvector.c
  clang/test/CodeGen/SystemZ/builtins-systemz-zvector2-constrained.c
  clang/test/CodeGen/SystemZ/builtins-systemz-zvector2.c
  llvm/lib/IR/ConstantFold.cpp
  llvm/lib/IR/Instructions.cpp
  llvm/test/Transforms/InstCombine/shufflevec-constant.ll

Index: llvm/test/Transforms/InstCombine/shufflevec-constant.ll
===
--- llvm/test/Transforms/InstCombine/shufflevec-constant.ll
+++ llvm/test/Transforms/InstCombine/shufflevec-constant.ll
@@ -18,7 +18,7 @@
 
 define <4 x i1> @shuffle_undef_fixed() {
 ; CHECK-LABEL: @shuffle_undef_fixed(
-; CHECK-NEXT:ret <4 x i1> undef
+; CHECK-NEXT:ret <4 x i1> poison
 ;
   %S = shufflevector <4 x i1> zeroinitializer, <4 x i1> zeroinitializer, <4 x i32> undef
   ret <4 x i1> %S
@@ -26,7 +26,7 @@
 
 define  @suffle_undef_scalable() {
 ; CHECK-LABEL: @suffle_undef_scalable(
-; CHECK-NEXT:ret  undef
+; CHECK-NEXT:ret  poison
 ;
   %S = shufflevector  zeroinitializer,  zeroinitializer,  undef
   ret  %S
Index: llvm/lib/IR/Instructions.cpp
===
--- llvm/lib/IR/Instructions.cpp
+++ llvm/lib/IR/Instructions.cpp
@@ -2290,7 +2290,7 @@
   SmallVector MaskConst;
   for (int Elem : Mask) {
 if (Elem == PoisonMaskElem)
-  MaskConst.push_back(UndefValue::get(Int32Ty));
+  MaskConst.push_back(PoisonValue::get(Int32Ty));
 else
   MaskConst.push_back(ConstantInt::get(Int32Ty, Elem));
   }
Index: llvm/lib/IR/ConstantFold.cpp
===
--- llvm/lib/IR/ConstantFold.cpp
+++ llvm/lib/IR/ConstantFold.cpp
@@ -710,9 +710,9 @@
   ElementCount::get(MaskNumElts, isa(V1VTy));
   Type *EltTy = V1VTy->getElementType();
 
-  // Undefined shuffle mask -> undefined value.
+  // Poison shuffle mask -> poison value.
   if (all_of(Mask, [](int Elt) { return Elt == PoisonMaskElem; })) {
-return UndefValue::get(VectorType::get(EltTy, MaskEltCount));
+return PoisonValue::get(VectorType::get(EltTy, MaskEltCount));
   }
 
   // If the mask is all zeros this is a splat, no need to go through all
Index: clang/test/CodeGen/SystemZ/builtins-systemz-zvector2.c
===
--- clang/test/CodeGen/SystemZ/builtins-systemz-zvector2.c
+++ clang/test/CodeGen/SystemZ/builtins-systemz-zvector2.c
@@ -102,10 +102,10 @@
   // CHECK-ASM: vlvgg
 
   vf = vec_promote(f, idx);
-  // CHECK: insertelement <4 x float> undef, float %{{.*}}, i32 %{{.*}}
+  // CHECK: insertelement <4 x float> poison, float %{{.*}}, i32 %{{.*}}
   // CHECK-ASM: vlvgf
   vd = vec_promote(d, idx);
-  // CHECK: insertelement <2 x double> undef, double %{{.*}}, i32 %{{.*}}
+  // CHECK: insertelement <2 x double> poison, double %{{.*}}, i32 %{{.*}}
   // CHECK-ASM: vlvgg
 
   vf = vec_insert_and_zero(cptrf);
Index: clang/test/CodeGen/SystemZ/builtins-systemz-zvector2-constrained.c
===
--- clang/test/CodeGen/SystemZ/builtins-systemz-zvector2-constrained.c
+++ clang/test/CodeGen/SystemZ/builtins-systemz-zvector2-constrained.c
@@ -70,10 +70,10 @@
   // CHECK-ASM: vlvgg
 
   vf = vec_promote(f, idx);
-  // CHECK: insertelement <4 x float> undef, float %{{.*}}, i32 %{{.*}}
+  // CHECK: insertelement <4 x float> poison, float %{{.*}}, i32 %{{.*}}
   // CHECK-ASM: vlvgf
   vd = vec_promote(d, idx);
-  // CHECK: insertelement <2 x double> undef, double %{{.*}}, i32 %{{.*}}
+  // CHECK: insertelement <2 x double> poison, double %{{.*}}, i32 %{{.*}}
   // CHECK-ASM: vlvgg
 
   vf = vec_insert_and_zero(cptrf);
Index: clang/test/CodeGen/SystemZ/builtins-systemz-zvector.c
===
--- clang/test/CodeGen/SystemZ/builtins-systemz-zvector.c
+++ clang/test/CodeGen/SystemZ/builtins-systemz-zvector.c
@@ -166,31 +166,31 @@
   // CHECK-ASM: vlvgg
 
   vsc = vec_promote(sc, idx);
-  // CHECK: insertelement <16 x i8> undef, i8 %{{.*}}, i32 %{{

[PATCH] D142388: [clang] Add builtin_nondeterministic_value

2023-06-09 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added inline comments.



Comment at: clang/test/CodeGen/builtins-nondeterministic-value.c:26
+// CHECK-LABEL: entry
+// CHECK: [[A:%.*]] = alloca double, align 8
+// CHECK: store double [[X:%.*]], ptr [[A]], align 8

zixuan-wu wrote:
> hi, @ManuelJBrito , because double is 4 alignment in CSKY target, could you 
> please update this with capture match pattern which makes it more adaptable?
Hi, I'll make a patch dropping the alignment since it's not relevant for what 
we are testing here.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142388/new/

https://reviews.llvm.org/D142388

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D152547: [clang][NFC] Drop alignment in builtin-nondeterministic-value test

2023-06-09 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito created this revision.
ManuelJBrito added reviewers: zixuan-wu, erichkeane.
Herald added a subscriber: mgrang.
Herald added a project: All.
ManuelJBrito requested review of this revision.
Herald added a project: clang.
Herald added a subscriber: cfe-commits.

Drop alignment to allow tests to run in different platforms.
Following @zixuan-wu 's request in D142388 .


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D152547

Files:
  clang/test/CodeGen/builtins-nondeterministic-value.c


Index: clang/test/CodeGen/builtins-nondeterministic-value.c
===
--- clang/test/CodeGen/builtins-nondeterministic-value.c
+++ clang/test/CodeGen/builtins-nondeterministic-value.c
@@ -5,8 +5,8 @@
 
 int clang_nondet_i( int x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca i32, align 4
-// CHECK: store i32 [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[A:%.*]] = alloca i32, align
+// CHECK: store i32 [[X:%.*]], ptr [[A]], align
 // CHECK: [[R:%.*]] = freeze i32 poison
 // CHECK: ret i32 [[R]]
   return __builtin_nondeterministic_value(x);
@@ -14,8 +14,8 @@
 
 float clang_nondet_f( float x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca float, align 4
-// CHECK: store float [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[A:%.*]] = alloca float, align
+// CHECK: store float [[X:%.*]], ptr [[A]], align
 // CHECK: [[R:%.*]] = freeze float poison
 // CHECK: ret float [[R]]
   return __builtin_nondeterministic_value(x);
@@ -23,8 +23,8 @@
 
 double clang_nondet_d( double x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca double, align 8
-// CHECK: store double [[X:%.*]], ptr [[A]], align 8
+// CHECK: [[A:%.*]] = alloca double, align
+// CHECK: store double [[X:%.*]], ptr [[A]], align
 // CHECK: [[R:%.*]] = freeze double poison
 // CHECK: ret double [[R]]
   return __builtin_nondeterministic_value(x);
@@ -32,9 +32,9 @@
 
 _Bool clang_nondet_b( _Bool x) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[A:%.*]] = alloca i8, align
 // CHECK: [[B:%.*]] = zext i1 %x to i8
-// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: store i8 [[B]], ptr [[A]], align
 // CHECK: [[R:%.*]] = freeze i1 poison
 // CHECK: ret i1 [[R]]
   return __builtin_nondeterministic_value(x);


Index: clang/test/CodeGen/builtins-nondeterministic-value.c
===
--- clang/test/CodeGen/builtins-nondeterministic-value.c
+++ clang/test/CodeGen/builtins-nondeterministic-value.c
@@ -5,8 +5,8 @@
 
 int clang_nondet_i( int x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca i32, align 4
-// CHECK: store i32 [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[A:%.*]] = alloca i32, align
+// CHECK: store i32 [[X:%.*]], ptr [[A]], align
 // CHECK: [[R:%.*]] = freeze i32 poison
 // CHECK: ret i32 [[R]]
   return __builtin_nondeterministic_value(x);
@@ -14,8 +14,8 @@
 
 float clang_nondet_f( float x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca float, align 4
-// CHECK: store float [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[A:%.*]] = alloca float, align
+// CHECK: store float [[X:%.*]], ptr [[A]], align
 // CHECK: [[R:%.*]] = freeze float poison
 // CHECK: ret float [[R]]
   return __builtin_nondeterministic_value(x);
@@ -23,8 +23,8 @@
 
 double clang_nondet_d( double x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca double, align 8
-// CHECK: store double [[X:%.*]], ptr [[A]], align 8
+// CHECK: [[A:%.*]] = alloca double, align
+// CHECK: store double [[X:%.*]], ptr [[A]], align
 // CHECK: [[R:%.*]] = freeze double poison
 // CHECK: ret double [[R]]
   return __builtin_nondeterministic_value(x);
@@ -32,9 +32,9 @@
 
 _Bool clang_nondet_b( _Bool x) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[A:%.*]] = alloca i8, align
 // CHECK: [[B:%.*]] = zext i1 %x to i8
-// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: store i8 [[B]], ptr [[A]], align
 // CHECK: [[R:%.*]] = freeze i1 poison
 // CHECK: ret i1 [[R]]
   return __builtin_nondeterministic_value(x);
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D149548: [IR] Update to use new shufflevector semantics

2023-06-09 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

In D149548#4405173 , @qiucf wrote:

> Why this changes IR output of following case?
>
>   // RUN: clang vecpromote.c -S -o - -O0 -target s390x-linux-gnu -fzvector 
> -emit-llvm
>   #include 
>   
>   vector int si;
>   int g;
>   int i;
>   
>   void foo() {
> si = vec_promote(g, i);
>   }
>   
>   // store <4 x i32> undef, ptr %__vec.i, align 16, !noalias !6  ; this undef 
> becomes poison after this patch
>   // %2 = load i32, ptr %__scalar.addr.i, align 4, !noalias !6
>   // %3 = load i32, ptr %__index.addr.i, align 4, !noalias !6
>   // %4 = load <4 x i32>, ptr %__vec.i, align 16, !noalias !6
>   // %vecins.i = insertelement <4 x i32> %4, i32 %2, i32 %and.i

I think it's because vec_promote is implemented in clang using a shufflevector 
with an undefined mask (-1). 
ConstantFold sees that we have a fully undefined mask and it folds the 
shufflevector to poison.(previously undef)
My concern : is it Ok to have the undefined values of the vector be poison or 
was the undef used to abide to some specific semantics?
Hope this makes sense.

> I see no PowerPC related case changes. Maybe adding SystemZ folks for 
> comments @uweigand @jonpa

My bad I pinged the wrong people, thanks for adding SystemZ folks.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D149548/new/

https://reviews.llvm.org/D149548

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D152547: [clang][NFC] Drop alignment in builtin-nondeterministic-value test

2023-06-11 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 530321.
ManuelJBrito added a comment.

Completely drop alignment.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D152547/new/

https://reviews.llvm.org/D152547

Files:
  clang/test/CodeGen/builtins-nondeterministic-value.c


Index: clang/test/CodeGen/builtins-nondeterministic-value.c
===
--- clang/test/CodeGen/builtins-nondeterministic-value.c
+++ clang/test/CodeGen/builtins-nondeterministic-value.c
@@ -5,8 +5,8 @@
 
 int clang_nondet_i( int x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca i32, align 4
-// CHECK: store i32 [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[A:%.*]] = alloca i32
+// CHECK: store i32 [[X:%.*]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze i32 poison
 // CHECK: ret i32 [[R]]
   return __builtin_nondeterministic_value(x);
@@ -14,8 +14,8 @@
 
 float clang_nondet_f( float x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca float, align 4
-// CHECK: store float [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[A:%.*]] = alloca float
+// CHECK: store float [[X:%.*]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze float poison
 // CHECK: ret float [[R]]
   return __builtin_nondeterministic_value(x);
@@ -23,8 +23,8 @@
 
 double clang_nondet_d( double x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca double, align 8
-// CHECK: store double [[X:%.*]], ptr [[A]], align 8
+// CHECK: [[A:%.*]] = alloca double
+// CHECK: store double [[X:%.*]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze double poison
 // CHECK: ret double [[R]]
   return __builtin_nondeterministic_value(x);
@@ -32,9 +32,9 @@
 
 _Bool clang_nondet_b( _Bool x) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[A:%.*]] = alloca i8
 // CHECK: [[B:%.*]] = zext i1 %x to i8
-// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: store i8 [[B]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze i1 poison
 // CHECK: ret i1 [[R]]
   return __builtin_nondeterministic_value(x);
@@ -42,19 +42,19 @@
 
 void clang_nondet_fv( ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca <4 x float>, align
+// CHECK: [[A:%.*]] = alloca <4 x float>
 // CHECK: [[R:%.*]] = freeze <4 x float> poison
-// CHECK: store <4 x float> [[R]], ptr [[A]], align
+// CHECK: store <4 x float> [[R]], ptr [[A]]
 // CHECK: ret void
   float4 x = __builtin_nondeterministic_value(x);
 }
 
 void clang_nondet_bv( ) {
-// CHECK: [[A:%.*]] = alloca i8, align
+// CHECK: [[A:%.*]] = alloca i8
 // CHECK: [[V:%.*]] = freeze <4 x i1> poison
 // CHECK: [[SV:%.*]] = shufflevector <4 x i1> [[V]], <4 x i1> poison, <8 x 
i32> 
 // CHECK: [[BC:%.*]] = bitcast <8 x i1> [[SV]] to i8
-// CHECK: store i8 [[BC]], ptr [[A]], align
+// CHECK: store i8 [[BC]], ptr [[A]]
 // CHECK: ret void
   bool4 x = __builtin_nondeterministic_value(x);
 }


Index: clang/test/CodeGen/builtins-nondeterministic-value.c
===
--- clang/test/CodeGen/builtins-nondeterministic-value.c
+++ clang/test/CodeGen/builtins-nondeterministic-value.c
@@ -5,8 +5,8 @@
 
 int clang_nondet_i( int x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca i32, align 4
-// CHECK: store i32 [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[A:%.*]] = alloca i32
+// CHECK: store i32 [[X:%.*]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze i32 poison
 // CHECK: ret i32 [[R]]
   return __builtin_nondeterministic_value(x);
@@ -14,8 +14,8 @@
 
 float clang_nondet_f( float x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca float, align 4
-// CHECK: store float [[X:%.*]], ptr [[A]], align 4
+// CHECK: [[A:%.*]] = alloca float
+// CHECK: store float [[X:%.*]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze float poison
 // CHECK: ret float [[R]]
   return __builtin_nondeterministic_value(x);
@@ -23,8 +23,8 @@
 
 double clang_nondet_d( double x ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca double, align 8
-// CHECK: store double [[X:%.*]], ptr [[A]], align 8
+// CHECK: [[A:%.*]] = alloca double
+// CHECK: store double [[X:%.*]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze double poison
 // CHECK: ret double [[R]]
   return __builtin_nondeterministic_value(x);
@@ -32,9 +32,9 @@
 
 _Bool clang_nondet_b( _Bool x) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca i8, align 1
+// CHECK: [[A:%.*]] = alloca i8
 // CHECK: [[B:%.*]] = zext i1 %x to i8
-// CHECK: store i8 [[B]], ptr [[A]], align 1
+// CHECK: store i8 [[B]], ptr [[A]]
 // CHECK: [[R:%.*]] = freeze i1 poison
 // CHECK: ret i1 [[R]]
   return __builtin_nondeterministic_value(x);
@@ -42,19 +42,19 @@
 
 void clang_nondet_fv( ) {
 // CHECK-LABEL: entry
-// CHECK: [[A:%.*]] = alloca <4 x float>, align
+// CHECK: [[A:%.*]] = alloca <4 x float>
 // CHECK: [[R:%.*]] = freeze <4 x float> poison
-// CHECK: store <4 x float> [[R]], ptr [[A]], align
+// CHECK: store <4 x float> [[R]], ptr [[A]]
 // CHECK: ret void
   float4 x = __builtin_nondeterministic_value(x);

[PATCH] D149548: [IR] Update to use new shufflevector semantics

2023-06-13 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

In D149548#4413762 , @uweigand wrote:

> In D149548#4413639 , @nlopes wrote:
>
>> If a vector is fully initialized with `insertvector` (i.e., one operation 
>> per index), then the value of the base vector is irrelevant. It can be 
>> poison.
>> Poison in vectors is element-wise.  doesn't propagate to 
>> .
>
> OK, that should be fine for SystemZ then.  Thanks!

Thanks! I'll move forward then.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D149548/new/

https://reviews.llvm.org/D149548

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D149548: [IR] Update to use new shufflevector semantics

2023-06-13 Thread Manuel Brito via Phabricator via cfe-commits 
This revision was landed with ongoing or failed builds.
This revision was automatically updated to reflect the committed changes.
Closed by commit rG22f5dc7501b5: [IR] Update to use new shufflevector semantics 
(authored by ManuelJBrito).

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D149548/new/

https://reviews.llvm.org/D149548

Files:
  clang/test/CodeGen/SystemZ/builtins-systemz-zvector-constrained.c
  clang/test/CodeGen/SystemZ/builtins-systemz-zvector.c
  clang/test/CodeGen/SystemZ/builtins-systemz-zvector2-constrained.c
  clang/test/CodeGen/SystemZ/builtins-systemz-zvector2.c
  llvm/lib/IR/ConstantFold.cpp
  llvm/lib/IR/Instructions.cpp
  llvm/test/Transforms/InstCombine/shufflevec-constant.ll

Index: llvm/test/Transforms/InstCombine/shufflevec-constant.ll
===
--- llvm/test/Transforms/InstCombine/shufflevec-constant.ll
+++ llvm/test/Transforms/InstCombine/shufflevec-constant.ll
@@ -18,7 +18,7 @@
 
 define <4 x i1> @shuffle_undef_fixed() {
 ; CHECK-LABEL: @shuffle_undef_fixed(
-; CHECK-NEXT:ret <4 x i1> undef
+; CHECK-NEXT:ret <4 x i1> poison
 ;
   %S = shufflevector <4 x i1> zeroinitializer, <4 x i1> zeroinitializer, <4 x i32> undef
   ret <4 x i1> %S
@@ -26,7 +26,7 @@
 
 define  @suffle_undef_scalable() {
 ; CHECK-LABEL: @suffle_undef_scalable(
-; CHECK-NEXT:ret  undef
+; CHECK-NEXT:ret  poison
 ;
   %S = shufflevector  zeroinitializer,  zeroinitializer,  undef
   ret  %S
Index: llvm/lib/IR/Instructions.cpp
===
--- llvm/lib/IR/Instructions.cpp
+++ llvm/lib/IR/Instructions.cpp
@@ -2290,7 +2290,7 @@
   SmallVector MaskConst;
   for (int Elem : Mask) {
 if (Elem == PoisonMaskElem)
-  MaskConst.push_back(UndefValue::get(Int32Ty));
+  MaskConst.push_back(PoisonValue::get(Int32Ty));
 else
   MaskConst.push_back(ConstantInt::get(Int32Ty, Elem));
   }
Index: llvm/lib/IR/ConstantFold.cpp
===
--- llvm/lib/IR/ConstantFold.cpp
+++ llvm/lib/IR/ConstantFold.cpp
@@ -710,9 +710,9 @@
   ElementCount::get(MaskNumElts, isa(V1VTy));
   Type *EltTy = V1VTy->getElementType();
 
-  // Undefined shuffle mask -> undefined value.
+  // Poison shuffle mask -> poison value.
   if (all_of(Mask, [](int Elt) { return Elt == PoisonMaskElem; })) {
-return UndefValue::get(VectorType::get(EltTy, MaskEltCount));
+return PoisonValue::get(VectorType::get(EltTy, MaskEltCount));
   }
 
   // If the mask is all zeros this is a splat, no need to go through all
Index: clang/test/CodeGen/SystemZ/builtins-systemz-zvector2.c
===
--- clang/test/CodeGen/SystemZ/builtins-systemz-zvector2.c
+++ clang/test/CodeGen/SystemZ/builtins-systemz-zvector2.c
@@ -102,10 +102,10 @@
   // CHECK-ASM: vlvgg
 
   vf = vec_promote(f, idx);
-  // CHECK: insertelement <4 x float> undef, float %{{.*}}, i32 %{{.*}}
+  // CHECK: insertelement <4 x float> poison, float %{{.*}}, i32 %{{.*}}
   // CHECK-ASM: vlvgf
   vd = vec_promote(d, idx);
-  // CHECK: insertelement <2 x double> undef, double %{{.*}}, i32 %{{.*}}
+  // CHECK: insertelement <2 x double> poison, double %{{.*}}, i32 %{{.*}}
   // CHECK-ASM: vlvgg
 
   vf = vec_insert_and_zero(cptrf);
Index: clang/test/CodeGen/SystemZ/builtins-systemz-zvector2-constrained.c
===
--- clang/test/CodeGen/SystemZ/builtins-systemz-zvector2-constrained.c
+++ clang/test/CodeGen/SystemZ/builtins-systemz-zvector2-constrained.c
@@ -70,10 +70,10 @@
   // CHECK-ASM: vlvgg
 
   vf = vec_promote(f, idx);
-  // CHECK: insertelement <4 x float> undef, float %{{.*}}, i32 %{{.*}}
+  // CHECK: insertelement <4 x float> poison, float %{{.*}}, i32 %{{.*}}
   // CHECK-ASM: vlvgf
   vd = vec_promote(d, idx);
-  // CHECK: insertelement <2 x double> undef, double %{{.*}}, i32 %{{.*}}
+  // CHECK: insertelement <2 x double> poison, double %{{.*}}, i32 %{{.*}}
   // CHECK-ASM: vlvgg
 
   vf = vec_insert_and_zero(cptrf);
Index: clang/test/CodeGen/SystemZ/builtins-systemz-zvector.c
===
--- clang/test/CodeGen/SystemZ/builtins-systemz-zvector.c
+++ clang/test/CodeGen/SystemZ/builtins-systemz-zvector.c
@@ -166,31 +166,31 @@
   // CHECK-ASM: vlvgg
 
   vsc = vec_promote(sc, idx);
-  // CHECK: insertelement <16 x i8> undef, i8 %{{.*}}, i32 %{{.*}}
+  // CHECK: insertelement <16 x i8> poison, i8 %{{.*}}, i32 %{{.*}}
   // CHECK-ASM: vlvgb
   vuc = vec_promote(uc, idx);
-  // CHECK: insertelement <16 x i8> undef, i8 %{{.*}}, i32 %{{.*}}
+  // CHECK: insertelement <16 x i8> poison, i8 %{{.*}}, i32 %{{.*}}
   // CHECK-ASM: vlvgb
   vss = vec_promote(ss, idx);
-  // CHECK: insertelement <8 x i16> undef, i16 %{{.*}}, i32 %{{.*}}
+  // CHECK: insertelement <8 x i16> poison, i16 %{{.*}}, i32 %{{

[PATCH] D142388: [clang] Add builtin_nondeterministic_value

2023-06-16 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

In D142388#4425246 , @aaron.ballman 
wrote:

> In D142388#4415349 , @rsmith wrote:
>
>> I think `__builtin_any_value` works pretty well, and emphasizes that this 
>> can really return any value. I'd also be OK with 
>> `__builtin_convenient_value`, to emphasize that the compiler will pick 
>> something that's convenient for it, but I prefer `__builtin_any_value`.
>
> I can get behind `__builtin_any_value` as the name.
>
> @ManuelJBrito: would you be willing to make these modifications?

Yes, I can change the name to `__builtin_any_value` . 
Where do we stand in terms of further restricting the types for which we can 
use this? I'm a bit out of my depth in regards to all the subtleties.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142388/new/

https://reviews.llvm.org/D142388

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D152658: [InstCombine] Change SimplifyDemandedVectorElts to use PoisonElts instead of UndefElts

2023-06-16 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added inline comments.



Comment at: llvm/lib/Target/X86/X86InstCombineIntrinsic.cpp:3102-3105
   case Intrinsic::x86_sse4a_extrqi:
   case Intrinsic::x86_sse4a_insertq:
   case Intrinsic::x86_sse4a_insertqi:
+PoisonElts.setHighBits(VWidth / 2);

nlopes wrote:
> this change doesn't look correct. The bits are undefined, doesn't seem like 
> you can use poison here.
Thanks, I'll fix it. This will cause some regressions in the cases where only 
the upper half is demanded.
Eventually we could fold those cases to freeze(poison).



Comment at: llvm/test/Transforms/InstCombine/vec_shuffle.ll:1301
 ; CHECK-LABEL: @fmul_splat_constant(
-; CHECK-NEXT:[[TMP1:%.*]] = fmul <2 x float> [[X:%.*]], 
-; CHECK-NEXT:[[R:%.*]] = shufflevector <2 x float> [[TMP1]], <2 x float> 
poison, <2 x i32> zeroinitializer
+  ; CHECK-NEXT:[[TMP1:%.*]] = shufflevector <2 x float> [[X:%.*]], <2 x 
float> undef, <2 x i32> zeroinitializer
+; CHECK-NEXT:[[R:%.*]] = fmul <2 x float> [[TMP1]], 

nlopes wrote:
> Why does it regress here (goes from poison to undef)?
This regression is due to the change in foldVectorBinop in 
InstructionCombining.cpp
It seems that without this canonicalization, demanded elements doesn't trigger 
causing the regression.

I changed it to match poison instead of undef, because of the following:

shufflevector < 4 x i32 > %1, < 4 x i32 > undef , <1, 2, 6, 7>  --> 
shufflevector < 4 x i32 > %1, < 4 x i32 > undef , <1, 2, undef, undef>

Now we don't do this because it would insert poison.

Not simplifying the mask causes the following assert to fail:

```
assert(ShMask[I] < (int)NumElts && "Not expecting narrowing shuffle") 
```


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D152658/new/

https://reviews.llvm.org/D152658

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D138755: [Clang][CodeGen] Use poison instead of undef for extra argument in __builtin_amdgcn_mov_dpp [NFC]

2022-12-06 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito edited reviewers, added: arsenm; removed: yaxunl.
ManuelJBrito added a comment.
Herald added a subscriber: wdng.

Is this change OK? It seems to me that here undef is just a placeholder. Was 
hoping that some AMDGPU folks could review it.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D138755/new/

https://reviews.llvm.org/D138755

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D138755: [Clang][CodeGen] Use poison instead of undef for extra argument in __builtin_amdgcn_mov_dpp [NFC]

2022-12-06 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito added a comment.

Thanks for the review arsenm!


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D138755/new/

https://reviews.llvm.org/D138755

___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

[PATCH] D139745: Use poison instead of undef where its used as placeholder[NFC]

2022-12-09 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito created this revision.
ManuelJBrito added a reviewer: nlopes.
Herald added subscribers: Enna1, ThomasRaoux, hiraditya.
Herald added a project: All.
ManuelJBrito requested review of this revision.
Herald added a reviewer: zuban32.
Herald added projects: clang, LLVM.
Herald added subscribers: llvm-commits, cfe-commits.

Replacing the following instances of undef with poison where undef is used as a 
placeholder for aggregate construction.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D139745

Files:
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/lib/CodeGen/CGException.cpp
  clang/lib/CodeGen/CGExprScalar.cpp
  clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq-bfloat.c
  clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq.c
  clang/test/CodeGen/builtinshufflevector2.c
  clang/test/CodeGenCXX/nrvo.cpp
  clang/test/CodeGenCXX/pr40771-ctad-with-lambda-copy-capture.cpp
  clang/test/CodeGenCXX/vla-consruct.cpp
  clang/test/CodeGenCoroutines/coro-eh-cleanup-exp-namespace.cpp
  clang/test/CodeGenCoroutines/coro-eh-cleanup.cpp
  clang/test/OpenMP/distribute_parallel_for_num_threads_codegen.cpp
  clang/test/OpenMP/distribute_parallel_for_simd_num_threads_codegen.cpp
  clang/test/OpenMP/single_codegen.cpp
  clang/test/OpenMP/teams_distribute_parallel_for_num_threads_codegen.cpp
  clang/test/OpenMP/teams_distribute_parallel_for_simd_num_threads_codegen.cpp
  clang/test/OpenMP/threadprivate_codegen.cpp
  llvm/lib/Analysis/ValueTracking.cpp
  llvm/lib/CodeGen/SjLjEHPrepare.cpp
  llvm/lib/Target/SPIRV/SPIRVPrepareFunctions.cpp
  llvm/lib/Transforms/Instrumentation/ThreadSanitizer.cpp
  llvm/test/Instrumentation/ThreadSanitizer/atomic.ll

Index: llvm/test/Instrumentation/ThreadSanitizer/atomic.ll
===
--- llvm/test/Instrumentation/ThreadSanitizer/atomic.ll
+++ llvm/test/Instrumentation/ThreadSanitizer/atomic.ll
@@ -1645,7 +1645,7 @@
 ; CHECK-NEXT: {{.*}} = call i64 @__tsan_atomic64_compare_exchange_val(ptr {{.*}}, i64 {{.*}}, i64 {{.*}}, i32 5, i32 5), !dbg
 ; CHECK-NEXT: {{.*}} = icmp eq i64
 ; CHECK-NEXT: {{.*}} = inttoptr i64 {{.*}} to ptr
-; CHECK-NEXT: {{.*}} = insertvalue { ptr, i1 } undef, ptr {{.*}}, 0
+; CHECK-NEXT: {{.*}} = insertvalue { ptr, i1 } poison, ptr {{.*}}, 0
 ; CHECK-NEXT: {{.*}} = insertvalue { ptr, i1 } {{.*}}, i1 {{.*}}, 1
 
 define i128 @atomic128_load_unordered(ptr %a) nounwind uwtable {
Index: llvm/lib/Transforms/Instrumentation/ThreadSanitizer.cpp
===
--- llvm/lib/Transforms/Instrumentation/ThreadSanitizer.cpp
+++ llvm/lib/Transforms/Instrumentation/ThreadSanitizer.cpp
@@ -801,7 +801,7 @@
 }
 
 Value *Res =
-  IRB.CreateInsertValue(UndefValue::get(CASI->getType()), OldVal, 0);
+  IRB.CreateInsertValue(PoisonValue::get(CASI->getType()), OldVal, 0);
 Res = IRB.CreateInsertValue(Res, Success, 1);
 
 I->replaceAllUsesWith(Res);
Index: llvm/lib/Target/SPIRV/SPIRVPrepareFunctions.cpp
===
--- llvm/lib/Target/SPIRV/SPIRVPrepareFunctions.cpp
+++ llvm/lib/Target/SPIRV/SPIRVPrepareFunctions.cpp
@@ -283,7 +283,7 @@
   // umul.with.overflow intrinsic return a structure, where the first element
   // is the multiplication result, and the second is an overflow bit.
   Type *StructTy = UMulFunc->getReturnType();
-  Value *Agg = IRB.CreateInsertValue(UndefValue::get(StructTy), Mul, {0});
+  Value *Agg = IRB.CreateInsertValue(PoisonValue::get(StructTy), Mul, {0});
   Value *Res = IRB.CreateInsertValue(Agg, Overflow, {1});
   IRB.CreateRet(Res);
 }
Index: llvm/lib/CodeGen/SjLjEHPrepare.cpp
===
--- llvm/lib/CodeGen/SjLjEHPrepare.cpp
+++ llvm/lib/CodeGen/SjLjEHPrepare.cpp
@@ -164,7 +164,7 @@
   // There are still some uses of LPI. Construct an aggregate with the exception
   // values and replace the LPI with that aggregate.
   Type *LPadType = LPI->getType();
-  Value *LPadVal = UndefValue::get(LPadType);
+  Value *LPadVal = PoisonValue::get(LPadType);
   auto *SelI = cast(SelVal);
   IRBuilder<> Builder(SelI->getParent(), std::next(SelI->getIterator()));
   LPadVal = Builder.CreateInsertValue(LPadVal, ExnVal, 0, "lpad.val");
Index: llvm/lib/Analysis/ValueTracking.cpp
===
--- llvm/lib/Analysis/ValueTracking.cpp
+++ llvm/lib/Analysis/ValueTracking.cpp
@@ -4107,7 +4107,7 @@
   assert(InsertBefore && "Must have someplace to insert!");
   Type *IndexedType = ExtractValueInst::getIndexedType(From->getType(),
  idx_range);
-  Value *To = UndefValue::get(IndexedType);
+  Value *To = PoisonValue::get(IndexedType);
   SmallVector Idxs(idx_range.begin(), idx_range.end());
   unsigned IdxSkip = Idxs.size();
 
Index: clang/test/OpenMP/threadprivate_codegen.cpp
==

[PATCH] D139745: [Clang]Use poison instead of undef where its used as placeholder[NFC]

2022-12-11 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito updated this revision to Diff 481908.
ManuelJBrito retitled this revision from "Use poison instead of undef where its 
used as placeholder[NFC]" to "[Clang]Use poison instead of undef where its used 
as placeholder[NFC]".
ManuelJBrito added a comment.

Splitting the diff


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D139745/new/

https://reviews.llvm.org/D139745

Files:
  clang/lib/CodeGen/CGException.cpp
  clang/lib/CodeGen/CGExprScalar.cpp
  clang/test/CodeGen/builtinshufflevector2.c
  clang/test/CodeGenCXX/nrvo.cpp
  clang/test/CodeGenCXX/pr40771-ctad-with-lambda-copy-capture.cpp
  clang/test/CodeGenCXX/vla-consruct.cpp
  clang/test/CodeGenCoroutines/coro-eh-cleanup-exp-namespace.cpp
  clang/test/CodeGenCoroutines/coro-eh-cleanup.cpp
  clang/test/OpenMP/distribute_parallel_for_num_threads_codegen.cpp
  clang/test/OpenMP/distribute_parallel_for_simd_num_threads_codegen.cpp
  clang/test/OpenMP/single_codegen.cpp
  clang/test/OpenMP/teams_distribute_parallel_for_num_threads_codegen.cpp
  clang/test/OpenMP/teams_distribute_parallel_for_simd_num_threads_codegen.cpp
  clang/test/OpenMP/threadprivate_codegen.cpp

Index: clang/test/OpenMP/threadprivate_codegen.cpp
===
--- clang/test/OpenMP/threadprivate_codegen.cpp
+++ clang/test/OpenMP/threadprivate_codegen.cpp
@@ -1123,7 +1123,7 @@
 // CHECK1:   eh.resume:
 // CHECK1-NEXT:[[EXN:%.*]] = load ptr, ptr [[EXN_SLOT]], align 8
 // CHECK1-NEXT:[[SEL:%.*]] = load i32, ptr [[EHSELECTOR_SLOT]], align 4
-// CHECK1-NEXT:[[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } undef, ptr [[EXN]], 0
+// CHECK1-NEXT:[[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } poison, ptr [[EXN]], 0
 // CHECK1-NEXT:[[LPAD_VAL28:%.*]] = insertvalue { ptr, i32 } [[LPAD_VAL]], i32 [[SEL]], 1
 // CHECK1-NEXT:resume { ptr, i32 } [[LPAD_VAL28]]
 //
@@ -1331,7 +1331,7 @@
 // CHECK1:   eh.resume:
 // CHECK1-NEXT:[[EXN:%.*]] = load ptr, ptr [[EXN_SLOT]], align 8
 // CHECK1-NEXT:[[SEL:%.*]] = load i32, ptr [[EHSELECTOR_SLOT]], align 4
-// CHECK1-NEXT:[[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } undef, ptr [[EXN]], 0
+// CHECK1-NEXT:[[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } poison, ptr [[EXN]], 0
 // CHECK1-NEXT:[[LPAD_VAL22:%.*]] = insertvalue { ptr, i32 } [[LPAD_VAL]], i32 [[SEL]], 1
 // CHECK1-NEXT:resume { ptr, i32 } [[LPAD_VAL22]]
 //
@@ -1446,7 +1446,7 @@
 // CHECK1:   eh.resume:
 // CHECK1-NEXT:[[EXN:%.*]] = load ptr, ptr [[EXN_SLOT]], align 8
 // CHECK1-NEXT:[[SEL:%.*]] = load i32, ptr [[EHSELECTOR_SLOT]], align 4
-// CHECK1-NEXT:[[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } undef, ptr [[EXN]], 0
+// CHECK1-NEXT:[[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } poison, ptr [[EXN]], 0
 // CHECK1-NEXT:[[LPAD_VAL15:%.*]] = insertvalue { ptr, i32 } [[LPAD_VAL]], i32 [[SEL]], 1
 // CHECK1-NEXT:resume { ptr, i32 } [[LPAD_VAL15]]
 //
@@ -1854,7 +1854,7 @@
 // CHECK2:   eh.resume:
 // CHECK2-NEXT:[[EXN:%.*]] = load ptr, ptr [[EXN_SLOT]], align 8
 // CHECK2-NEXT:[[SEL:%.*]] = load i32, ptr [[EHSELECTOR_SLOT]], align 4
-// CHECK2-NEXT:[[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } undef, ptr [[EXN]], 0
+// CHECK2-NEXT:[[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } poison, ptr [[EXN]], 0
 // CHECK2-NEXT:[[LPAD_VAL22:%.*]] = insertvalue { ptr, i32 } [[LPAD_VAL]], i32 [[SEL]], 1
 // CHECK2-NEXT:resume { ptr, i32 } [[LPAD_VAL22]]
 //
@@ -1975,7 +1975,7 @@
 // CHECK2:   eh.resume:
 // CHECK2-NEXT:[[EXN:%.*]] = load ptr, ptr [[EXN_SLOT]], align 8
 // CHECK2-NEXT:[[SEL:%.*]] = load i32, ptr [[EHSELECTOR_SLOT]], align 4
-// CHECK2-NEXT:[[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } undef, ptr [[EXN]], 0
+// CHECK2-NEXT:[[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } poison, ptr [[EXN]], 0
 // CHECK2-NEXT:[[LPAD_VAL28:%.*]] = insertvalue { ptr, i32 } [[LPAD_VAL]], i32 [[SEL]], 1
 // CHECK2-NEXT:resume { ptr, i32 } [[LPAD_VAL28]]
 //
@@ -2100,7 +2100,7 @@
 // CHECK2:   eh.resume:
 // CHECK2-NEXT:[[EXN:%.*]] = load ptr, ptr [[EXN_SLOT]], align 8
 // CHECK2-NEXT:[[SEL:%.*]] = load i32, ptr [[EHSELECTOR_SLOT]], align 4
-// CHECK2-NEXT:[[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } undef, ptr [[EXN]], 0
+// CHECK2-NEXT:[[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } poison, ptr [[EXN]], 0
 // CHECK2-NEXT:[[LPAD_VAL15:%.*]] = insertvalue { ptr, i32 } [[LPAD_VAL]], i32 [[SEL]], 1
 // CHECK2-NEXT:resume { ptr, i32 } [[LPAD_VAL15]]
 //
@@ -2529,7 +2529,7 @@
 // SIMD1:   eh.resume:
 // SIMD1-NEXT:[[EXN:%.*]] = load ptr, ptr [[EXN_SLOT]], align 8
 // SIMD1-NEXT:[[SEL:%.*]] = load i32, ptr [[EHSELECTOR_SLOT]], align 4
-// SIMD1-NEXT:[[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } undef, ptr [[EXN]], 0
+// SIMD1-NEXT:[[LPAD_VAL:%.*]] = insertvalue { ptr, i32 } poison, ptr [[EXN]], 0
 // SIMD1-NEXT:[[LPAD_VAL22:%.*]] = insertvalue { ptr, i32 } [[LPAD_VAL]], i

[PATCH] D140090: [Clang][CodeGen] Use poison instead of undef in CodeGen for ARM Builtins [NFC]

2022-12-15 Thread Manuel Brito via Phabricator via cfe-commits 
ManuelJBrito created this revision.
ManuelJBrito added reviewers: peterwaller-arm, paulwalker-arm, david-arm.
Herald added subscribers: ctetreau, arphaman, kristof.beyls.
Herald added a project: All.
ManuelJBrito requested review of this revision.
Herald added subscribers: cfe-commits, alextsao1999.
Herald added a project: clang.

In light of the efforts to remove undef, this patch changes the following 
instances of undef to poison in the CodeGen for ARM Builtins:

- `EmitCommonNeonSISDBuiltinExpr` : uses overloaded intrinsics to implement 
SISD expressions, ops go in first entry, the rest of the vector is irrelevant.

- `EmitAArch64SVEBuiltinExpr` : cast from fixed vector to scalable vector with 
the same size therefore undef/poison is hidden.


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D140090

Files:
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/test/CodeGen/aarch64-neon-intrinsics.c
  clang/test/CodeGen/aarch64-neon-scalar-x-indexed-elem.c
  clang/test/CodeGen/aarch64-neon-vsqadd-float-conversion.c
  clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq-bfloat.c
  clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq.c
  clang/test/CodeGen/aarch64-v8.1a-neon-intrinsics.c
  
clang/test/CodeGen/aarch64_neon_sve_bridge_intrinsics/acle_neon_sve_bridge_dup_neonq.c

Index: clang/test/CodeGen/aarch64_neon_sve_bridge_intrinsics/acle_neon_sve_bridge_dup_neonq.c
===
--- clang/test/CodeGen/aarch64_neon_sve_bridge_intrinsics/acle_neon_sve_bridge_dup_neonq.c
+++ clang/test/CodeGen/aarch64_neon_sve_bridge_intrinsics/acle_neon_sve_bridge_dup_neonq.c
@@ -16,13 +16,13 @@
 
 // CHECK-LABEL: @test_svdup_neonq_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv16i8.v16i8( undef, <16 x i8> [[N:%.*]], i64 0)
+// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv16i8.v16i8( poison, <16 x i8> [[N:%.*]], i64 0)
 // CHECK-NEXT:[[TMP1:%.*]] = tail call  @llvm.aarch64.sve.dupq.lane.nxv16i8( [[TMP0]], i64 0)
 // CHECK-NEXT:ret  [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z19test_svdup_neonq_s811__Int8x16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv16i8.v16i8( undef, <16 x i8> [[N:%.*]], i64 0)
+// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv16i8.v16i8( poison, <16 x i8> [[N:%.*]], i64 0)
 // CPP-CHECK-NEXT:[[TMP1:%.*]] = tail call  @llvm.aarch64.sve.dupq.lane.nxv16i8( [[TMP0]], i64 0)
 // CPP-CHECK-NEXT:ret  [[TMP1]]
 //
@@ -32,13 +32,13 @@
 
 // CHECK-LABEL: @test_svdup_neonq_s16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv8i16.v8i16( undef, <8 x i16> [[N:%.*]], i64 0)
+// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv8i16.v8i16( poison, <8 x i16> [[N:%.*]], i64 0)
 // CHECK-NEXT:[[TMP1:%.*]] = tail call  @llvm.aarch64.sve.dupq.lane.nxv8i16( [[TMP0]], i64 0)
 // CHECK-NEXT:ret  [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z20test_svdup_neonq_s1611__Int16x8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv8i16.v8i16( undef, <8 x i16> [[N:%.*]], i64 0)
+// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv8i16.v8i16( poison, <8 x i16> [[N:%.*]], i64 0)
 // CPP-CHECK-NEXT:[[TMP1:%.*]] = tail call  @llvm.aarch64.sve.dupq.lane.nxv8i16( [[TMP0]], i64 0)
 // CPP-CHECK-NEXT:ret  [[TMP1]]
 //
@@ -48,13 +48,13 @@
 
 // CHECK-LABEL: @test_svdup_neonq_s32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv4i32.v4i32( undef, <4 x i32> [[N:%.*]], i64 0)
+// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv4i32.v4i32( poison, <4 x i32> [[N:%.*]], i64 0)
 // CHECK-NEXT:[[TMP1:%.*]] = tail call  @llvm.aarch64.sve.dupq.lane.nxv4i32( [[TMP0]], i64 0)
 // CHECK-NEXT:ret  [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z20test_svdup_neonq_s3211__Int32x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv4i32.v4i32( undef, <4 x i32> [[N:%.*]], i64 0)
+// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv4i32.v4i32( poison, <4 x i32> [[N:%.*]], i64 0)
 // CPP-CHECK-NEXT:[[TMP1:%.*]] = tail call  @llvm.aarch64.sve.dupq.lane.nxv4i32( [[TMP0]], i64 0)
 // CPP-CHECK-NEXT:ret  [[TMP1]]
 //
@@ -64,13 +64,13 @@
 
 // CHECK-LABEL: @test_svdup_neonq_s64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv2i64.v2i64( undef, <2 x i64> [[N:%.*]], i64 0)
+// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv2i64.v2i64( poison, <2 x i64> [[N:%.*]], i64 0)
 // CHECK-NEXT:[[TMP1:%.*]] = tail call  @llvm.aarch64.sve.dupq.lane.nxv2i64( [[TMP0]], i64 0)
 // CHECK-NEXT:ret  [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z20test_svdup_neonq_s6411__Int64x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:[[TMP0:%.

[PATCH] D140090: [Clang][CodeGen] Use poison instead of undef in CodeGen for ARM Builtins [NFC]

2022-12-15 Thread Manuel Brito via Phabricator via cfe-commits 
This revision was automatically updated to reflect the committed changes.
Closed by commit rG84b4ff24e94b: [Clang][CodeGen] Use poison instead of undef 
in CodeGen for ARM Builtins [NFC] (authored by ManuelJBrito).

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D140090/new/

https://reviews.llvm.org/D140090

Files:
  clang/lib/CodeGen/CGBuiltin.cpp
  clang/test/CodeGen/aarch64-neon-intrinsics.c
  clang/test/CodeGen/aarch64-neon-scalar-x-indexed-elem.c
  clang/test/CodeGen/aarch64-neon-vsqadd-float-conversion.c
  clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq-bfloat.c
  clang/test/CodeGen/aarch64-sve-intrinsics/acle_sve_dupq.c
  clang/test/CodeGen/aarch64-v8.1a-neon-intrinsics.c
  
clang/test/CodeGen/aarch64_neon_sve_bridge_intrinsics/acle_neon_sve_bridge_dup_neonq.c

Index: clang/test/CodeGen/aarch64_neon_sve_bridge_intrinsics/acle_neon_sve_bridge_dup_neonq.c
===
--- clang/test/CodeGen/aarch64_neon_sve_bridge_intrinsics/acle_neon_sve_bridge_dup_neonq.c
+++ clang/test/CodeGen/aarch64_neon_sve_bridge_intrinsics/acle_neon_sve_bridge_dup_neonq.c
@@ -16,13 +16,13 @@
 
 // CHECK-LABEL: @test_svdup_neonq_s8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv16i8.v16i8( undef, <16 x i8> [[N:%.*]], i64 0)
+// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv16i8.v16i8( poison, <16 x i8> [[N:%.*]], i64 0)
 // CHECK-NEXT:[[TMP1:%.*]] = tail call  @llvm.aarch64.sve.dupq.lane.nxv16i8( [[TMP0]], i64 0)
 // CHECK-NEXT:ret  [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z19test_svdup_neonq_s811__Int8x16_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv16i8.v16i8( undef, <16 x i8> [[N:%.*]], i64 0)
+// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv16i8.v16i8( poison, <16 x i8> [[N:%.*]], i64 0)
 // CPP-CHECK-NEXT:[[TMP1:%.*]] = tail call  @llvm.aarch64.sve.dupq.lane.nxv16i8( [[TMP0]], i64 0)
 // CPP-CHECK-NEXT:ret  [[TMP1]]
 //
@@ -32,13 +32,13 @@
 
 // CHECK-LABEL: @test_svdup_neonq_s16(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv8i16.v8i16( undef, <8 x i16> [[N:%.*]], i64 0)
+// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv8i16.v8i16( poison, <8 x i16> [[N:%.*]], i64 0)
 // CHECK-NEXT:[[TMP1:%.*]] = tail call  @llvm.aarch64.sve.dupq.lane.nxv8i16( [[TMP0]], i64 0)
 // CHECK-NEXT:ret  [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z20test_svdup_neonq_s1611__Int16x8_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv8i16.v8i16( undef, <8 x i16> [[N:%.*]], i64 0)
+// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv8i16.v8i16( poison, <8 x i16> [[N:%.*]], i64 0)
 // CPP-CHECK-NEXT:[[TMP1:%.*]] = tail call  @llvm.aarch64.sve.dupq.lane.nxv8i16( [[TMP0]], i64 0)
 // CPP-CHECK-NEXT:ret  [[TMP1]]
 //
@@ -48,13 +48,13 @@
 
 // CHECK-LABEL: @test_svdup_neonq_s32(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv4i32.v4i32( undef, <4 x i32> [[N:%.*]], i64 0)
+// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv4i32.v4i32( poison, <4 x i32> [[N:%.*]], i64 0)
 // CHECK-NEXT:[[TMP1:%.*]] = tail call  @llvm.aarch64.sve.dupq.lane.nxv4i32( [[TMP0]], i64 0)
 // CHECK-NEXT:ret  [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z20test_svdup_neonq_s3211__Int32x4_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv4i32.v4i32( undef, <4 x i32> [[N:%.*]], i64 0)
+// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv4i32.v4i32( poison, <4 x i32> [[N:%.*]], i64 0)
 // CPP-CHECK-NEXT:[[TMP1:%.*]] = tail call  @llvm.aarch64.sve.dupq.lane.nxv4i32( [[TMP0]], i64 0)
 // CPP-CHECK-NEXT:ret  [[TMP1]]
 //
@@ -64,13 +64,13 @@
 
 // CHECK-LABEL: @test_svdup_neonq_s64(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv2i64.v2i64( undef, <2 x i64> [[N:%.*]], i64 0)
+// CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv2i64.v2i64( poison, <2 x i64> [[N:%.*]], i64 0)
 // CHECK-NEXT:[[TMP1:%.*]] = tail call  @llvm.aarch64.sve.dupq.lane.nxv2i64( [[TMP0]], i64 0)
 // CHECK-NEXT:ret  [[TMP1]]
 //
 // CPP-CHECK-LABEL: @_Z20test_svdup_neonq_s6411__Int64x2_t(
 // CPP-CHECK-NEXT:  entry:
-// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv2i64.v2i64( undef, <2 x i64> [[N:%.*]], i64 0)
+// CPP-CHECK-NEXT:[[TMP0:%.*]] = tail call  @llvm.vector.insert.nxv2i64.v2i64( poison, <2 x i64> [[N:%.*]], i64 0)
 // CPP-CHECK-NEXT:[[TMP1:%.*]] = tail call  @llvm.aarch64.sve.dupq.lane.nxv2i64( [[TMP0]], i64 0)
 // CPP-CHECK-NEXT:ret  [[TMP1]]
 //
@@ -80,13 +80,13 @@
 
 // CHECK-LABEL: @test_svdup_neonq_u8(
 // CHECK-NEXT:  entry:
-// CHECK-NEXT:[[TMP0:%.*]] = tail