[HLSL][DirectX] Emit convergence control tokens when targeting DirectX (#188792)
This pr allows codegen to generate convergence control tokens. This
allows for a more accurate description of convergence behaviour to
prevent (or allow) invalid control flow graph transforms. As noted, the
use of convergence control tokens is the ideal norm and this follows
that by enabling it for `DirectX`.
This was done now under the precedent of preventing a convergent exit
condition of a loop from being illegally moved across control flow. Test
cases for this are explicitly added.
Please see the individual commits for logically similar chunks.
Unfortunately, it is tricky to stage this in smaller individual commits.
Resolves https://github.com/llvm/llvm-project/issues/180621.
https://github.com/llvm/llvm-project/pull/188537 is a pre-requisite of
this passing HLSL offload suite tests.
Assisted by: Github Copilot
diff --git a/clang/lib/CodeGen/CGExprAgg.cpp b/clang/lib/CodeGen/CGExprAgg.cpp
index 3a42917..d3dc101 100644
--- a/clang/lib/CodeGen/CGExprAgg.cpp
+++ b/clang/lib/CodeGen/CGExprAgg.cpp
@@ -715,6 +715,9 @@
Builder.CreatePHI(element->getType(), 2, "arrayinit.cur");
currentElement->addIncoming(element, entryBB);
+ if (CGF.CGM.shouldEmitConvergenceTokens())
+ CGF.ConvergenceTokenStack.push_back(CGF.emitConvergenceLoopToken(bodyBB));
+
// Emit the actual filler expression.
{
// C++1z [class.temporary]p5:
@@ -746,6 +749,9 @@
Builder.CreateCondBr(done, endBB, bodyBB);
currentElement->addIncoming(nextElement, Builder.GetInsertBlock());
+ if (CGF.CGM.shouldEmitConvergenceTokens())
+ CGF.ConvergenceTokenStack.pop_back();
+
CGF.EmitBlock(endBB);
}
}
@@ -1987,6 +1993,9 @@
llvm::Value *element =
Builder.CreateInBoundsGEP(llvmElementType, begin, index);
+ if (CGF.CGM.shouldEmitConvergenceTokens())
+ CGF.ConvergenceTokenStack.push_back(CGF.emitConvergenceLoopToken(bodyBB));
+
// Prepare for a cleanup.
QualType::DestructionKind dtorKind = elementType.isDestructedType();
EHScopeStack::stable_iterator cleanup;
@@ -2034,6 +2043,9 @@
llvm::BasicBlock *endBB = CGF.createBasicBlock("arrayinit.end");
Builder.CreateCondBr(done, endBB, bodyBB);
+ if (CGF.CGM.shouldEmitConvergenceTokens())
+ CGF.ConvergenceTokenStack.pop_back();
+
CGF.EmitBlock(endBB);
// Leave the partial-array cleanup if we entered one.
diff --git a/clang/lib/CodeGen/CGHLSLBuiltins.cpp b/clang/lib/CodeGen/CGHLSLBuiltins.cpp
index f510195..d0d8eed2 100644
--- a/clang/lib/CodeGen/CGHLSLBuiltins.cpp
+++ b/clang/lib/CodeGen/CGHLSLBuiltins.cpp
@@ -574,9 +574,13 @@
Value *IndexOp = EmitScalarExpr(E->getArg(1));
llvm::Type *RetTy = ConvertType(E->getType());
- return Builder.CreateIntrinsic(
- RetTy, CGM.getHLSLRuntime().getCreateResourceGetPointerIntrinsic(),
- ArrayRef<Value *>{HandleOp, IndexOp});
+ llvm::Function *IntrFn = llvm::Intrinsic::getOrInsertDeclaration(
+ &CGM.getModule(),
+ CGM.getHLSLRuntime().getCreateResourceGetPointerIntrinsic(),
+ {RetTy, HandleOp->getType(), IndexOp->getType()});
+ llvm::CallInst *CI = EmitRuntimeCall(IntrFn, {HandleOp, IndexOp});
+ CI->setCallingConv(IntrFn->getCallingConv());
+ return CI;
}
case Builtin::BI__builtin_hlsl_resource_sample: {
Value *HandleOp = EmitScalarExpr(E->getArg(0));
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.cpp b/clang/lib/CodeGen/CGHLSLRuntime.cpp
index 1e25172..99b3a14 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.cpp
+++ b/clang/lib/CodeGen/CGHLSLRuntime.cpp
@@ -657,8 +657,16 @@
llvm::PoisonValue::get(B.getInt32Ty())};
llvm::Intrinsic::ID IntrinsicID = llvm::Intrinsic::dx_load_input;
- llvm::Value *Value = B.CreateIntrinsic(/*ReturnType=*/Type, IntrinsicID, Args,
- nullptr, VariableName);
+
+ SmallVector<OperandBundleDef, 1> OB;
+ if (auto *Token = getConvergenceToken(*B.GetInsertBlock())) {
+ llvm::Value *bundleArgs[] = {Token};
+ OB.emplace_back("convergencectrl", bundleArgs);
+ }
+
+ llvm::Function *IntrFn = llvm::Intrinsic::getOrInsertDeclaration(
+ B.GetInsertBlock()->getModule(), IntrinsicID, {Type});
+ llvm::Value *Value = B.CreateCall(IntrFn, Args, OB, VariableName);
return Value;
}
@@ -676,7 +684,16 @@
Source};
llvm::Intrinsic::ID IntrinsicID = llvm::Intrinsic::dx_store_output;
- B.CreateIntrinsic(/*ReturnType=*/CGM.VoidTy, IntrinsicID, Args, nullptr);
+
+ SmallVector<OperandBundleDef, 1> OB;
+ if (auto *Token = getConvergenceToken(*B.GetInsertBlock())) {
+ llvm::Value *bundleArgs[] = {Token};
+ OB.emplace_back("convergencectrl", bundleArgs);
+ }
+
+ llvm::Function *IntrFn = llvm::Intrinsic::getOrInsertDeclaration(
+ B.GetInsertBlock()->getModule(), IntrinsicID, {Source->getType()});
+ B.CreateCall(IntrFn, Args, OB);
}
llvm::Value *CGHLSLRuntime::emitUserSemanticLoad(
diff --git a/clang/lib/CodeGen/CodeGenFunction.h b/clang/lib/CodeGen/CodeGenFunction.h
index 29b87a0..d1752b8 100644
--- a/clang/lib/CodeGen/CodeGenFunction.h
+++ b/clang/lib/CodeGen/CodeGenFunction.h
@@ -5437,11 +5437,11 @@
void maybeAttachRangeForLoad(llvm::LoadInst *Load, QualType Ty,
SourceLocation Loc);
-private:
// Emits a convergence_loop instruction for the given |BB|, with |ParentToken|
// as it's parent convergence instr.
llvm::ConvergenceControlInst *emitConvergenceLoopToken(llvm::BasicBlock *BB);
+private:
// Adds a convergence_ctrl token with |ParentToken| as parent convergence
// instr to the call |Input|.
llvm::CallBase *addConvergenceControlToken(llvm::CallBase *Input);
diff --git a/clang/lib/CodeGen/CodeGenModule.h b/clang/lib/CodeGen/CodeGenModule.h
index d62707a..dc29691 100644
--- a/clang/lib/CodeGen/CodeGenModule.h
+++ b/clang/lib/CodeGen/CodeGenModule.h
@@ -1815,7 +1815,7 @@
bool shouldEmitConvergenceTokens() const {
// TODO: this should probably become unconditional once the controlled
// convergence becomes the norm.
- return getTriple().isSPIRVLogical();
+ return getTriple().isSPIRVLogical() || getTriple().isDXIL();
}
void addUndefinedGlobalForTailCall(
diff --git a/clang/test/CodeGenDirectX/Builtins/dot2add.c b/clang/test/CodeGenDirectX/Builtins/dot2add.c
index 4275a28..bc50739 100644
--- a/clang/test/CodeGenDirectX/Builtins/dot2add.c
+++ b/clang/test/CodeGenDirectX/Builtins/dot2add.c
@@ -8,6 +8,7 @@
// CHECK-LABEL: define float @test_dot2add(
// CHECK-SAME: <2 x half> noundef [[X:%.*]], <2 x half> noundef [[Y:%.*]], float noundef [[Z:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[X_ADDR:%.*]] = alloca <2 x half>, align 2
// CHECK-NEXT: [[Y_ADDR:%.*]] = alloca <2 x half>, align 2
// CHECK-NEXT: [[Z_ADDR:%.*]] = alloca float, align 4
diff --git a/clang/test/CodeGenHLSL/BasicFeatures/ArrayReturn.hlsl b/clang/test/CodeGenHLSL/BasicFeatures/ArrayReturn.hlsl
index 832c4ac..b4235eed 100644
--- a/clang/test/CodeGenHLSL/BasicFeatures/ArrayReturn.hlsl
+++ b/clang/test/CodeGenHLSL/BasicFeatures/ArrayReturn.hlsl
@@ -3,12 +3,14 @@
typedef int Foo[2];
// CHECK-LABEL: define void {{.*}}boop{{.*}}(ptr dead_on_unwind noalias writable sret([2 x i32]) align 4 %agg.result)
+// CHECK: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK: [[G:%.*]] = alloca [2 x i32], align 4
// CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 4 [[G]], ptr align 4 {{.*}}, i32 8, i1 false)
// CHECK-NEXT: [[AIB:%.*]] = getelementptr inbounds [2 x i32], ptr %agg.result, i32 0, i32 0
// CHECK-NEXT: br label %arrayinit.body
// CHECK: arrayinit.body:
// CHECK-NEXT: [[AII:%.*]] = phi i32 [ 0, %entry ], [ %arrayinit.next, %arrayinit.body ]
+// CHECK-NEXT: %[[#CV_LOOP:]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token %[[#C_ENTRY]]) ]
// CHECK-NEXT: [[X:%.*]] = getelementptr inbounds i32, ptr [[AIB]], i32 [[AII]]
// CHECK-NEXT: [[AI:%.*]] = getelementptr inbounds nuw [2 x i32], ptr [[G]], i32 0, i32 [[AII]]
// CHECK-NEXT: [[Y:%.*]] = load i32, ptr [[AI]], align 4
diff --git a/clang/test/CodeGenHLSL/BasicFeatures/InitLists.hlsl b/clang/test/CodeGenHLSL/BasicFeatures/InitLists.hlsl
index 9c42da8..40f32c2 100644
--- a/clang/test/CodeGenHLSL/BasicFeatures/InitLists.hlsl
+++ b/clang/test/CodeGenHLSL/BasicFeatures/InitLists.hlsl
@@ -66,6 +66,7 @@
// CHECK-LABEL: define hidden void @_Z5case1v(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_TWOFLOATS:%.*]]) align 1 [[AGG_RESULT:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 1 [[AGG_RESULT]], ptr align 1 @__const._Z5case1v.TF1, i32 8, i1 false)
// CHECK-NEXT: ret void
//
@@ -78,6 +79,7 @@
// CHECK-LABEL: define hidden void @_Z5case2v(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_TWOFLOATS:%.*]]) align 1 [[AGG_RESULT:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 1 [[AGG_RESULT]], ptr align 1 @__const._Z5case2v.TF2, i32 8, i1 false)
// CHECK-NEXT: ret void
//
@@ -90,6 +92,7 @@
// CHECK-LABEL: define hidden void @_Z5case3i(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_TWOFLOATS:%.*]]) align 1 [[AGG_RESULT:%.*]], i32 noundef [[VAL:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[VAL_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store i32 [[VAL]], ptr [[VAL_ADDR]], align 4
// CHECK-NEXT: [[X:%.*]] = getelementptr inbounds nuw [[STRUCT_TWOFLOATS]], ptr [[AGG_RESULT]], i32 0, i32 0
@@ -110,6 +113,7 @@
// CHECK-LABEL: define hidden void @_Z5case4Dv2_i(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_TWOFLOATS:%.*]]) align 1 [[AGG_RESULT:%.*]], <2 x i32> noundef [[TWOVALS:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[TWOVALS_ADDR:%.*]] = alloca <2 x i32>, align 4
// CHECK-NEXT: store <2 x i32> [[TWOVALS]], ptr [[TWOVALS_ADDR]], align 4
// CHECK-NEXT: [[X:%.*]] = getelementptr inbounds nuw [[STRUCT_TWOFLOATS]], ptr [[AGG_RESULT]], i32 0, i32 0
@@ -133,6 +137,7 @@
// CHECK-LABEL: define hidden void @_Z5case5Dv2_i(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_TWOINTS:%.*]]) align 1 [[AGG_RESULT:%.*]], <2 x i32> noundef [[TWOVALS:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[TWOVALS_ADDR:%.*]] = alloca <2 x i32>, align 4
// CHECK-NEXT: store <2 x i32> [[TWOVALS]], ptr [[TWOVALS_ADDR]], align 4
// CHECK-NEXT: [[Z:%.*]] = getelementptr inbounds nuw [[STRUCT_TWOINTS]], ptr [[AGG_RESULT]], i32 0, i32 0
@@ -155,6 +160,7 @@
// CHECK-LABEL: define hidden void @_Z5case69TwoFloats(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_TWOINTS:%.*]]) align 1 [[AGG_RESULT:%.*]], ptr noundef byval([[STRUCT_TWOFLOATS:%.*]]) align 1 [[TF4:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[Z:%.*]] = getelementptr inbounds nuw [[STRUCT_TWOINTS]], ptr [[AGG_RESULT]], i32 0, i32 0
// CHECK-NEXT: [[X:%.*]] = getelementptr inbounds nuw [[STRUCT_TWOFLOATS]], ptr [[TF4]], i32 0, i32 0
// CHECK-NEXT: [[TMP0:%.*]] = load float, ptr [[X]], align 1
@@ -177,6 +183,7 @@
// CHECK-LABEL: define hidden void @_Z5case77TwoIntsS_i9TwoFloatsS0_S0_S0_(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_DOGGO:%.*]]) align 1 [[AGG_RESULT:%.*]], ptr noundef byval([[STRUCT_TWOINTS:%.*]]) align 1 [[TI1:%.*]], ptr noundef byval([[STRUCT_TWOINTS]]) align 1 [[TI2:%.*]], i32 noundef [[VAL:%.*]], ptr noundef byval([[STRUCT_TWOFLOATS:%.*]]) align 1 [[TF1:%.*]], ptr noundef byval([[STRUCT_TWOFLOATS]]) align 1 [[TF2:%.*]], ptr noundef byval([[STRUCT_TWOFLOATS]]) align 1 [[TF3:%.*]], ptr noundef byval([[STRUCT_TWOFLOATS]]) align 1 [[TF4:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[VAL_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store i32 [[VAL]], ptr [[VAL_ADDR]], align 4
// CHECK-NEXT: [[LEGSTATE:%.*]] = getelementptr inbounds nuw [[STRUCT_DOGGO]], ptr [[AGG_RESULT]], i32 0, i32 0
@@ -241,6 +248,7 @@
// CHECK-LABEL: define hidden void @_Z5case85Doggo(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_ANIMALBITS:%.*]]) align 1 [[AGG_RESULT:%.*]], ptr noundef byval([[STRUCT_DOGGO:%.*]]) align 1 [[D1:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[LEGS:%.*]] = getelementptr inbounds nuw [[STRUCT_ANIMALBITS]], ptr [[AGG_RESULT]], i32 0, i32 0
// CHECK-NEXT: [[LEGSTATE:%.*]] = getelementptr inbounds nuw [[STRUCT_DOGGO]], ptr [[D1]], i32 0, i32 0
// CHECK-NEXT: [[TMP0:%.*]] = load <4 x i32>, ptr [[LEGSTATE]], align 1
@@ -327,6 +335,7 @@
// CHECK-LABEL: define hidden void @_Z5case95Doggo10AnimalBits(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_ZOO:%.*]]) align 1 [[AGG_RESULT:%.*]], ptr noundef byval([[STRUCT_DOGGO:%.*]]) align 1 [[D1:%.*]], ptr noundef byval([[STRUCT_ANIMALBITS:%.*]]) align 1 [[A1:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[DOGS:%.*]] = getelementptr inbounds nuw [[STRUCT_ZOO]], ptr [[AGG_RESULT]], i32 0, i32 0
// CHECK-NEXT: [[LEGSTATE:%.*]] = getelementptr inbounds nuw [[STRUCT_DOGGO]], ptr [[DOGS]], i32 0, i32 0
// CHECK-NEXT: [[LEGSTATE1:%.*]] = getelementptr inbounds nuw [[STRUCT_DOGGO]], ptr [[D1]], i32 0, i32 0
@@ -743,6 +752,7 @@
// CHECK-LABEL: define hidden void @_Z6case109TwoFloatsS_(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_FOURFLOATS:%.*]]) align 1 [[AGG_RESULT:%.*]], ptr noundef byval([[STRUCT_TWOFLOATS:%.*]]) align 1 [[TF1:%.*]], ptr noundef byval([[STRUCT_TWOFLOATS]]) align 1 [[TF2:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[X:%.*]] = getelementptr inbounds nuw [[STRUCT_TWOFLOATS]], ptr [[AGG_RESULT]], i32 0, i32 0
// CHECK-NEXT: [[X1:%.*]] = getelementptr inbounds nuw [[STRUCT_TWOFLOATS]], ptr [[TF1]], i32 0, i32 0
// CHECK-NEXT: [[TMP0:%.*]] = load float, ptr [[X1]], align 1
@@ -770,6 +780,7 @@
// CHECK-LABEL: define hidden void @_Z6case11f(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_FOURFLOATS:%.*]]) align 1 [[AGG_RESULT:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: [[REF_TMP:%.*]] = alloca <4 x float>, align 4
// CHECK-NEXT: [[REF_TMP1:%.*]] = alloca <4 x float>, align 4
@@ -819,6 +830,7 @@
// CHECK-LABEL: define hidden void @_Z6case12ii(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_SLICYBITS:%.*]]) align 1 [[AGG_RESULT:%.*]], i32 noundef [[I:%.*]], i32 noundef [[J:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: [[J_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store i32 [[I]], ptr [[I_ADDR]], align 4
@@ -841,6 +853,7 @@
// CHECK-LABEL: define hidden void @_Z6case137TwoInts(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_SLICYBITS:%.*]]) align 1 [[AGG_RESULT:%.*]], ptr noundef byval([[STRUCT_TWOINTS:%.*]]) align 1 [[TI:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[Z:%.*]] = getelementptr inbounds nuw [[STRUCT_TWOINTS]], ptr [[TI]], i32 0, i32 0
// CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr [[Z]], align 1
// CHECK-NEXT: [[TMP1:%.*]] = trunc i32 [[TMP0]] to i8
@@ -861,6 +874,7 @@
// CHECK-LABEL: define hidden void @_Z6case149SlicyBits(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_TWOINTS:%.*]]) align 1 [[AGG_RESULT:%.*]], ptr noundef byval([[STRUCT_SLICYBITS:%.*]]) align 1 [[SB:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[Z:%.*]] = getelementptr inbounds nuw [[STRUCT_TWOINTS]], ptr [[AGG_RESULT]], i32 0, i32 0
// CHECK-NEXT: [[BF_LOAD:%.*]] = load i8, ptr [[SB]], align 1
// CHECK-NEXT: [[BF_CAST:%.*]] = sext i8 [[BF_LOAD]] to i32
@@ -881,6 +895,7 @@
// CHECK-LABEL: define hidden void @_Z6case159SlicyBits(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_TWOFLOATS:%.*]]) align 1 [[AGG_RESULT:%.*]], ptr noundef byval([[STRUCT_SLICYBITS:%.*]]) align 1 [[SB:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[X:%.*]] = getelementptr inbounds nuw [[STRUCT_TWOFLOATS]], ptr [[AGG_RESULT]], i32 0, i32 0
// CHECK-NEXT: [[BF_LOAD:%.*]] = load i8, ptr [[SB]], align 1
// CHECK-NEXT: [[BF_CAST:%.*]] = sext i8 [[BF_LOAD]] to i32
@@ -904,6 +919,7 @@
// CHECK-LABEL: define hidden void @_Z7makeTwoRf(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_TWOFLOATS:%.*]]) align 1 [[AGG_RESULT:%.*]], ptr noalias noundef nonnull align 4 dereferenceable(4) [[X:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[X_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: store ptr [[X]], ptr [[X_ADDR]], align 4
// CHECK-NEXT: [[X1:%.*]] = getelementptr inbounds nuw [[STRUCT_TWOFLOATS]], ptr [[AGG_RESULT]], i32 0, i32 0
@@ -930,6 +946,7 @@
// CHECK-LABEL: define hidden void @_Z6case16v(
// CHECK-SAME: ptr dead_on_unwind noalias writable sret([[STRUCT_FOURFLOATS:%.*]]) align 1 [[AGG_RESULT:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[X:%.*]] = alloca float, align 4
// CHECK-NEXT: [[REF_TMP:%.*]] = alloca [[STRUCT_TWOFLOATS:%.*]], align 1
// CHECK-NEXT: [[TMP:%.*]] = alloca float, align 4
@@ -963,6 +980,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z12case17Helperi(
// CHECK-SAME: i32 noundef [[X:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[X_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store i32 [[X]], ptr [[X_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load i32, ptr [[X_ADDR]], align 4
@@ -976,6 +994,7 @@
// CHECK-LABEL: define hidden void @_Z6case17v(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[X:%.*]] = alloca <2 x i32>, align 4
// CHECK-NEXT: [[CALL:%.*]] = call noundef i32 @_Z12case17Helperi(i32 noundef 0) #[[ATTR2]]
// CHECK-NEXT: [[CALL1:%.*]] = call noundef i32 @_Z12case17Helperi(i32 noundef 1) #[[ATTR2]]
@@ -992,6 +1011,7 @@
// CHECK-LABEL: define hidden void @_Z6case18v(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[U:%.*]] = alloca [[STRUCT_UNNAMED:%.*]], align 1
// CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 1 [[U]], ptr align 1 @__const._Z6case18v.U, i32 5, i1 false)
// CHECK-NEXT: ret void
@@ -1004,6 +1024,7 @@
// CHECK-LABEL: define hidden void @_Z6case197Unnamed(
// CHECK-SAME: ptr noundef byval([[STRUCT_UNNAMED:%.*]]) align 1 [[U:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[TI:%.*]] = alloca [[STRUCT_TWOINTS:%.*]], align 1
// CHECK-NEXT: [[Z:%.*]] = getelementptr inbounds nuw [[STRUCT_TWOINTS]], ptr [[TI]], i32 0, i32 0
// CHECK-NEXT: [[A:%.*]] = getelementptr inbounds nuw [[STRUCT_UNNAMED]], ptr [[U]], i32 0, i32 0
@@ -1021,6 +1042,7 @@
// CHECK-LABEL: define hidden void @_Z6case20v(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[E:%.*]] = alloca [[STRUCT_EMPTY:%.*]], align 1
// CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 1 [[E]], ptr align 1 @__const._Z6case20v.E, i32 1, i1 false)
// CHECK-NEXT: ret void
@@ -1033,6 +1055,7 @@
// CHECK-LABEL: define hidden void @_Z6case215Empty(
// CHECK-SAME: ptr noundef byval([[STRUCT_EMPTY:%.*]]) align 1 [[E:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[TI:%.*]] = alloca [[STRUCT_TWOINTS:%.*]], align 1
// CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 1 [[TI]], ptr align 1 @__const._Z6case215Empty.TI, i32 8, i1 false)
// CHECK-NEXT: ret void
@@ -1045,6 +1068,7 @@
// CHECK-LABEL: define hidden void @_Z6case22v(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[UO:%.*]] = alloca [[STRUCT_UNNAMEDONLY:%.*]], align 1
// CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 1 [[UO]], ptr align 1 @__const._Z6case22v.UO, i32 1, i1 false)
// CHECK-NEXT: ret void
@@ -1057,6 +1081,7 @@
// CHECK-LABEL: define hidden void @_Z6case2311UnnamedOnly(
// CHECK-SAME: ptr noundef byval([[STRUCT_UNNAMEDONLY:%.*]]) align 1 [[UO:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[TI:%.*]] = alloca [[STRUCT_TWOINTS:%.*]], align 1
// CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 1 [[TI]], ptr align 1 @__const._Z6case2311UnnamedOnly.TI, i32 8, i1 false)
// CHECK-NEXT: ret void
@@ -1070,6 +1095,7 @@
// CHECK-LABEL: define hidden void @_Z6case24v(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[ED:%.*]] = alloca [[STRUCT_EMPTYDERIVED:%.*]], align 1
// CHECK-NEXT: [[UD:%.*]] = alloca [[STRUCT_UNNAMEDDERIVED:%.*]], align 1
// CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 1 [[ED]], ptr align 1 @__const._Z6case24v.ED, i32 1, i1 false)
@@ -1084,6 +1110,7 @@
// CHECK-LABEL: define hidden void @_Z6case2512EmptyDerived14UnnamedDerived(
// CHECK-SAME: ptr noundef byval([[STRUCT_EMPTYDERIVED:%.*]]) align 1 [[ED:%.*]], ptr noundef byval([[STRUCT_UNNAMEDDERIVED:%.*]]) align 1 [[UD:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[TI1:%.*]] = alloca [[STRUCT_TWOINTS:%.*]], align 1
// CHECK-NEXT: [[TI2:%.*]] = alloca [[STRUCT_TWOINTS]], align 1
// CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 1 [[TI1]], ptr align 1 @__const._Z6case2512EmptyDerived14UnnamedDerived.TI1, i32 8, i1 false)
@@ -1098,6 +1125,7 @@
// CHECK-LABEL: define hidden void @_Z6case267TwoInts(
// CHECK-SAME: ptr noundef byval([[STRUCT_TWOINTS:%.*]]) align 1 [[TI:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[F:%.*]] = alloca <4 x float>, align 4
// CHECK-NEXT: [[F2:%.*]] = alloca <3 x float>, align 4
// CHECK-NEXT: [[Z:%.*]] = getelementptr inbounds nuw [[STRUCT_TWOINTS]], ptr [[TI]], i32 0, i32 0
@@ -1136,6 +1164,7 @@
// CHECK-LABEL: define hidden void @_Z6case2714CustomResource(
// CHECK-SAME: ptr noundef byval([[STRUCT_CUSTOMRESOURCE:%.*]]) align 1 [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[B:%.*]] = alloca [[STRUCT_CUSTOMRESOURCE]], align 1
// CHECK-NEXT: [[H:%.*]] = getelementptr inbounds nuw [[STRUCT_CUSTOMRESOURCE]], ptr [[B]], i32 0, i32 0
// CHECK-NEXT: [[H1:%.*]] = getelementptr inbounds nuw [[STRUCT_CUSTOMRESOURCE]], ptr [[A]], i32 0, i32 0
@@ -1152,6 +1181,7 @@
// CHECK-LABEL: define hidden void @_Z6case289TwoFloats(
// CHECK-SAME: ptr noundef byval([[STRUCT_TWOFLOATS:%.*]]) align 1 [[TF:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[TI:%.*]] = alloca [[STRUCT_TWOINTS:%.*]], align 1
// CHECK-NEXT: [[REF_TMP:%.*]] = alloca [[STRUCT_TWOINTS]], align 1
// CHECK-NEXT: [[AGG_TEMP:%.*]] = alloca [[STRUCT_TWOFLOATS]], align 1
@@ -1196,6 +1226,7 @@
// CHECK-LABEL: define hidden void @_Z6case2910FourFloats(
// CHECK-SAME: ptr noundef byval([[STRUCT_FOURFLOATS:%.*]]) align 1 [[FF:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[INTS:%.*]] = alloca [2 x i32], align 4
// CHECK-NEXT: [[REF_TMP:%.*]] = alloca [2 x i32], align 4
// CHECK-NEXT: [[AGG_TEMP:%.*]] = alloca [[STRUCT_FOURFLOATS]], align 1
@@ -1243,6 +1274,7 @@
// CHECK-LABEL: define hidden void @_Z6case30v(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[ARR:%.*]] = alloca [4 x float], align 4
// CHECK-NEXT: [[TI:%.*]] = alloca [[STRUCT_TWOINTS:%.*]], align 1
// CHECK-NEXT: [[REF_TMP:%.*]] = alloca [[STRUCT_TWOINTS]], align 1
@@ -1294,6 +1326,7 @@
// CHECK-LABEL: define hidden void @_Z6case31v(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[F:%.*]] = alloca float, align 4
// CHECK-NEXT: [[SB:%.*]] = alloca [[STRUCT_SLICYBITS:%.*]], align 1
// CHECK-NEXT: store float 1.000000e+00, ptr [[F]], align 4
diff --git a/clang/test/CodeGenHLSL/BasicFeatures/MatrixConstructor.hlsl b/clang/test/CodeGenHLSL/BasicFeatures/MatrixConstructor.hlsl
index 467fb38..3d32149 100644
--- a/clang/test/CodeGenHLSL/BasicFeatures/MatrixConstructor.hlsl
+++ b/clang/test/CodeGenHLSL/BasicFeatures/MatrixConstructor.hlsl
@@ -4,6 +4,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <6 x float> @_Z5case1v(
// CHECK-SAME: ) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: ret <6 x float> <float 0.000000e+00, float 2.000000e+00, float 4.000000e+00, float 1.000000e+00, float 3.000000e+00, float 5.000000e+00>
//
float3x2 case1() {
@@ -24,6 +25,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <6 x float> @_Z5case2v(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[CALL:%.*]] = call noundef nonnull align 4 dereferenceable(4) ptr @_ZN4hlsl18RWStructuredBufferIfEixEj(ptr noundef nonnull align 4 dereferenceable(8) @_ZL2In, i32 noundef 0) #[[ATTR4:[0-9]+]]
// CHECK-NEXT: [[CALL1:%.*]] = call noundef nonnull align 4 dereferenceable(4) ptr @_ZN4hlsl18RWStructuredBufferIfEixEj(ptr noundef nonnull align 4 dereferenceable(8) @_ZL2In, i32 noundef 1) #[[ATTR4]]
// CHECK-NEXT: [[CALL2:%.*]] = call noundef nonnull align 4 dereferenceable(4) ptr @_ZN4hlsl18RWStructuredBufferIfEixEj(ptr noundef nonnull align 4 dereferenceable(8) @_ZL2In, i32 noundef 2) #[[ATTR4]]
@@ -60,6 +62,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <6 x float> @_Z5case3Dv3_fS_(
// CHECK-SAME: <3 x float> noundef nofpclass(nan inf) [[A:%.*]], <3 x float> noundef nofpclass(nan inf) [[B:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca <3 x float>, align 4
// CHECK-NEXT: [[B_ADDR:%.*]] = alloca <3 x float>, align 4
// CHECK-NEXT: store <3 x float> [[A]], ptr [[A_ADDR]], align 4
diff --git a/clang/test/CodeGenHLSL/BasicFeatures/MatrixElementTypeCast.hlsl b/clang/test/CodeGenHLSL/BasicFeatures/MatrixElementTypeCast.hlsl
index dd9dd70..ec8c794 100644
--- a/clang/test/CodeGenHLSL/BasicFeatures/MatrixElementTypeCast.hlsl
+++ b/clang/test/CodeGenHLSL/BasicFeatures/MatrixElementTypeCast.hlsl
@@ -5,6 +5,7 @@
// CHECK-LABEL: define hidden noundef <6 x i32> @_Z22elementwise_type_cast0u11matrix_typeILm3ELm2EfE(
// CHECK-SAME: <6 x float> noundef nofpclass(nan inf) [[F32:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// ROW-CHECK-NEXT: [[F32_ADDR:%.*]] = alloca [3 x <2 x float>], align 4
// ROW-CHECK-NEXT: [[I32:%.*]] = alloca [3 x <2 x i32>], align 4
// COL-CHECK-NEXT: [[F32_ADDR:%.*]] = alloca [2 x <3 x float>], align 4
@@ -24,6 +25,7 @@
// CHECK-LABEL: define hidden noundef <6 x i32> @_Z22elementwise_type_cast1u11matrix_typeILm3ELm2EsE(
// CHECK-SAME: <6 x i16> noundef [[I16_32:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// ROW-CHECK-NEXT: [[I16_32_ADDR:%.*]] = alloca [3 x <2 x i16>], align 2
// ROW-CHECK-NEXT: [[I32:%.*]] = alloca [3 x <2 x i32>], align 4
// COL-CHECK-NEXT: [[I16_32_ADDR:%.*]] = alloca [2 x <3 x i16>], align 2
@@ -43,6 +45,7 @@
// CHECK-LABEL: define hidden noundef <6 x i32> @_Z22elementwise_type_cast2u11matrix_typeILm3ELm2ElE(
// CHECK-SAME: <6 x i64> noundef [[I64_32:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// ROW-CHECK-NEXT: [[I64_32_ADDR:%.*]] = alloca [3 x <2 x i64>], align 8
// ROW-CHECK-NEXT: [[I32:%.*]] = alloca [3 x <2 x i32>], align 4
// COL-CHECK-NEXT: [[I64_32_ADDR:%.*]] = alloca [2 x <3 x i64>], align 8
@@ -62,6 +65,7 @@
// CHECK-LABEL: define hidden noundef <6 x i16> @_Z22elementwise_type_cast3u11matrix_typeILm2ELm3EDhE(
// CHECK-SAME: <6 x half> noundef nofpclass(nan inf) [[H23:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// ROW-CHECK-NEXT: [[H23_ADDR:%.*]] = alloca [2 x <3 x half>], align 2
// ROW-CHECK-NEXT: [[I23:%.*]] = alloca [2 x <3 x i16>], align 2
// COL-CHECK-NEXT: [[H23_ADDR:%.*]] = alloca [3 x <2 x half>], align 2
@@ -81,6 +85,7 @@
// CHECK-LABEL: define hidden noundef <6 x i32> @_Z22elementwise_type_cast4u11matrix_typeILm3ELm2EdE(
// CHECK-SAME: <6 x double> noundef nofpclass(nan inf) [[D32:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// ROW-CHECK-NEXT: [[D32_ADDR:%.*]] = alloca [3 x <2 x double>], align 8
// ROW-CHECK-NEXT: [[I32:%.*]] = alloca [3 x <2 x i32>], align 4
// COL-CHECK-NEXT: [[D32_ADDR:%.*]] = alloca [2 x <3 x double>], align 8
@@ -100,6 +105,7 @@
// CHECK-LABEL: define hidden void @_Z5call2v(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A:%.*]] = alloca [2 x [1 x i32]], align 4
// ROW-CHECK-NEXT: [[B:%.*]] = alloca [2 x <1 x i32>], align 4
// COL-CHECK-NEXT: [[B:%.*]] = alloca [1 x <2 x i32>], align 4
@@ -130,6 +136,7 @@
// CHECK-LABEL: define hidden void @_Z5call3v(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[S:%.*]] = alloca [[STRUCT_S:%.*]], align 1
// ROW-CHECK-NEXT: [[A:%.*]] = alloca [2 x <1 x i32>], align 4
// COL-CHECK-NEXT: [[A:%.*]] = alloca [1 x <2 x i32>], align 4
@@ -167,6 +174,7 @@
// CHECK-LABEL: define hidden void @_Z5call47Derived(
// CHECK-SAME: ptr noundef byval([[STRUCT_DERIVED:%.*]]) align 1 [[D:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A:%.*]] = alloca [2 x <2 x i32>], align 4
// CHECK-NEXT: [[AGG_TEMP:%.*]] = alloca [[STRUCT_DERIVED]], align 1
// CHECK-NEXT: [[FLATCAST_TMP:%.*]] = alloca <4 x i32>, align 4
@@ -202,6 +210,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <4 x float> @_Z5call5Dv4_f(
// CHECK-SAME: <4 x float> noundef nofpclass(nan inf) [[V:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <4 x float>, align 4
// CHECK-NEXT: [[M:%.*]] = alloca [2 x <2 x float>], align 4
// CHECK-NEXT: [[HLSL_EWCAST_SRC:%.*]] = alloca <4 x float>, align 4
diff --git a/clang/test/CodeGenHLSL/BasicFeatures/MatrixExplicitTruncation.hlsl b/clang/test/CodeGenHLSL/BasicFeatures/MatrixExplicitTruncation.hlsl
index 587ccfe..4ab82fa 100644
--- a/clang/test/CodeGenHLSL/BasicFeatures/MatrixExplicitTruncation.hlsl
+++ b/clang/test/CodeGenHLSL/BasicFeatures/MatrixExplicitTruncation.hlsl
@@ -5,6 +5,7 @@
// CHECK-LABEL: define hidden noundef <12 x i32> @_Z10trunc_castu11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// ROW-CHECK-NEXT: [[I34:%.*]] = alloca [3 x <4 x i32>], align 4
// COL-CHECK-NEXT: [[I34:%.*]] = alloca [4 x <3 x i32>], align 4
@@ -24,6 +25,7 @@
// CHECK-LABEL: define hidden noundef <12 x i32> @_Z11trunc_cast0u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// ROW-CHECK-NEXT: [[I43:%.*]] = alloca [4 x <3 x i32>], align 4
// COL-CHECK-NEXT: [[I43:%.*]] = alloca [3 x <4 x i32>], align 4
@@ -43,6 +45,7 @@
// CHECK-LABEL: define hidden noundef <9 x i32> @_Z11trunc_cast1u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: [[I33:%.*]] = alloca [3 x <3 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[I44]], ptr [[I44_ADDR]], align 4
@@ -60,6 +63,7 @@
// CHECK-LABEL: define hidden noundef <6 x i32> @_Z11trunc_cast2u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// ROW-CHECK-NEXT: [[I32:%.*]] = alloca [3 x <2 x i32>], align 4
// COL-CHECK-NEXT: [[I32:%.*]] = alloca [2 x <3 x i32>], align 4
@@ -79,6 +83,7 @@
// CHECK-LABEL: define hidden noundef <6 x i32> @_Z11trunc_cast3u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// ROW-CHECK-NEXT: [[I23:%.*]] = alloca [2 x <3 x i32>], align 4
// COL-CHECK-NEXT: [[I23:%.*]] = alloca [3 x <2 x i32>], align 4
@@ -98,6 +103,7 @@
// CHECK-LABEL: define hidden noundef <4 x i32> @_Z11trunc_cast4u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: [[I22:%.*]] = alloca [2 x <2 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[I44]], ptr [[I44_ADDR]], align 4
@@ -115,6 +121,7 @@
// CHECK-LABEL: define hidden noundef <2 x i32> @_Z11trunc_cast5u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// ROW-CHECK-NEXT: [[I21:%.*]] = alloca [2 x <1 x i32>], align 4
// COL-CHECK-NEXT: [[I21:%.*]] = alloca [1 x <2 x i32>], align 4
@@ -134,6 +141,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z11trunc_cast6u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: [[I1:%.*]] = alloca i32, align 4
// CHECK-NEXT: store <16 x i32> [[I44]], ptr [[I44_ADDR]], align 4
@@ -151,6 +159,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z16trunc_multi_castu11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: [[I1:%.*]] = alloca i32, align 4
// CHECK-NEXT: store <16 x i32> [[I44]], ptr [[I44_ADDR]], align 4
diff --git a/clang/test/CodeGenHLSL/BasicFeatures/MatrixImplicitTruncation.hlsl b/clang/test/CodeGenHLSL/BasicFeatures/MatrixImplicitTruncation.hlsl
index 1a9a953..52105b8 100644
--- a/clang/test/CodeGenHLSL/BasicFeatures/MatrixImplicitTruncation.hlsl
+++ b/clang/test/CodeGenHLSL/BasicFeatures/MatrixImplicitTruncation.hlsl
@@ -5,6 +5,7 @@
// CHECK-LABEL: define hidden noundef <12 x i32> @_Z10trunc_castu11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// ROW-CHECK-NEXT: [[I34:%.*]] = alloca [3 x <4 x i32>], align 4
// COL-CHECK-NEXT: [[I34:%.*]] = alloca [4 x <3 x i32>], align 4
@@ -24,6 +25,7 @@
// CHECK-LABEL: define hidden noundef <12 x i32> @_Z11trunc_cast0u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// ROW-CHECK-NEXT: [[I43:%.*]] = alloca [4 x <3 x i32>], align 4
// COL-CHECK-NEXT: [[I43:%.*]] = alloca [3 x <4 x i32>], align 4
@@ -43,6 +45,7 @@
// CHECK-LABEL: define hidden noundef <9 x i32> @_Z11trunc_cast1u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: [[I33:%.*]] = alloca [3 x <3 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[I44]], ptr [[I44_ADDR]], align 4
@@ -60,6 +63,7 @@
// CHECK-LABEL: define hidden noundef <6 x i32> @_Z11trunc_cast2u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// ROW-CHECK-NEXT: [[I32:%.*]] = alloca [3 x <2 x i32>], align 4
// COL-CHECK-NEXT: [[I32:%.*]] = alloca [2 x <3 x i32>], align 4
@@ -79,6 +83,7 @@
// CHECK-LABEL: define hidden noundef <6 x i32> @_Z11trunc_cast3u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// ROW-CHECK-NEXT: [[I23:%.*]] = alloca [2 x <3 x i32>], align 4
// COL-CHECK-NEXT: [[I23:%.*]] = alloca [3 x <2 x i32>], align 4
@@ -98,6 +103,7 @@
// CHECK-LABEL: define hidden noundef <4 x i32> @_Z11trunc_cast4u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: [[I22:%.*]] = alloca [2 x <2 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[I44]], ptr [[I44_ADDR]], align 4
@@ -115,6 +121,7 @@
// CHECK-LABEL: define hidden noundef <2 x i32> @_Z11trunc_cast5u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// ROW-CHECK-NEXT: [[I21:%.*]] = alloca [2 x <1 x i32>], align 4
// COL-CHECK-NEXT: [[I21:%.*]] = alloca [1 x <2 x i32>], align 4
@@ -134,6 +141,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z11trunc_cast6u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[I44:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[I44_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: [[I1:%.*]] = alloca i32, align 4
// CHECK-NEXT: store <16 x i32> [[I44]], ptr [[I44_ADDR]], align 4
diff --git a/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptConstSwizzle.hlsl b/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptConstSwizzle.hlsl
index 378bd04..3df51ea 100644
--- a/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptConstSwizzle.hlsl
+++ b/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptConstSwizzle.hlsl
@@ -4,6 +4,7 @@
// CHECK-LABEL: define hidden void @_Z10setMatrix1Ru11matrix_typeILm4ELm4EfEDv4_f(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[M:%.*]], <4 x float> noundef nofpclass(nan inf) [[V:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <4 x float>, align 4
// CHECK-NEXT: store ptr [[M]], ptr [[M_ADDR]], align 4
@@ -31,6 +32,7 @@
// CHECK-LABEL: define hidden void @_Z10setMatrix2Ru11matrix_typeILm4ELm4EiEDv4_i(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[M:%.*]], <4 x i32> noundef [[V:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <4 x i32>, align 4
// CHECK-NEXT: store ptr [[M]], ptr [[M_ADDR]], align 4
@@ -58,6 +60,7 @@
// CHECK-LABEL: define hidden void @_Z22setMatrixVectorSwizzleRu11matrix_typeILm2ELm3EiEDv3_i(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(24) [[M:%.*]], <3 x i32> noundef [[V:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <3 x i32>, align 4
// CHECK-NEXT: store ptr [[M]], ptr [[M_ADDR]], align 4
@@ -83,6 +86,7 @@
// CHECK-LABEL: define hidden void @_Z24setVectorOnMatrixSwizzleRu11matrix_typeILm2ELm3EiEDv3_i(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(24) [[M:%.*]], <3 x i32> noundef [[V:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <3 x i32>, align 4
// CHECK-NEXT: store ptr [[M]], ptr [[M_ADDR]], align 4
@@ -107,6 +111,7 @@
// CHECK-LABEL: define hidden void @_Z19setMatrixFromMatrixRu11matrix_typeILm2ELm3EiES_i(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(24) [[M:%.*]], <6 x i32> noundef [[N:%.*]], i32 noundef [[MINDEX:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[N_ADDR:%.*]] = alloca [3 x <2 x i32>], align 4
// CHECK-NEXT: [[MINDEX_ADDR:%.*]] = alloca i32, align 4
@@ -138,6 +143,7 @@
// CHECK-LABEL: define hidden void @_Z26setMatrixSwizzleFromMatrixRu11matrix_typeILm2ELm3EiES_i(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(24) [[M:%.*]], <6 x i32> noundef [[N:%.*]], i32 noundef [[NINDEX:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[N_ADDR:%.*]] = alloca [3 x <2 x i32>], align 4
// CHECK-NEXT: [[NINDEX_ADDR:%.*]] = alloca i32, align 4
@@ -175,6 +181,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <2 x float> @_Z2fnu11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[M:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[M]], ptr [[M_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[M_ADDR]], align 4
diff --git a/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptDynamicSwizzle.hlsl b/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptDynamicSwizzle.hlsl
index 36e3515..cf8f2fa 100644
--- a/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptDynamicSwizzle.hlsl
+++ b/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptDynamicSwizzle.hlsl
@@ -4,6 +4,7 @@
// CHECK-LABEL: define hidden void @_Z9setMatrixRu11matrix_typeILm4ELm4EfEiDv4_f(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[M:%.*]], i32 noundef [[INDEX:%.*]], <4 x float> noundef nofpclass(nan inf) [[V:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[INDEX_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <4 x float>, align 4
@@ -38,6 +39,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <3 x float> @_Z9getMatrixu11matrix_typeILm4ELm4EfEi(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[M:%.*]], i32 noundef [[INDEX:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: [[INDEX_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store <16 x float> [[M]], ptr [[M_ADDR]], align 4
@@ -62,6 +64,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <4 x float> @_Z9getMatrixu11matrix_typeILm3ELm3EfEi(
// CHECK-SAME: <9 x float> noundef nofpclass(nan inf) [[M:%.*]], i32 noundef [[INDEX:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca [3 x <3 x float>], align 4
// CHECK-NEXT: [[INDEX_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store <9 x float> [[M]], ptr [[M_ADDR]], align 4
@@ -89,6 +92,7 @@
// CHECK-LABEL: define hidden noundef <3 x i32> @_Z19getMatrixSwizzle2x3Ru11matrix_typeILm2ELm3EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(24) [[M:%.*]], i32 noundef [[INDEX:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[INDEX_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[M]], ptr [[M_ADDR]], align 4
@@ -114,6 +118,7 @@
// CHECK-LABEL: define hidden void @_Z26setMatrixSwizzleFromMatrixRu11matrix_typeILm2ELm3EiES_i(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(24) [[M:%.*]], <6 x i32> noundef [[N:%.*]], i32 noundef [[INDEX:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[N_ADDR:%.*]] = alloca [3 x <2 x i32>], align 4
// CHECK-NEXT: [[INDEX_ADDR:%.*]] = alloca i32, align 4
diff --git a/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptGetter.hlsl b/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptGetter.hlsl
index 7358849..03c20d5 100644
--- a/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptGetter.hlsl
+++ b/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptGetter.hlsl
@@ -4,6 +4,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <4 x float> @_Z24getFloatVecMatrixDynamicu11matrix_typeILm4ELm4EfEi(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[M:%.*]], i32 noundef [[INDEX:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: [[INDEX_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store <16 x float> [[M]], ptr [[M_ADDR]], align 4
@@ -31,6 +32,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z27getFloatScalarMatrixDynamicu11matrix_typeILm2ELm1EfEi(
// CHECK-SAME: <2 x float> noundef nofpclass(nan inf) [[M:%.*]], i32 noundef [[INDEX:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca [1 x <2 x float>], align 4
// CHECK-NEXT: [[INDEX_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store <2 x float> [[M]], ptr [[M_ADDR]], align 4
@@ -50,6 +52,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z28getFloatScalarMatrixConstantu11matrix_typeILm2ELm1EfE(
// CHECK-SAME: <2 x float> noundef nofpclass(nan inf) [[M:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca [1 x <2 x float>], align 4
// CHECK-NEXT: store <2 x float> [[M]], ptr [[M_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <2 x float>, ptr [[M_ADDR]], align 4
@@ -65,6 +68,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z29getFloatScalarMatrixConstant2u11matrix_typeILm2ELm1EfE(
// CHECK-SAME: <2 x float> noundef nofpclass(nan inf) [[M:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca [1 x <2 x float>], align 4
// CHECK-NEXT: store <2 x float> [[M]], ptr [[M_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <2 x float>, ptr [[M_ADDR]], align 4
@@ -80,6 +84,7 @@
// CHECK-LABEL: define hidden noundef <4 x i32> @_Z19getIntMatrixDynamicu11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: <16 x i32> noundef [[M:%.*]], i32 noundef [[INDEX:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: [[INDEX_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store <16 x i32> [[M]], ptr [[M_ADDR]], align 4
@@ -107,6 +112,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <4 x float> @_Z22AddFloatMatrixConstantu11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[M:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[M]], ptr [[M_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[M_ADDR]], align 4
@@ -157,6 +163,7 @@
// CHECK-LABEL: define hidden noundef <4 x i32> @_Z20AddIntMatrixConstantu11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[M:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[M]], ptr [[M_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[M_ADDR]], align 4
@@ -207,6 +214,7 @@
// CHECK-LABEL: define hidden noundef <3 x i1> @_Z23getBoolVecMatrixDynamicu11matrix_typeILm2ELm3EbEi(
// CHECK-SAME: <6 x i1> noundef [[M:%.*]], i32 noundef [[INDEX:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca [3 x <2 x i32>], align 4
// CHECK-NEXT: [[INDEX_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: [[TMP0:%.*]] = zext <6 x i1> [[M]] to <6 x i32>
@@ -233,6 +241,7 @@
// CHECK-LABEL: define hidden noundef <4 x i1> @_Z24getBoolVecMatrixConstantu11matrix_typeILm4ELm4EbE(
// CHECK-SAME: <16 x i1> noundef [[M:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: [[TMP0:%.*]] = zext <16 x i1> [[M]] to <16 x i32>
// CHECK-NEXT: store <16 x i32> [[TMP0]], ptr [[M_ADDR]], align 4
@@ -255,6 +264,7 @@
// CHECK-LABEL: define hidden noundef i1 @_Z27getBoolScalarMatrixConstantu11matrix_typeILm3ELm1EbE(
// CHECK-SAME: <3 x i1> noundef [[M:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca [1 x <3 x i32>], align 4
// CHECK-NEXT: [[TMP0:%.*]] = zext <3 x i1> [[M]] to <3 x i32>
// CHECK-NEXT: store <3 x i32> [[TMP0]], ptr [[M_ADDR]], align 4
diff --git a/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptSetter.hlsl b/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptSetter.hlsl
index 8595bb93..50f4305 100644
--- a/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptSetter.hlsl
+++ b/clang/test/CodeGenHLSL/BasicFeatures/MatrixSingleSubscriptSetter.hlsl
@@ -4,6 +4,7 @@
// CHECK-LABEL: define hidden void @_Z9setMatrixRu11matrix_typeILm4ELm4EfEiDv4_f(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[M:%.*]], i32 noundef [[INDEX:%.*]], <4 x float> noundef nofpclass(nan inf) [[V:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[INDEX_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <4 x float>, align 4
@@ -38,6 +39,7 @@
// CHECK-LABEL: define hidden void @_Z15setMatrixScalarRu11matrix_typeILm2ELm1EfEif(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(8) [[M:%.*]], i32 noundef [[INDEX:%.*]], float noundef nofpclass(nan inf) [[S:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[INDEX_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: [[S_ADDR:%.*]] = alloca float, align 4
@@ -62,6 +64,7 @@
// CHECK-LABEL: define hidden void @_Z13setBoolMatrixRu11matrix_typeILm4ELm4EbEiDv4_b(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[M:%.*]], i32 noundef [[INDEX:%.*]], <4 x i1> noundef [[V:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[INDEX_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <4 x i32>, align 4
@@ -99,6 +102,7 @@
// CHECK-LABEL: define hidden void @_Z19setBoolMatrixScalarRu11matrix_typeILm2ELm1EbEib(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(8) [[M:%.*]], i32 noundef [[INDEX:%.*]], i1 noundef [[S:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[INDEX_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: [[S_ADDR:%.*]] = alloca i32, align 4
@@ -126,6 +130,7 @@
// CHECK-LABEL: define hidden void @_Z19setMatrixConstIndexRu11matrix_typeILm4ELm4EiES_(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[M:%.*]], <16 x i32> noundef [[N:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[N_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store ptr [[M]], ptr [[M_ADDR]], align 4
diff --git a/clang/test/CodeGenHLSL/BasicFeatures/MatrixSplat.hlsl b/clang/test/CodeGenHLSL/BasicFeatures/MatrixSplat.hlsl
index b65bcdf..8ddd575 100644
--- a/clang/test/CodeGenHLSL/BasicFeatures/MatrixSplat.hlsl
+++ b/clang/test/CodeGenHLSL/BasicFeatures/MatrixSplat.hlsl
@@ -4,6 +4,7 @@
// CHECK-LABEL: define hidden void @_Z13ConstantSplatv(
// CHECK-SAME: ) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> splat (i32 1), ptr [[M]], align 4
// CHECK-NEXT: ret void
@@ -15,6 +16,7 @@
// CHECK-LABEL: define hidden void @_Z18ConstantFloatSplatv(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M:%.*]] = alloca [2 x <2 x float>], align 4
// CHECK-NEXT: store <4 x float> splat (float 3.250000e+00), ptr [[M]], align 4
// CHECK-NEXT: ret void
@@ -26,6 +28,7 @@
// CHECK-LABEL: define hidden void @_Z21ConstantTrueBoolSplatv(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M:%.*]] = alloca [3 x <3 x i32>], align 4
// CHECK-NEXT: store <9 x i32> splat (i32 1), ptr [[M]], align 4
// CHECK-NEXT: ret void
@@ -37,6 +40,7 @@
// CHECK-LABEL: define hidden void @_Z22ConstantFalseBoolSplatv(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M:%.*]] = alloca [3 x <3 x i32>], align 4
// CHECK-NEXT: store <9 x i32> zeroinitializer, ptr [[M]], align 4
// CHECK-NEXT: ret void
@@ -48,6 +52,7 @@
// CHECK-LABEL: define hidden void @_Z12DynamicSplatf(
// CHECK-SAME: float noundef nofpclass(nan inf) [[VALUE:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[VALUE_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: [[M:%.*]] = alloca [3 x <3 x float>], align 4
// CHECK-NEXT: store float [[VALUE]], ptr [[VALUE_ADDR]], align 4
@@ -64,6 +69,7 @@
// CHECK-LABEL: define hidden void @_Z16DynamicBoolSplatb(
// CHECK-SAME: i1 noundef [[VALUE:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[VALUE_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: [[M:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: [[STOREDV:%.*]] = zext i1 [[VALUE]] to i32
@@ -83,6 +89,7 @@
// CHECK-LABEL: define hidden void @_Z13CastThenSplatDv4_f(
// CHECK-SAME: <4 x float> noundef nofpclass(nan inf) [[VALUE:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[VALUE_ADDR:%.*]] = alloca <4 x float>, align 4
// CHECK-NEXT: [[M:%.*]] = alloca [3 x <3 x float>], align 4
// CHECK-NEXT: store <4 x float> [[VALUE]], ptr [[VALUE_ADDR]], align 4
@@ -100,6 +107,7 @@
// CHECK-LABEL: define hidden void @_Z30ExplicitIntToBoolCastThenSplatDv3_i(
// CHECK-SAME: <3 x i32> noundef [[VALUE:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[VALUE_ADDR:%.*]] = alloca <3 x i32>, align 4
// CHECK-NEXT: [[M:%.*]] = alloca [2 x <2 x i32>], align 4
// CHECK-NEXT: store <3 x i32> [[VALUE]], ptr [[VALUE_ADDR]], align 4
@@ -119,6 +127,7 @@
// CHECK-LABEL: define hidden void @_Z32ExplicitFloatToBoolCastThenSplatDv2_f(
// CHECK-SAME: <2 x float> noundef nofpclass(nan inf) [[VALUE:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[VALUE_ADDR:%.*]] = alloca <2 x float>, align 4
// CHECK-NEXT: [[M:%.*]] = alloca [3 x <2 x i32>], align 4
// CHECK-NEXT: store <2 x float> [[VALUE]], ptr [[VALUE_ADDR]], align 4
@@ -138,6 +147,7 @@
// CHECK-LABEL: define hidden void @_Z32ExplicitBoolToFloatCastThenSplatb(
// CHECK-SAME: i1 noundef [[VALUE:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[VALUE_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: [[M:%.*]] = alloca [2 x <3 x float>], align 4
// CHECK-NEXT: [[STOREDV:%.*]] = zext i1 [[VALUE]] to i32
@@ -157,6 +167,7 @@
// CHECK-LABEL: define hidden void @_Z32ImplicitFloatToBoolCastThenSplatf(
// CHECK-SAME: float noundef nofpclass(nan inf) [[VALUE:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[VALUE_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: [[M:%.*]] = alloca [3 x <2 x i32>], align 4
// CHECK-NEXT: store float [[VALUE]], ptr [[VALUE_ADDR]], align 4
@@ -175,6 +186,7 @@
// CHECK-LABEL: define hidden void @_Z32ImplicitBoolToFloatCastThenSplatb(
// CHECK-SAME: i1 noundef [[VALUE:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[VALUE_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: [[M:%.*]] = alloca [2 x <3 x float>], align 4
// CHECK-NEXT: [[STOREDV:%.*]] = zext i1 [[VALUE]] to i32
diff --git a/clang/test/CodeGenHLSL/BasicFeatures/MatrixToAndFromVectorConstructors.hlsl b/clang/test/CodeGenHLSL/BasicFeatures/MatrixToAndFromVectorConstructors.hlsl
index 2b3fafe..96859ca 100644
--- a/clang/test/CodeGenHLSL/BasicFeatures/MatrixToAndFromVectorConstructors.hlsl
+++ b/clang/test/CodeGenHLSL/BasicFeatures/MatrixToAndFromVectorConstructors.hlsl
@@ -4,6 +4,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <4 x float> @_Z2fnu11matrix_typeILm2ELm2EfE(
// CHECK-SAME: <4 x float> noundef nofpclass(nan inf) [[M:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca [2 x <2 x float>], align 4
// CHECK-NEXT: [[V:%.*]] = alloca <4 x float>, align 4
// CHECK-NEXT: store <4 x float> [[M]], ptr [[M_ADDR]], align 4
@@ -33,6 +34,7 @@
// CHECK-LABEL: define hidden noundef <4 x i32> @_Z2fnDv4_i(
// CHECK-SAME: <4 x i32> noundef [[V:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <4 x i32>, align 4
// CHECK-NEXT: [[M:%.*]] = alloca [2 x <2 x i32>], align 4
// CHECK-NEXT: store <4 x i32> [[V]], ptr [[V_ADDR]], align 4
@@ -62,6 +64,7 @@
// CHECK-LABEL: define hidden noundef <2 x i32> @_Z3fn1Dv2_i(
// CHECK-SAME: <2 x i32> noundef [[V:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <2 x i32>, align 4
// CHECK-NEXT: store <2 x i32> [[V]], ptr [[V_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <2 x i32>, ptr [[V_ADDR]], align 4
@@ -79,6 +82,7 @@
// CHECK-LABEL: define hidden noundef <3 x i1> @_Z3fn2Dv3_b(
// CHECK-SAME: <3 x i1> noundef [[B:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[B_ADDR:%.*]] = alloca <3 x i32>, align 4
// CHECK-NEXT: [[TMP0:%.*]] = zext <3 x i1> [[B]] to <3 x i32>
// CHECK-NEXT: store <3 x i32> [[TMP0]], ptr [[B_ADDR]], align 4
@@ -103,6 +107,7 @@
// CHECK-LABEL: define hidden noundef <3 x i32> @_Z3fn3u11matrix_typeILm1ELm3EbE(
// CHECK-SAME: <3 x i1> noundef [[B:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// COL-CHECK-NEXT: [[B_ADDR:%.*]] = alloca [3 x <1 x i32>], align 4
// ROW-CHECK-NEXT: [[B_ADDR:%.*]] = alloca [1 x <3 x i32>], align 4
// CHECK-NEXT: [[TMP0:%.*]] = zext <3 x i1> [[B]] to <3 x i32>
diff --git a/clang/test/CodeGenHLSL/BoolMatrix.hlsl b/clang/test/CodeGenHLSL/BoolMatrix.hlsl
index e9841ae..f87a054 100644
--- a/clang/test/CodeGenHLSL/BoolMatrix.hlsl
+++ b/clang/test/CodeGenHLSL/BoolMatrix.hlsl
@@ -10,6 +10,7 @@
// CHECK-LABEL: define hidden noundef i1 @_Z3fn1v(
// CHECK-SAME: ) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[RETVAL:%.*]] = alloca i1, align 4
// CHECK-NEXT: [[B:%.*]] = alloca [2 x <2 x i32>], align 4
// CHECK-NEXT: store <4 x i32> splat (i32 1), ptr [[B]], align 4
@@ -27,6 +28,7 @@
// CHECK-LABEL: define hidden noundef <4 x i1> @_Z3fn2b(
// CHECK-SAME: i1 noundef [[V:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[RETVAL:%.*]] = alloca <4 x i1>, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: [[A:%.*]] = alloca [2 x <2 x i32>], align 4
@@ -55,6 +57,7 @@
// CHECK-LABEL: define hidden noundef i1 @_Z3fn3v(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[RETVAL:%.*]] = alloca i1, align 4
// CHECK-NEXT: [[S:%.*]] = alloca [[STRUCT_S:%.*]], align 1
// CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 1 [[S]], ptr align 1 @__const._Z3fn3v.s, i32 20, i1 false)
@@ -73,6 +76,7 @@
// CHECK-LABEL: define hidden noundef i1 @_Z3fn4v(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[RETVAL:%.*]] = alloca i1, align 4
// CHECK-NEXT: [[ARR:%.*]] = alloca [2 x [2 x <2 x i32>]], align 4
// CHECK-NEXT: store <4 x i32> splat (i32 1), ptr [[ARR]], align 4
@@ -93,6 +97,7 @@
// CHECK-LABEL: define hidden void @_Z3fn5v(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[M:%.*]] = alloca [2 x <2 x i32>], align 4
// CHECK-NEXT: store <4 x i32> splat (i32 1), ptr [[M]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = getelementptr <4 x i32>, ptr [[M]], i32 0, i32 3
@@ -107,6 +112,7 @@
// CHECK-LABEL: define hidden void @_Z3fn6v(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[V:%.*]] = alloca i32, align 4
// CHECK-NEXT: [[S:%.*]] = alloca [[STRUCT_S:%.*]], align 1
// CHECK-NEXT: store i32 0, ptr [[V]], align 4
@@ -128,6 +134,7 @@
// CHECK-LABEL: define hidden void @_Z3fn7v(
// CHECK-SAME: ) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[ARR:%.*]] = alloca [2 x [2 x <2 x i32>]], align 4
// CHECK-NEXT: store <4 x i32> splat (i32 1), ptr [[ARR]], align 4
// CHECK-NEXT: [[ARRAYINIT_ELEMENT:%.*]] = getelementptr inbounds [2 x <2 x i32>], ptr [[ARR]], i32 1
@@ -145,6 +152,7 @@
// CHECK-LABEL: define hidden noundef <16 x i1> @_Z3fn8u11matrix_typeILm4ELm4EbE(
// CHECK-SAME: <16 x i1> noundef [[M:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[RETVAL:%.*]] = alloca <16 x i1>, align 4
// CHECK-NEXT: [[M_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: [[TMP0:%.*]] = zext <16 x i1> [[M]] to <16 x i32>
diff --git a/clang/test/CodeGenHLSL/GlobalConstructorFunction.hlsl b/clang/test/CodeGenHLSL/GlobalConstructorFunction.hlsl
index 5553f8c..2b039b1 100644
--- a/clang/test/CodeGenHLSL/GlobalConstructorFunction.hlsl
+++ b/clang/test/CodeGenHLSL/GlobalConstructorFunction.hlsl
@@ -24,22 +24,23 @@
// CHECK: define void @main()
// CHECK-NEXT: entry:
+// INLINE-NEXT: alloca
+// CHECK: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// Verify function constructors are emitted
// NOINLINE-NEXT: call void @_Z13call_me_firstv()
// NOINLINE-NEXT: call void @_Z12then_call_mev()
// NOINLINE-NEXT: call void @_GLOBAL__sub_I_GlobalConstructorFunction.hlsl()
-// NOINLINE-NEXT: %0 = call i32 @llvm.dx.flattened.thread.id.in.group()
-// NOINLINE-NEXT: call void @_Z4mainj(i32 %0)
+// NOINLINE-NEXT: %[[#THREAD_ID:]] = call i32 @llvm.dx.flattened.thread.id.in.group()
+// NOINLINE-NEXT: call void @_Z4mainj(i32 %[[#THREAD_ID]])
// NOINLINE-NEXT: call void @_Z12call_me_lastv(
// NOINLINE-NEXT: ret void
// Verify constructor calls are inlined when AlwaysInline is run
-// INLINE-NEXT: alloca
// INLINE-NEXT: store i32 12
// INLINE-NEXT: store i32 13
// INLINE-NEXT: %[[HANDLE:.*]] = call target("dx.CBuffer", %"__cblayout_$Globals") @"llvm.dx.resource.handlefromimplicitbinding.tdx.CBuffer_s___cblayout_$Globalsst"(i32 0, i32 0, i32 1, i32 0, ptr @"$Globals.str")
// INLINE-NEXT: store target("dx.CBuffer", %"__cblayout_$Globals") %[[HANDLE]], ptr @"$Globals.cb", align 4
-// INLINE-NEXT: %0 = call i32 @llvm.dx.flattened.thread.id.in.group()
-// INLINE-NEXT: store i32 %
+// INLINE-NEXT: %[[#THREAD_ID:]] = call i32 @llvm.dx.flattened.thread.id.in.group()
+// INLINE-NEXT: store i32 %[[#THREAD_ID]]
// INLINE-NEXT: store i32 0
// INLINE: ret void
diff --git a/clang/test/CodeGenHLSL/GlobalConstructorLib.hlsl b/clang/test/CodeGenHLSL/GlobalConstructorLib.hlsl
index 0ef4b43..93d4492 100644
--- a/clang/test/CodeGenHLSL/GlobalConstructorLib.hlsl
+++ b/clang/test/CodeGenHLSL/GlobalConstructorLib.hlsl
@@ -12,6 +12,7 @@
// CHECK: define void @FirstEntry()
// CHECK-NEXT: entry:
+// CHECK: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// NOINLINE-NEXT: call void @_GLOBAL__sub_I_GlobalConstructorLib.hlsl()
// NOINLINE-NEXT: call void @FirstEntry()
// Verify inlining leaves only calls to "llvm." intrinsics
@@ -24,6 +25,7 @@
// CHECK: define void @SecondEntry()
// CHECK-NEXT: entry:
+// CHECK: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// NOINLINE-NEXT: call void @_GLOBAL__sub_I_GlobalConstructorLib.hlsl()
// NOINLINE-NEXT: call void @SecondEntry()
// Verify inlining leaves only calls to "llvm." intrinsics
diff --git a/clang/test/CodeGenHLSL/GlobalConstructors.hlsl b/clang/test/CodeGenHLSL/GlobalConstructors.hlsl
index 7b26dba..cd79c92 100644
--- a/clang/test/CodeGenHLSL/GlobalConstructors.hlsl
+++ b/clang/test/CodeGenHLSL/GlobalConstructors.hlsl
@@ -10,7 +10,8 @@
// CHECK-NOT:@llvm.global_dtors
//CHECK: define void @main()
//CHECK-NEXT: entry:
+//CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
//CHECK-NEXT: call void @_GLOBAL__sub_I_GlobalConstructors.hlsl()
-//CHECK-NEXT: %0 = call i32 @llvm.dx.flattened.thread.id.in.group()
-//CHECK-NEXT: call void @_Z4mainj(i32 %0)
+//CHECK-NEXT: %[[#THREAD_ID:]] = call i32 @llvm.dx.flattened.thread.id.in.group()
+//CHECK-NEXT: call void @_Z4mainj(i32 %[[#THREAD_ID]])
//CHECK-NEXT: ret void
diff --git a/clang/test/CodeGenHLSL/GlobalDestructors.hlsl b/clang/test/CodeGenHLSL/GlobalDestructors.hlsl
index 9d8c2e6..4b0fbe8 100644
--- a/clang/test/CodeGenHLSL/GlobalDestructors.hlsl
+++ b/clang/test/CodeGenHLSL/GlobalDestructors.hlsl
@@ -57,18 +57,18 @@
// CHECK: define void @main()
// CHECK-NEXT: entry:
+// CHECK: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// Verify destructor is emitted
-// NOINLINE-DXIL-NEXT: call void @_GLOBAL__sub_I_GlobalDestructors.hlsl()
-// NOINLINE-DXIL-NEXT: %0 = call i32 @llvm.dx.flattened.thread.id.in.group()
-// NOINLINE-DXIL-NEXT: call void @_Z4mainj(i32 %0)
-// NOINLINE-DXIL-NEXT: call void @_GLOBAL__D_a()
+// NOINLINE-DXIL-NEXT: call void @_GLOBAL__sub_I_GlobalDestructors.hlsl() [ "convergencectrl"(token %[[#C_ENTRY]]) ]
+// NOINLINE-DXIL-NEXT: %[[#THREAD_ID:]] = call i32 @llvm.dx.flattened.thread.id.in.group()
+// NOINLINE-DXIL-NEXT: call void @_Z4mainj(i32 %[[#THREAD_ID]]) [ "convergencectrl"(token %[[#C_ENTRY]]) ]
+// NOINLINE-DXIL-NEXT: call void @_GLOBAL__D_a() [ "convergencectrl"(token %[[#C_ENTRY]]) ]
// NOINLINE-DXIL-NEXT: ret void
-// NOINLINE-SPIRV-NEXT: %0 = call token @llvm.experimental.convergence.entry()
-// NOINLINE-SPIRV-NEXT: call spir_func void @_GLOBAL__sub_I_GlobalDestructors.hlsl() [ "convergencectrl"(token %0) ]
-// NOINLINE-SPIRV-NEXT: %1 = call i32 @llvm.spv.flattened.thread.id.in.group()
-// NOINLINE-SPIRV-NEXT: call spir_func void @_Z4mainj(i32 %1) [ "convergencectrl"(token %0) ]
-// NOINLINE-SPIRV-NEXT: call spir_func void @_GLOBAL__D_a() [ "convergencectrl"(token %0) ]
+// NOINLINE-SPIRV-NEXT: call spir_func void @_GLOBAL__sub_I_GlobalDestructors.hlsl() [ "convergencectrl"(token %[[#C_ENTRY]]) ]
+// NOINLINE-SPIRV-NEXT: %[[#THREAD_ID:]] = call i32 @llvm.spv.flattened.thread.id.in.group()
+// NOINLINE-SPIRV-NEXT: call spir_func void @_Z4mainj(i32 %[[#THREAD_ID]]) [ "convergencectrl"(token %[[#C_ENTRY]]) ]
+// NOINLINE-SPIRV-NEXT: call spir_func void @_GLOBAL__D_a() [ "convergencectrl"(token %[[#C_ENTRY]]) ]
// NOINLINE-SPIRV-NEXT: ret void
// Verify inlining leaves only calls to "llvm." intrinsics
@@ -80,15 +80,16 @@
// NOINLINE-DXIL: define internal void @_GLOBAL__D_a() [[IntAttr:\#[0-9]+]]
// NOINLINE-DXIL-NEXT: entry:
-// NOINLINE-DXIL-NEXT: call void @_ZN4TailD1Ev(ptr @_ZZ3WagvE1T)
-// NOINLINE-DXIL-NEXT: call void @_ZN6PupperD1Ev(ptr @GlobalPup)
+// NOINLINE-DXIL: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
+// NOINLINE-DXIL-NEXT: call void @_ZN4TailD1Ev(ptr @_ZZ3WagvE1T) [ "convergencectrl"(token %[[#C_ENTRY]]) ]
+// NOINLINE-DXIL-NEXT: call void @_ZN6PupperD1Ev(ptr @GlobalPup) [ "convergencectrl"(token %[[#C_ENTRY]]) ]
// NOINLINE-DXIL-NEXT: ret void
// NOINLINE-SPIRV: define internal spir_func void @_GLOBAL__D_a() [[IntAttr:\#[0-9]+]]
// NOINLINE-SPIRV-NEXT: entry:
-// NOINLINE-SPIRV-NEXT: %0 = call token @llvm.experimental.convergence.entry()
-// NOINLINE-SPIRV-NEXT: call spir_func void @_ZN4TailD1Ev(ptr addrspacecast (ptr addrspace(10) @_ZZ3WagvE1T to ptr)) [ "convergencectrl"(token %0) ]
-// NOINLINE-SPIRV-NEXT: call spir_func void @_ZN6PupperD1Ev(ptr @GlobalPup) [ "convergencectrl"(token %0) ]
+// NOINLINE-SPIRV-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
+// NOINLINE-SPIRV-NEXT: call spir_func void @_ZN4TailD1Ev(ptr addrspacecast (ptr addrspace(10) @_ZZ3WagvE1T to ptr)) [ "convergencectrl"(token %[[#C_ENTRY]]) ]
+// NOINLINE-SPIRV-NEXT: call spir_func void @_ZN6PupperD1Ev(ptr @GlobalPup) [ "convergencectrl"(token %[[#C_ENTRY]]) ]
// NOINLINE-SPIRV-NEXT: ret void
// NOINLINE: attributes [[IntAttr]] = {{.*}} alwaysinline
diff --git a/clang/test/CodeGenHLSL/builtins/AddUint64.hlsl b/clang/test/CodeGenHLSL/builtins/AddUint64.hlsl
index 5ec6720..ba866f9 100644
--- a/clang/test/CodeGenHLSL/builtins/AddUint64.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/AddUint64.hlsl
@@ -7,6 +7,7 @@
// CHECK-LABEL: define hidden noundef <2 x i32> @_Z20test_AddUint64_uint2Dv2_jS_(
// CHECK-SAME: <2 x i32> noundef [[A:%.*]], <2 x i32> noundef [[B:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca <2 x i32>, align 4
// CHECK-NEXT: [[B_ADDR:%.*]] = alloca <2 x i32>, align 4
// CHECK-NEXT: store <2 x i32> [[A]], ptr [[A_ADDR]], align 4
@@ -34,6 +35,7 @@
// CHECK-LABEL: define hidden noundef <4 x i32> @_Z20test_AddUint64_uint4Dv4_jS_(
// CHECK-SAME: <4 x i32> noundef [[A:%.*]], <4 x i32> noundef [[B:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca <4 x i32>, align 4
// CHECK-NEXT: [[B_ADDR:%.*]] = alloca <4 x i32>, align 4
// CHECK-NEXT: store <4 x i32> [[A]], ptr [[A_ADDR]], align 4
diff --git a/clang/test/CodeGenHLSL/builtins/ScalarSwizzles.hlsl b/clang/test/CodeGenHLSL/builtins/ScalarSwizzles.hlsl
index 484c3ba..0087580 100644
--- a/clang/test/CodeGenHLSL/builtins/ScalarSwizzles.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/ScalarSwizzles.hlsl
@@ -233,7 +233,7 @@
// CHECK: lor.end:
// CHECK-NEXT: [[H:%.*]] = phi i1 [ true, %entry ], [ [[G]], %lor.rhs ]
-// CHECK-NEXT: [[J:%.*]] = zext i1 %9 to i32
+// CHECK-NEXT: [[J:%.*]] = zext i1 [[H]] to i32
// CHECK-NEXT: store i32 [[J]], ptr [[XAddr]], align 4
// CHECK-NEXT: [[I:%.*]] = load i32, ptr [[XAddr]], align 4
// CHECK-NEXT: [[LoadV:%.*]] = trunc i32 [[I]] to i1
diff --git a/clang/test/CodeGenHLSL/builtins/abs.hlsl b/clang/test/CodeGenHLSL/builtins/abs.hlsl
index 45cc907..1eb8ff4 100644
--- a/clang/test/CodeGenHLSL/builtins/abs.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/abs.hlsl
@@ -53,7 +53,7 @@
// NATIVE_HALF-LABEL: define hidden noundef nofpclass(nan inf) half @_Z13test_abs_half
// NATIVE_HALF: call reassoc nnan ninf nsz arcp afn half @llvm.fabs.f16(
// NO_HALF-LABEL: define hidden noundef nofpclass(nan inf) float @_Z13test_abs_half
-// NO_HALF: call reassoc nnan ninf nsz arcp afn float @llvm.fabs.f32(float %0)
+// NO_HALF: call reassoc nnan ninf nsz arcp afn float @llvm.fabs.f32(float %[[#]])
half test_abs_half(half p0) { return abs(p0); }
// NATIVE_HALF-LABEL: define hidden noundef nofpclass(nan inf) <2 x half> @_Z14test_abs_half2
// NATIVE_HALF: call reassoc nnan ninf nsz arcp afn <2 x half> @llvm.fabs.v2f16(
diff --git a/clang/test/CodeGenHLSL/builtins/ceil.hlsl b/clang/test/CodeGenHLSL/builtins/ceil.hlsl
index d87d56e..bda9e35 100644
--- a/clang/test/CodeGenHLSL/builtins/ceil.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/ceil.hlsl
@@ -10,7 +10,7 @@
// NATIVE_HALF-LABEL: define hidden noundef nofpclass(nan inf) half @_Z14test_ceil_half
// NATIVE_HALF: call reassoc nnan ninf nsz arcp afn half @llvm.ceil.f16(
// NO_HALF-LABEL: define hidden noundef nofpclass(nan inf) float @_Z14test_ceil_half
-// NO_HALF: call reassoc nnan ninf nsz arcp afn float @llvm.ceil.f32(float %0)
+// NO_HALF: call reassoc nnan ninf nsz arcp afn float @llvm.ceil.f32(float %[[#]])
half test_ceil_half(half p0) { return ceil(p0); }
// NATIVE_HALF-LABEL: define hidden noundef nofpclass(nan inf) <2 x half> @_Z15test_ceil_half2
// NATIVE_HALF: call reassoc nnan ninf nsz arcp afn <2 x half> @llvm.ceil.v2f16(
diff --git a/clang/test/CodeGenHLSL/builtins/f16tof32-builtin.hlsl b/clang/test/CodeGenHLSL/builtins/f16tof32-builtin.hlsl
index e7be39a..e98a8f8 100644
--- a/clang/test/CodeGenHLSL/builtins/f16tof32-builtin.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/f16tof32-builtin.hlsl
@@ -6,7 +6,7 @@
// RUN: FileCheck %s --check-prefix=SPIRV
// CHECK: define hidden noundef nofpclass(nan inf) float
-// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn float @llvm.dx.legacyf16tof32.i32(i32 %0)
+// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn float @llvm.dx.legacyf16tof32.i32(i32 %[[#]])
// CHECK: ret float %hlsl.f16tof32
// CHECK: declare float @llvm.dx.legacyf16tof32.i32(i32)
//
@@ -18,7 +18,7 @@
float test_scalar(uint p0) { return __builtin_hlsl_elementwise_f16tof32(p0); }
// CHECK: define hidden noundef nofpclass(nan inf) <2 x float>
-// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn <2 x float> @llvm.dx.legacyf16tof32.v2i32(<2 x i32> %0)
+// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn <2 x float> @llvm.dx.legacyf16tof32.v2i32(<2 x i32> %[[#]])
// CHECK: ret <2 x float> %hlsl.f16tof32
// CHECK: declare <2 x float> @llvm.dx.legacyf16tof32.v2i32(<2 x i32>)
//
@@ -29,7 +29,7 @@
float2 test_uint2(uint2 p0) { return __builtin_hlsl_elementwise_f16tof32(p0); }
// CHECK: define hidden noundef nofpclass(nan inf) <3 x float> @_Z10test_uint3Dv3_j(<3 x i32> noundef %p0) #0 {
-// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn <3 x float> @llvm.dx.legacyf16tof32.v3i32(<3 x i32> %0)
+// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn <3 x float> @llvm.dx.legacyf16tof32.v3i32(<3 x i32> %[[#]])
// CHECK: ret <3 x float> %hlsl.f16tof32
// CHECK: declare <3 x float> @llvm.dx.legacyf16tof32.v3i32(<3 x i32>)
//
@@ -40,7 +40,7 @@
float3 test_uint3(uint3 p0) { return __builtin_hlsl_elementwise_f16tof32(p0); }
// CHECK: define hidden noundef nofpclass(nan inf) <4 x float> @_Z10test_uint4Dv4_j(<4 x i32> noundef %p0) #0 {
-// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn <4 x float> @llvm.dx.legacyf16tof32.v4i32(<4 x i32> %0)
+// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn <4 x float> @llvm.dx.legacyf16tof32.v4i32(<4 x i32> %[[#]])
// CHECK: ret <4 x float> %hlsl.f16tof32
// CHECK: declare <4 x float> @llvm.dx.legacyf16tof32.v4i32(<4 x i32>)
//
diff --git a/clang/test/CodeGenHLSL/builtins/f16tof32.hlsl b/clang/test/CodeGenHLSL/builtins/f16tof32.hlsl
index d4ab863..0e1d3ba 100644
--- a/clang/test/CodeGenHLSL/builtins/f16tof32.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/f16tof32.hlsl
@@ -6,7 +6,7 @@
// RUN: FileCheck %s --check-prefix=SPIRV
// CHECK: define hidden noundef nofpclass(nan inf) float
-// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn float @llvm.dx.legacyf16tof32.i32(i32 %0)
+// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn float @llvm.dx.legacyf16tof32.i32(i32 %[[#]])
// CHECK: ret float %hlsl.f16tof32
// CHECK: declare float @llvm.dx.legacyf16tof32.i32(i32)
//
@@ -18,7 +18,7 @@
float test_scalar(uint p0) { return f16tof32(p0); }
// CHECK: define hidden noundef nofpclass(nan inf) <2 x float>
-// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn <2 x float> @llvm.dx.legacyf16tof32.v2i32(<2 x i32> %0)
+// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn <2 x float> @llvm.dx.legacyf16tof32.v2i32(<2 x i32> %[[#]])
// CHECK: ret <2 x float> %hlsl.f16tof32
// CHECK: declare <2 x float> @llvm.dx.legacyf16tof32.v2i32(<2 x i32>)
//
@@ -29,7 +29,7 @@
float2 test_uint2(uint2 p0) { return f16tof32(p0); }
// CHECK: define hidden noundef nofpclass(nan inf) <3 x float> @_Z10test_uint3Dv3_j(<3 x i32> noundef %p0) #0 {
-// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn <3 x float> @llvm.dx.legacyf16tof32.v3i32(<3 x i32> %0)
+// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn <3 x float> @llvm.dx.legacyf16tof32.v3i32(<3 x i32> %[[#]])
// CHECK: ret <3 x float> %hlsl.f16tof32
// CHECK: declare <3 x float> @llvm.dx.legacyf16tof32.v3i32(<3 x i32>)
//
@@ -40,7 +40,7 @@
float3 test_uint3(uint3 p0) { return f16tof32(p0); }
// CHECK: define hidden noundef nofpclass(nan inf) <4 x float> @_Z10test_uint4Dv4_j(<4 x i32> noundef %p0) #0 {
-// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn <4 x float> @llvm.dx.legacyf16tof32.v4i32(<4 x i32> %0)
+// CHECK: %hlsl.f16tof32 = call reassoc nnan ninf nsz arcp afn <4 x float> @llvm.dx.legacyf16tof32.v4i32(<4 x i32> %[[#]])
// CHECK: ret <4 x float> %hlsl.f16tof32
// CHECK: declare <4 x float> @llvm.dx.legacyf16tof32.v4i32(<4 x i32>)
//
diff --git a/clang/test/CodeGenHLSL/builtins/f32tof16-builtin.hlsl b/clang/test/CodeGenHLSL/builtins/f32tof16-builtin.hlsl
index da46af1..c4b9033 100644
--- a/clang/test/CodeGenHLSL/builtins/f32tof16-builtin.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/f32tof16-builtin.hlsl
@@ -6,9 +6,9 @@
// RUN: FileCheck %s --check-prefix=SPIRV
// CHECK: define hidden noundef i32 @_Z11test_scalarf(float noundef nofpclass(nan inf) %p0) #0 {
-// CHECK: %hlsl.f32tof16 = call i32 @llvm.dx.legacyf32tof16.f32(float %0)
+// CHECK: %hlsl.f32tof16 = call i32 @llvm.dx.legacyf32tof16.f32(float %[[#]])
// CHECK: ret i32 %hlsl.f32tof16
-// CHECK: declare i32 @llvm.dx.legacyf32tof16.f32(float) #1
+// CHECK: declare i32 @llvm.dx.legacyf32tof16.f32(float)
//
// SPIRV: define hidden spir_func noundef i32 @_Z11test_scalarf(float noundef nofpclass(nan inf) %p0) #0 {
// SPIRV: call i32 @llvm.spv.packhalf2x16.i32.v2f32(<2 x float> %[[#]])
@@ -18,9 +18,9 @@
uint test_scalar(float p0) { return __builtin_hlsl_elementwise_f32tof16(p0); }
// CHECK: define hidden noundef <2 x i32> @_Z10test_uint2Dv2_f(<2 x float> noundef nofpclass(nan inf) %p0) #0 {
-// CHECK: %hlsl.f32tof16 = call <2 x i32> @llvm.dx.legacyf32tof16.v2f32(<2 x float> %0)
+// CHECK: %hlsl.f32tof16 = call <2 x i32> @llvm.dx.legacyf32tof16.v2f32(<2 x float> %[[#]])
// CHECK: ret <2 x i32> %hlsl.f32tof16
-// CHECK: declare <2 x i32> @llvm.dx.legacyf32tof16.v2f32(<2 x float>) #1
+// CHECK: declare <2 x i32> @llvm.dx.legacyf32tof16.v2f32(<2 x float>)
//
// SPIRV: define hidden spir_func noundef <2 x i32> @_Z10test_uint2Dv2_f(<2 x float> noundef nofpclass(nan inf) %p0) #0 {
// SPIRV-COUNT-2: call i32 @llvm.spv.packhalf2x16.i32.v2f32(<2 x float> %[[#]])
@@ -29,9 +29,9 @@
uint2 test_uint2(float2 p0) { return __builtin_hlsl_elementwise_f32tof16(p0); }
// CHECK: define hidden noundef <3 x i32> @_Z10test_uint3Dv3_f(<3 x float> noundef nofpclass(nan inf) %p0) #0 {
-// CHECK: %hlsl.f32tof16 = call <3 x i32> @llvm.dx.legacyf32tof16.v3f32(<3 x float> %0)
+// CHECK: %hlsl.f32tof16 = call <3 x i32> @llvm.dx.legacyf32tof16.v3f32(<3 x float> %[[#]])
// CHECK: ret <3 x i32> %hlsl.f32tof16
-// CHECK: declare <3 x i32> @llvm.dx.legacyf32tof16.v3f32(<3 x float>) #1
+// CHECK: declare <3 x i32> @llvm.dx.legacyf32tof16.v3f32(<3 x float>)
//
// SPIRV: define hidden spir_func noundef <3 x i32> @_Z10test_uint3Dv3_f(<3 x float> noundef nofpclass(nan inf) %p0) #0 {
// SPIRV-COUNT-3: call i32 @llvm.spv.packhalf2x16.i32.v2f32(<2 x float> %[[#]])
@@ -40,9 +40,9 @@
uint3 test_uint3(float3 p0) { return __builtin_hlsl_elementwise_f32tof16(p0); }
// CHECK: define hidden noundef <4 x i32> @_Z10test_uint4Dv4_f(<4 x float> noundef nofpclass(nan inf) %p0) #0 {
-// CHECK: %hlsl.f32tof16 = call <4 x i32> @llvm.dx.legacyf32tof16.v4f32(<4 x float> %0)
+// CHECK: %hlsl.f32tof16 = call <4 x i32> @llvm.dx.legacyf32tof16.v4f32(<4 x float> %[[#]])
// CHECK: ret <4 x i32> %hlsl.f32tof16
-// CHECK: declare <4 x i32> @llvm.dx.legacyf32tof16.v4f32(<4 x float>) #1
+// CHECK: declare <4 x i32> @llvm.dx.legacyf32tof16.v4f32(<4 x float>)
//
// SPIRV: define hidden spir_func noundef <4 x i32> @_Z10test_uint4Dv4_f(<4 x float> noundef nofpclass(nan inf) %p0) #0 {
// SPIRV-COUNT-4: call i32 @llvm.spv.packhalf2x16.i32.v2f32(<2 x float> %[[#]])
diff --git a/clang/test/CodeGenHLSL/builtins/f32tof16.hlsl b/clang/test/CodeGenHLSL/builtins/f32tof16.hlsl
index 93bcea3..551f30e 100644
--- a/clang/test/CodeGenHLSL/builtins/f32tof16.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/f32tof16.hlsl
@@ -6,9 +6,9 @@
// RUN: FileCheck %s --check-prefix=SPIRV
// CHECK: define hidden noundef i32 @_Z11test_scalarf(float noundef nofpclass(nan inf) %p0) #0 {
-// CHECK: %hlsl.f32tof16 = call i32 @llvm.dx.legacyf32tof16.f32(float %0)
+// CHECK: %hlsl.f32tof16 = call i32 @llvm.dx.legacyf32tof16.f32(float %[[#]])
// CHECK: ret i32 %hlsl.f32tof16
-// CHECK: declare i32 @llvm.dx.legacyf32tof16.f32(float) #1
+// CHECK: declare i32 @llvm.dx.legacyf32tof16.f32(float)
//
// SPIRV: define hidden spir_func noundef i32 @_Z11test_scalarf(float noundef nofpclass(nan inf) %p0) #0 {
// SPIRV: call i32 @llvm.spv.packhalf2x16.i32.v2f32(<2 x float> %[[#]])
@@ -18,9 +18,9 @@
uint test_scalar(float p0) { return f32tof16(p0); }
// CHECK: define hidden noundef <2 x i32> @_Z10test_uint2Dv2_f(<2 x float> noundef nofpclass(nan inf) %p0) #0 {
-// CHECK: %hlsl.f32tof16 = call <2 x i32> @llvm.dx.legacyf32tof16.v2f32(<2 x float> %0)
+// CHECK: %hlsl.f32tof16 = call <2 x i32> @llvm.dx.legacyf32tof16.v2f32(<2 x float> %[[#]])
// CHECK: ret <2 x i32> %hlsl.f32tof16
-// CHECK: declare <2 x i32> @llvm.dx.legacyf32tof16.v2f32(<2 x float>) #1
+// CHECK: declare <2 x i32> @llvm.dx.legacyf32tof16.v2f32(<2 x float>)
//
// SPIRV: define hidden spir_func noundef <2 x i32> @_Z10test_uint2Dv2_f(<2 x float> noundef nofpclass(nan inf) %p0) #0 {
// SPIRV-COUNT-2: call i32 @llvm.spv.packhalf2x16.i32.v2f32(<2 x float> %[[#]])
@@ -29,9 +29,9 @@
uint2 test_uint2(float2 p0) { return f32tof16(p0); }
// CHECK: define hidden noundef <3 x i32> @_Z10test_uint3Dv3_f(<3 x float> noundef nofpclass(nan inf) %p0) #0 {
-// CHECK: %hlsl.f32tof16 = call <3 x i32> @llvm.dx.legacyf32tof16.v3f32(<3 x float> %0)
+// CHECK: %hlsl.f32tof16 = call <3 x i32> @llvm.dx.legacyf32tof16.v3f32(<3 x float> %[[#]])
// CHECK: ret <3 x i32> %hlsl.f32tof16
-// CHECK: declare <3 x i32> @llvm.dx.legacyf32tof16.v3f32(<3 x float>) #1
+// CHECK: declare <3 x i32> @llvm.dx.legacyf32tof16.v3f32(<3 x float>)
//
// SPIRV: define hidden spir_func noundef <3 x i32> @_Z10test_uint3Dv3_f(<3 x float> noundef nofpclass(nan inf) %p0) #0 {
// SPIRV-COUNT-3: call i32 @llvm.spv.packhalf2x16.i32.v2f32(<2 x float> %[[#]])
@@ -40,9 +40,9 @@
uint3 test_uint3(float3 p0) { return f32tof16(p0); }
// CHECK: define hidden noundef <4 x i32> @_Z10test_uint4Dv4_f(<4 x float> noundef nofpclass(nan inf) %p0) #0 {
-// CHECK: %hlsl.f32tof16 = call <4 x i32> @llvm.dx.legacyf32tof16.v4f32(<4 x float> %0)
+// CHECK: %hlsl.f32tof16 = call <4 x i32> @llvm.dx.legacyf32tof16.v4f32(<4 x float> %[[#]])
// CHECK: ret <4 x i32> %hlsl.f32tof16
-// CHECK: declare <4 x i32> @llvm.dx.legacyf32tof16.v4f32(<4 x float>) #1
+// CHECK: declare <4 x i32> @llvm.dx.legacyf32tof16.v4f32(<4 x float>)
//
// SPIRV: define hidden spir_func noundef <4 x i32> @_Z10test_uint4Dv4_f(<4 x float> noundef nofpclass(nan inf) %p0) #0 {
// SPIRV-COUNT-4: call i32 @llvm.spv.packhalf2x16.i32.v2f32(<2 x float> %[[#]])
diff --git a/clang/test/CodeGenHLSL/builtins/floor.hlsl b/clang/test/CodeGenHLSL/builtins/floor.hlsl
index 4763e54..9a91545 100644
--- a/clang/test/CodeGenHLSL/builtins/floor.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/floor.hlsl
@@ -10,7 +10,7 @@
// NATIVE_HALF-LABEL: define hidden noundef nofpclass(nan inf) half @_Z15test_floor_half
// NATIVE_HALF: call reassoc nnan ninf nsz arcp afn half @llvm.floor.f16(
// NO_HALF-LABEL: define hidden noundef nofpclass(nan inf) float @_Z15test_floor_half
-// NO_HALF: call reassoc nnan ninf nsz arcp afn float @llvm.floor.f32(float %0)
+// NO_HALF: call reassoc nnan ninf nsz arcp afn float @llvm.floor.f32(float %[[#]])
half test_floor_half(half p0) { return floor(p0); }
// NATIVE_HALF-LABEL: define hidden noundef nofpclass(nan inf) <2 x half> @_Z16test_floor_half2
// NATIVE_HALF: call reassoc nnan ninf nsz arcp afn <2 x half> @llvm.floor.v2f16(
diff --git a/clang/test/CodeGenHLSL/builtins/mad.hlsl b/clang/test/CodeGenHLSL/builtins/mad.hlsl
index f501445..82ec276 100644
--- a/clang/test/CodeGenHLSL/builtins/mad.hlsl
+++ b/clang/test/CodeGenHLSL/builtins/mad.hlsl
@@ -15,49 +15,49 @@
// RUN: -o - | FileCheck %s --check-prefixes=CHECK,NO_HALF,SPIR_CHECK
#ifdef __HLSL_ENABLE_16_BIT
-// DXIL_NATIVE_HALF: %dx.umad = call i16 @llvm.dx.umad.i16(i16 %0, i16 %1, i16 %2)
+// DXIL_NATIVE_HALF: %dx.umad = call i16 @llvm.dx.umad.i16(i16 %[[#]], i16 %[[#]], i16 %[[#]])
// DXIL_NATIVE_HALF: ret i16 %dx.umad
// SPIR_NATIVE_HALF: mul nuw i16 %{{.*}}, %{{.*}}
// SPIR_NATIVE_HALF: add nuw i16 %{{.*}}, %{{.*}}
uint16_t test_mad_uint16_t(uint16_t p0, uint16_t p1, uint16_t p2) { return mad(p0, p1, p2); }
-// DXIL_NATIVE_HALF: %dx.umad = call <2 x i16> @llvm.dx.umad.v2i16(<2 x i16> %0, <2 x i16> %1, <2 x i16> %2)
+// DXIL_NATIVE_HALF: %dx.umad = call <2 x i16> @llvm.dx.umad.v2i16(<2 x i16> %[[#]], <2 x i16> %[[#]], <2 x i16> %[[#]])
// DXIL_NATIVE_HALF: ret <2 x i16> %dx.umad
// SPIR_NATIVE_HALF: mul nuw <2 x i16> %{{.*}}, %{{.*}}
// SPIR_NATIVE_HALF: add nuw <2 x i16> %{{.*}}, %{{.*}}
uint16_t2 test_mad_uint16_t2(uint16_t2 p0, uint16_t2 p1, uint16_t2 p2) { return mad(p0, p1, p2); }
-// DXIL_NATIVE_HALF: %dx.umad = call <3 x i16> @llvm.dx.umad.v3i16(<3 x i16> %0, <3 x i16> %1, <3 x i16> %2)
+// DXIL_NATIVE_HALF: %dx.umad = call <3 x i16> @llvm.dx.umad.v3i16(<3 x i16> %[[#]], <3 x i16> %[[#]], <3 x i16> %[[#]])
// DXIL_NATIVE_HALF: ret <3 x i16> %dx.umad
// SPIR_NATIVE_HALF: mul nuw <3 x i16> %{{.*}}, %{{.*}}
// SPIR_NATIVE_HALF: add nuw <3 x i16> %{{.*}}, %{{.*}}
uint16_t3 test_mad_uint16_t3(uint16_t3 p0, uint16_t3 p1, uint16_t3 p2) { return mad(p0, p1, p2); }
-// DXIL_NATIVE_HALF: %dx.umad = call <4 x i16> @llvm.dx.umad.v4i16(<4 x i16> %0, <4 x i16> %1, <4 x i16> %2)
+// DXIL_NATIVE_HALF: %dx.umad = call <4 x i16> @llvm.dx.umad.v4i16(<4 x i16> %[[#]], <4 x i16> %[[#]], <4 x i16> %[[#]])
// DXIL_NATIVE_HALF: ret <4 x i16> %dx.umad
// SPIR_NATIVE_HALF: mul nuw <4 x i16> %{{.*}}, %{{.*}}
// SPIR_NATIVE_HALF: add nuw <4 x i16> %{{.*}}, %{{.*}}
uint16_t4 test_mad_uint16_t4(uint16_t4 p0, uint16_t4 p1, uint16_t4 p2) { return mad(p0, p1, p2); }
-// DXIL_NATIVE_HALF: %dx.imad = call i16 @llvm.dx.imad.i16(i16 %0, i16 %1, i16 %2)
+// DXIL_NATIVE_HALF: %dx.imad = call i16 @llvm.dx.imad.i16(i16 %[[#]], i16 %[[#]], i16 %[[#]])
// DXIL_NATIVE_HALF: ret i16 %dx.imad
// SPIR_NATIVE_HALF: mul nsw i16 %{{.*}}, %{{.*}}
// SPIR_NATIVE_HALF: add nsw i16 %{{.*}}, %{{.*}}
int16_t test_mad_int16_t(int16_t p0, int16_t p1, int16_t p2) { return mad(p0, p1, p2); }
-// DXIL_NATIVE_HALF: %dx.imad = call <2 x i16> @llvm.dx.imad.v2i16(<2 x i16> %0, <2 x i16> %1, <2 x i16> %2)
+// DXIL_NATIVE_HALF: %dx.imad = call <2 x i16> @llvm.dx.imad.v2i16(<2 x i16> %[[#]], <2 x i16> %[[#]], <2 x i16> %[[#]])
// DXIL_NATIVE_HALF: ret <2 x i16> %dx.imad
// SPIR_NATIVE_HALF: mul nsw <2 x i16> %{{.*}}, %{{.*}}
// SPIR_NATIVE_HALF: add nsw <2 x i16> %{{.*}}, %{{.*}}
int16_t2 test_mad_int16_t2(int16_t2 p0, int16_t2 p1, int16_t2 p2) { return mad(p0, p1, p2); }
-// DXIL_NATIVE_HALF: %dx.imad = call <3 x i16> @llvm.dx.imad.v3i16(<3 x i16> %0, <3 x i16> %1, <3 x i16> %2)
+// DXIL_NATIVE_HALF: %dx.imad = call <3 x i16> @llvm.dx.imad.v3i16(<3 x i16> %[[#]], <3 x i16> %[[#]], <3 x i16> %[[#]])
// DXIL_NATIVE_HALF: ret <3 x i16> %dx.imad
// SPIR_NATIVE_HALF: mul nsw <3 x i16> %{{.*}}, %{{.*}}
// SPIR_NATIVE_HALF: add nsw <3 x i16> %{{.*}}, %{{.*}}
int16_t3 test_mad_int16_t3(int16_t3 p0, int16_t3 p1, int16_t3 p2) { return mad(p0, p1, p2); }
-// DXIL_NATIVE_HALF: %dx.imad = call <4 x i16> @llvm.dx.imad.v4i16(<4 x i16> %0, <4 x i16> %1, <4 x i16> %2)
+// DXIL_NATIVE_HALF: %dx.imad = call <4 x i16> @llvm.dx.imad.v4i16(<4 x i16> %[[#]], <4 x i16> %[[#]], <4 x i16> %[[#]])
// DXIL_NATIVE_HALF: ret <4 x i16> %dx.imad
// SPIR_NATIVE_HALF: mul nsw <4 x i16> %{{.*}}, %{{.*}}
// SPIR_NATIVE_HALF: add nsw <4 x i16> %{{.*}}, %{{.*}}
@@ -168,97 +168,97 @@
// CHECK: ret <4 x double> %hlsl.fmad
double4 test_mad_double4(double4 p0, double4 p1, double4 p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.imad = call i32 @llvm.dx.imad.i32(i32 %0, i32 %1, i32 %2)
+// DXIL_CHECK: %dx.imad = call i32 @llvm.dx.imad.i32(i32 %[[#]], i32 %[[#]], i32 %[[#]])
// DXIL_CHECK: ret i32 %dx.imad
// SPIR_CHECK: mul nsw i32 %{{.*}}, %{{.*}}
// SPIR_CHECK: add nsw i32 %{{.*}}, %{{.*}}
int test_mad_int(int p0, int p1, int p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.imad = call <2 x i32> @llvm.dx.imad.v2i32(<2 x i32> %0, <2 x i32> %1, <2 x i32> %2)
+// DXIL_CHECK: %dx.imad = call <2 x i32> @llvm.dx.imad.v2i32(<2 x i32> %[[#]], <2 x i32> %[[#]], <2 x i32> %[[#]])
// DXIL_CHECK: ret <2 x i32> %dx.imad
// SPIR_CHECK: mul nsw <2 x i32> %{{.*}}, %{{.*}}
// SPIR_CHECK: add nsw <2 x i32> %{{.*}}, %{{.*}}
int2 test_mad_int2(int2 p0, int2 p1, int2 p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.imad = call <3 x i32> @llvm.dx.imad.v3i32(<3 x i32> %0, <3 x i32> %1, <3 x i32> %2)
+// DXIL_CHECK: %dx.imad = call <3 x i32> @llvm.dx.imad.v3i32(<3 x i32> %[[#]], <3 x i32> %[[#]], <3 x i32> %[[#]])
// DXIL_CHECK: ret <3 x i32> %dx.imad
// SPIR_CHECK: mul nsw <3 x i32> %{{.*}}, %{{.*}}
// SPIR_CHECK: add nsw <3 x i32> %{{.*}}, %{{.*}}
int3 test_mad_int3(int3 p0, int3 p1, int3 p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.imad = call <4 x i32> @llvm.dx.imad.v4i32(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2)
+// DXIL_CHECK: %dx.imad = call <4 x i32> @llvm.dx.imad.v4i32(<4 x i32> %[[#]], <4 x i32> %[[#]], <4 x i32> %[[#]])
// DXIL_CHECK: ret <4 x i32> %dx.imad
// SPIR_CHECK: mul nsw <4 x i32> %{{.*}}, %{{.*}}
// SPIR_CHECK: add nsw <4 x i32> %{{.*}}, %{{.*}}
int4 test_mad_int4(int4 p0, int4 p1, int4 p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.imad = call i64 @llvm.dx.imad.i64(i64 %0, i64 %1, i64 %2)
+// DXIL_CHECK: %dx.imad = call i64 @llvm.dx.imad.i64(i64 %[[#]], i64 %[[#]], i64 %[[#]])
// DXIL_CHECK: ret i64 %dx.imad
// SPIR_CHECK: mul nsw i64 %{{.*}}, %{{.*}}
// SPIR_CHECK: add nsw i64 %{{.*}}, %{{.*}}
int64_t test_mad_int64_t(int64_t p0, int64_t p1, int64_t p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.imad = call <2 x i64> @llvm.dx.imad.v2i64(<2 x i64> %0, <2 x i64> %1, <2 x i64> %2)
+// DXIL_CHECK: %dx.imad = call <2 x i64> @llvm.dx.imad.v2i64(<2 x i64> %[[#]], <2 x i64> %[[#]], <2 x i64> %[[#]])
// DXIL_CHECK: ret <2 x i64> %dx.imad
// SPIR_CHECK: mul nsw <2 x i64> %{{.*}}, %{{.*}}
// SPIR_CHECK: add nsw <2 x i64> %{{.*}}, %{{.*}}
int64_t2 test_mad_int64_t2(int64_t2 p0, int64_t2 p1, int64_t2 p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.imad = call <3 x i64> @llvm.dx.imad.v3i64(<3 x i64> %0, <3 x i64> %1, <3 x i64> %2)
+// DXIL_CHECK: %dx.imad = call <3 x i64> @llvm.dx.imad.v3i64(<3 x i64> %[[#]], <3 x i64> %[[#]], <3 x i64> %[[#]])
// DXIL_CHECK: ret <3 x i64> %dx.imad
// SPIR_CHECK: mul nsw <3 x i64> %{{.*}}, %{{.*}}
// SPIR_CHECK: add nsw <3 x i64> %{{.*}}, %{{.*}}
int64_t3 test_mad_int64_t3(int64_t3 p0, int64_t3 p1, int64_t3 p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.imad = call <4 x i64> @llvm.dx.imad.v4i64(<4 x i64> %0, <4 x i64> %1, <4 x i64> %2)
+// DXIL_CHECK: %dx.imad = call <4 x i64> @llvm.dx.imad.v4i64(<4 x i64> %[[#]], <4 x i64> %[[#]], <4 x i64> %[[#]])
// DXIL_CHECK: ret <4 x i64> %dx.imad
// SPIR_CHECK: mul nsw <4 x i64> %{{.*}}, %{{.*}}
// SPIR_CHECK: add nsw <4 x i64> %{{.*}}, %{{.*}}
int64_t4 test_mad_int64_t4(int64_t4 p0, int64_t4 p1, int64_t4 p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.umad = call i32 @llvm.dx.umad.i32(i32 %0, i32 %1, i32 %2)
+// DXIL_CHECK: %dx.umad = call i32 @llvm.dx.umad.i32(i32 %[[#]], i32 %[[#]], i32 %[[#]])
// DXIL_CHECK: ret i32 %dx.umad
// SPIR_CHECK: mul nuw i32 %{{.*}}, %{{.*}}
// SPIR_CHECK: add nuw i32 %{{.*}}, %{{.*}}
uint test_mad_uint(uint p0, uint p1, uint p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.umad = call <2 x i32> @llvm.dx.umad.v2i32(<2 x i32> %0, <2 x i32> %1, <2 x i32> %2)
+// DXIL_CHECK: %dx.umad = call <2 x i32> @llvm.dx.umad.v2i32(<2 x i32> %[[#]], <2 x i32> %[[#]], <2 x i32> %[[#]])
// DXIL_CHECK: ret <2 x i32> %dx.umad
// SPIR_CHECK: mul nuw <2 x i32> %{{.*}}, %{{.*}}
// SPIR_CHECK: add nuw <2 x i32> %{{.*}}, %{{.*}}
uint2 test_mad_uint2(uint2 p0, uint2 p1, uint2 p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.umad = call <3 x i32> @llvm.dx.umad.v3i32(<3 x i32> %0, <3 x i32> %1, <3 x i32> %2)
+// DXIL_CHECK: %dx.umad = call <3 x i32> @llvm.dx.umad.v3i32(<3 x i32> %[[#]], <3 x i32> %[[#]], <3 x i32> %[[#]])
// DXIL_CHECK: ret <3 x i32> %dx.umad
// SPIR_CHECK: mul nuw <3 x i32> %{{.*}}, %{{.*}}
// SPIR_CHECK: add nuw <3 x i32> %{{.*}}, %{{.*}}
uint3 test_mad_uint3(uint3 p0, uint3 p1, uint3 p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.umad = call <4 x i32> @llvm.dx.umad.v4i32(<4 x i32> %0, <4 x i32> %1, <4 x i32> %2)
+// DXIL_CHECK: %dx.umad = call <4 x i32> @llvm.dx.umad.v4i32(<4 x i32> %[[#]], <4 x i32> %[[#]], <4 x i32> %[[#]])
// DXIL_CHECK: ret <4 x i32> %dx.umad
// SPIR_CHECK: mul nuw <4 x i32> %{{.*}}, %{{.*}}
// SPIR_CHECK: add nuw <4 x i32> %{{.*}}, %{{.*}}
uint4 test_mad_uint4(uint4 p0, uint4 p1, uint4 p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.umad = call i64 @llvm.dx.umad.i64(i64 %0, i64 %1, i64 %2)
+// DXIL_CHECK: %dx.umad = call i64 @llvm.dx.umad.i64(i64 %[[#]], i64 %[[#]], i64 %[[#]])
// DXIL_CHECK: ret i64 %dx.umad
// SPIR_CHECK: mul nuw i64 %{{.*}}, %{{.*}}
// SPIR_CHECK: add nuw i64 %{{.*}}, %{{.*}}
uint64_t test_mad_uint64_t(uint64_t p0, uint64_t p1, uint64_t p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.umad = call <2 x i64> @llvm.dx.umad.v2i64(<2 x i64> %0, <2 x i64> %1, <2 x i64> %2)
+// DXIL_CHECK: %dx.umad = call <2 x i64> @llvm.dx.umad.v2i64(<2 x i64> %[[#]], <2 x i64> %[[#]], <2 x i64> %[[#]])
// DXIL_CHECK: ret <2 x i64> %dx.umad
// SPIR_CHECK: mul nuw <2 x i64> %{{.*}}, %{{.*}}
// SPIR_CHECK: add nuw <2 x i64> %{{.*}}, %{{.*}}
uint64_t2 test_mad_uint64_t2(uint64_t2 p0, uint64_t2 p1, uint64_t2 p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.umad = call <3 x i64> @llvm.dx.umad.v3i64(<3 x i64> %0, <3 x i64> %1, <3 x i64> %2)
+// DXIL_CHECK: %dx.umad = call <3 x i64> @llvm.dx.umad.v3i64(<3 x i64> %[[#]], <3 x i64> %[[#]], <3 x i64> %[[#]])
// DXIL_CHECK: ret <3 x i64> %dx.umad
// SPIR_CHECK: mul nuw <3 x i64> %{{.*}}, %{{.*}}
// SPIR_CHECK: add nuw <3 x i64> %{{.*}}, %{{.*}}
uint64_t3 test_mad_uint64_t3(uint64_t3 p0, uint64_t3 p1, uint64_t3 p2) { return mad(p0, p1, p2); }
-// DXIL_CHECK: %dx.umad = call <4 x i64> @llvm.dx.umad.v4i64(<4 x i64> %0, <4 x i64> %1, <4 x i64> %2)
+// DXIL_CHECK: %dx.umad = call <4 x i64> @llvm.dx.umad.v4i64(<4 x i64> %[[#]], <4 x i64> %[[#]], <4 x i64> %[[#]])
// DXIL_CHECK: ret <4 x i64> %dx.umad
// SPIR_CHECK: mul nuw <4 x i64> %{{.*}}, %{{.*}}
// SPIR_CHECK: add nuw <4 x i64> %{{.*}}, %{{.*}}
diff --git a/clang/test/CodeGenHLSL/convergence/cf.for.plain.hlsl b/clang/test/CodeGenHLSL/convergence/cf.for.plain.hlsl
index 2f08854..ea27d9d 100644
--- a/clang/test/CodeGenHLSL/convergence/cf.for.plain.hlsl
+++ b/clang/test/CodeGenHLSL/convergence/cf.for.plain.hlsl
@@ -1,5 +1,7 @@
// RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -x hlsl -triple \
// RUN: spirv-pc-vulkan-library %s -emit-llvm -disable-llvm-passes -o - | FileCheck %s
+// RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -x hlsl -triple \
+// RUN: dxil-pc-shadermodel6.3-library %s -emit-llvm -disable-llvm-passes -o - | FileCheck %s
int process() {
// CHECK: entry:
diff --git a/clang/test/CodeGenHLSL/convergence/do.while.hlsl b/clang/test/CodeGenHLSL/convergence/do.while.hlsl
index 9aabbfd..f196600 100644
--- a/clang/test/CodeGenHLSL/convergence/do.while.hlsl
+++ b/clang/test/CodeGenHLSL/convergence/do.while.hlsl
@@ -1,5 +1,7 @@
// RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -x hlsl -triple \
// RUN: spirv-pc-vulkan-library %s -emit-llvm -disable-llvm-passes -o - | FileCheck %s
+// RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -x hlsl -triple \
+// RUN: dxil-pc-shadermodel6.3-library %s -emit-llvm -disable-llvm-passes -o - | FileCheck %s
bool cond();
void foo();
@@ -8,29 +10,29 @@
do {
} while (cond());
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test1v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test1v()
// CHECK-SAME: [[A0:#[0-9]+]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
// CHECK: do.body:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
// CHECK: do.cond:
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3:#[0-9]+]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3:#[0-9]+]] [ "convergencectrl"(token [[T1]]) ]
void test2() {
do {
foo();
} while (cond());
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test2v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test2v()
// CHECK-SAME: [[A0]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
// CHECK: do.body:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: do.cond:
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
void test3() {
do {
@@ -38,16 +40,16 @@
foo();
} while (cond());
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test3v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test3v()
// CHECK-SAME: [[A0]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
// CHECK: do.body:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
// CHECK: if.then:
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: do.cond:
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
void test4() {
do {
@@ -57,16 +59,16 @@
}
} while (cond());
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test4v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test4v()
// CHECK-SAME: [[A0]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
// CHECK: do.body:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
// CHECK: if.then:
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: do.cond:
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
void test5() {
do {
@@ -78,7 +80,7 @@
}
} while (cond());
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test5v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test5v()
// CHECK-SAME: [[A0]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
@@ -87,9 +89,9 @@
// CHECK: while.cond:
// CHECK: [[T2:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T1]]) ]
// CHECK: if.then:
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T2]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T2]]) ]
// CHECK: do.cond:
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK-DAG: attributes [[A0]] = { {{.*}}convergent{{.*}} }
// CHECK-DAG: attributes [[A3]] = { {{.*}}convergent{{.*}} }
diff --git a/clang/test/CodeGenHLSL/convergence/entry.point.hlsl b/clang/test/CodeGenHLSL/convergence/entry.point.hlsl
index 337a9ad..9a8ae83 100644
--- a/clang/test/CodeGenHLSL/convergence/entry.point.hlsl
+++ b/clang/test/CodeGenHLSL/convergence/entry.point.hlsl
@@ -1,9 +1,10 @@
// RUN: %clang_cc1 -triple spirv-pc-vulkan-compute -finclude-default-header -disable-llvm-passes -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.0-compute -finclude-default-header -disable-llvm-passes -emit-llvm -o - %s | FileCheck %s
// CHECK-LABEL: define void @main()
// CHECK-NEXT: entry:
-// CHECK-NEXT: [[token:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
-// CHECK-NEXT: call spir_func void @_Z4mainv() [ "convergencectrl"(token [[token]]) ]
+// CHECK-NEXT: %[[#TOKEN:]] = call token @llvm.experimental.convergence.entry()
+// CHECK-NEXT: call{{.*}}void @_Z4mainv() [ "convergencectrl"(token %[[#TOKEN]]) ]
[numthreads(1,1,1)]
void main() {
diff --git a/clang/test/CodeGenHLSL/convergence/for.hlsl b/clang/test/CodeGenHLSL/convergence/for.hlsl
index b7b11e9..c4c1770 100644
--- a/clang/test/CodeGenHLSL/convergence/for.hlsl
+++ b/clang/test/CodeGenHLSL/convergence/for.hlsl
@@ -1,5 +1,7 @@
// RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -x hlsl -triple \
// RUN: spirv-pc-vulkan-library %s -emit-llvm -disable-llvm-passes -o - | FileCheck %s
+// RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -x hlsl -triple \
+// RUN: dxil-pc-shadermodel6.3-library %s -emit-llvm -disable-llvm-passes -o - | FileCheck %s
bool cond();
bool cond2();
@@ -10,73 +12,73 @@
foo();
}
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test1v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test1v()
// CHECK-SAME: [[A0:#[0-9]+]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
// CHECK: for.cond:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
-// CHECK: call spir_func void @_Z3foov() [[A3:#[0-9]+]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3:#[0-9]+]] [ "convergencectrl"(token [[T1]]) ]
void test2() {
for (;cond();) {
foo();
}
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test2v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test2v()
// CHECK-SAME: [[A0]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
// CHECK: for.cond:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: for.body:
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
void test3() {
for (cond();;) {
foo();
}
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test3v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test3v()
// CHECK-SAME: [[A0]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T0]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T0]]) ]
// CHECK: for.cond:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
void test4() {
for (cond();cond2();) {
foo();
}
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test4v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test4v()
// CHECK-SAME: [[A0]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T0]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T0]]) ]
// CHECK: for.cond:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
-// CHECK: call spir_func noundef i1 @_Z5cond2v() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z5cond2v() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: for.body:
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
void test5() {
for (cond();cond2();foo()) {
}
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test5v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test5v()
// CHECK-SAME: [[A0]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T0]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T0]]) ]
// CHECK: for.cond:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
-// CHECK: call spir_func noundef i1 @_Z5cond2v() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z5cond2v() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: for.inc:
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
void test6() {
for (cond();cond2();foo()) {
@@ -86,24 +88,24 @@
}
}
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test6v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test6v()
// CHECK-SAME: [[A0]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T0]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T0]]) ]
// CHECK: for.cond:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
-// CHECK: call spir_func noundef i1 @_Z5cond2v() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z5cond2v() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: for.body:
-// CHECK: [[C1:%[a-zA-Z0-9]+]] = call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: [[C1:%[a-zA-Z0-9]+]] = call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: br i1 [[C1]], label %if.then, label %if.end
// CHECK: if.then:
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: br label %for.end
// CHECK: if.end:
// CHECK: br label %for.inc
// CHECK: for.inc:
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
void test7() {
for (cond();;) {
@@ -112,17 +114,17 @@
}
}
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test7v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test7v()
// CHECK-SAME: [[A0]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T0]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T0]]) ]
// CHECK: for.cond:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: for.cond3:
// CHECK: [[T2:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T1]]) ]
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T2]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T2]]) ]
// CHECK-DAG: attributes [[A0]] = { {{.*}}convergent{{.*}} }
// CHECK-DAG: attributes [[A3]] = { {{.*}}convergent{{.*}} }
diff --git a/clang/test/CodeGenHLSL/convergence/global_array.hlsl b/clang/test/CodeGenHLSL/convergence/global_array.hlsl
index c594e3a..851b61e 100644
--- a/clang/test/CodeGenHLSL/convergence/global_array.hlsl
+++ b/clang/test/CodeGenHLSL/convergence/global_array.hlsl
@@ -1,12 +1,13 @@
// RUN: %clang_cc1 -finclude-default-header -triple spirv-unknown-vulkan-compute -emit-llvm -disable-llvm-passes -o - %s | FileCheck %s
+// RUN: %clang_cc1 -finclude-default-header -triple dxil-pc-shadermodel6.0-compute -emit-llvm -disable-llvm-passes -o - %s | FileCheck %s
-// CHECK: define internal spir_func void @__cxx_global_var_init()
+// CHECK: define internal{{.*}} void @__cxx_global_var_init()
// CHECK: [[entry_token:%.*]] = call token @llvm.experimental.convergence.entry()
// CHECK: br label %[[loop_entry:.*]]
// CHECK: [[loop_entry]]:
// CHECK: [[loop_token:%.*]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[entry_token]]) ]
-// CHECK: call spir_func void {{.*}} [ "convergencectrl"(token [[loop_token]]) ]
+// CHECK: call{{.*}} void {{.*}} [ "convergencectrl"(token [[loop_token]]) ]
// CHECK: br i1 {{%.*}} label {{%.*}} label %[[loop_entry]]
struct S {
diff --git a/clang/test/CodeGenHLSL/convergence/while.hlsl b/clang/test/CodeGenHLSL/convergence/while.hlsl
index 32579e8..4e8cdcd 100644
--- a/clang/test/CodeGenHLSL/convergence/while.hlsl
+++ b/clang/test/CodeGenHLSL/convergence/while.hlsl
@@ -1,5 +1,7 @@
// RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -x hlsl -triple \
// RUN: spirv-pc-vulkan-library %s -emit-llvm -disable-llvm-passes -o - | FileCheck %s
+// RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -x hlsl -triple \
+// RUN: dxil-pc-shadermodel6.3-library %s -emit-llvm -disable-llvm-passes -o - | FileCheck %s
bool cond();
void foo();
@@ -8,28 +10,28 @@
while (cond()) {
}
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test1v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test1v()
// CHECK-SAME: [[A0:#[0-9]+]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
// CHECK: while.cond:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3:#[0-9]+]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3:#[0-9]+]] [ "convergencectrl"(token [[T1]]) ]
void test2() {
while (cond()) {
foo();
}
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test2v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test2v()
// CHECK-SAME: [[A0]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
// CHECK: while.cond:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: while.body:
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
void test3() {
while (cond()) {
@@ -38,17 +40,17 @@
foo();
}
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test3v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test3v()
// CHECK-SAME: [[A0]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
// CHECK: while.cond:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: if.then:
// CHECK: br label %while.end
// CHECK: if.end:
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: br label %while.cond
void test4() {
@@ -59,15 +61,15 @@
}
}
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test4v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test4v()
// CHECK-SAME: [[A0]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
// CHECK: while.cond:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: if.then:
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: br label %while.end
// CHECK: if.end:
// CHECK: br label %while.cond
@@ -82,18 +84,18 @@
}
}
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test5v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test5v()
// CHECK-SAME: [[A0]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
// CHECK: while.cond:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: while.cond2:
// CHECK: [[T2:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T1]]) ]
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T2]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T2]]) ]
// CHECK: if.then:
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T2]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T2]]) ]
// CHECK: br label %while.end
void test6() {
@@ -107,18 +109,18 @@
}
}
}
-// CHECK-LABEL: define hidden spir_func void @_Z5test6v()
+// CHECK-LABEL: define hidden{{.*}} void @_Z5test6v()
// CHECK-SAME: [[A0]] {
// CHECK: entry:
// CHECK: [[T0:%[0-9]+]] = call token @llvm.experimental.convergence.entry()
// CHECK: while.cond:
// CHECK: [[T1:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T0]]) ]
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: while.cond2:
// CHECK: [[T2:%[0-9]+]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[T1]]) ]
-// CHECK: call spir_func noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T2]]) ]
+// CHECK: call{{.*}} noundef i1 @_Z4condv() [[A3]] [ "convergencectrl"(token [[T2]]) ]
// CHECK: if.then:
-// CHECK: call spir_func void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
+// CHECK: call{{.*}} void @_Z3foov() [[A3]] [ "convergencectrl"(token [[T1]]) ]
// CHECK: br label %while.end
// CHECK-DAG: attributes [[A0]] = { {{.*}}convergent{{.*}} }
diff --git a/clang/test/CodeGenHLSL/inline-constructors.hlsl b/clang/test/CodeGenHLSL/inline-constructors.hlsl
index b0d5a78..2116a63 100644
--- a/clang/test/CodeGenHLSL/inline-constructors.hlsl
+++ b/clang/test/CodeGenHLSL/inline-constructors.hlsl
@@ -46,10 +46,11 @@
// CHECK: define void @main()
// CHECK-NEXT: entry:
+// NOINLINE-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// Verify constructor is emitted
// NOINLINE-NEXT: call void @_GLOBAL__sub_I_inline_constructors.hlsl()
-// NOINLINE-NEXT: %0 = call i32 @llvm.dx.flattened.thread.id.in.group()
-// NOINLINE-NEXT: call void @_Z4mainj(i32 %0)
+// NOINLINE-NEXT: %[[#THREAD_ID:]] = call i32 @llvm.dx.flattened.thread.id.in.group()
+// NOINLINE-NEXT: call void @_Z4mainj(i32 %[[#THREAD_ID]])
// Verify inlining leaves only calls to "llvm." intrinsics
// INLINE-NOT: call {{[^@]*}} @{{[^l][^l][^v][^m][^\.]}}
// CHECK: ret void
@@ -62,6 +63,7 @@
// CHECK: define void @rainyMain()
// CHECK-NEXT: entry:
+// NOINLINE-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// Verify constructor is emitted
// NOINLINE-NEXT: call void @_GLOBAL__sub_I_inline_constructors.hlsl()
// NOINLINE-NEXT: call void @_Z9rainyMainv()
diff --git a/clang/test/CodeGenHLSL/matrix-member-one-based-accessor-scalar-load.hlsl b/clang/test/CodeGenHLSL/matrix-member-one-based-accessor-scalar-load.hlsl
index 6d8a3ce..09b40e2 100644
--- a/clang/test/CodeGenHLSL/matrix-member-one-based-accessor-scalar-load.hlsl
+++ b/clang/test/CodeGenHLSL/matrix-member-one-based-accessor-scalar-load.hlsl
@@ -7,6 +7,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return11u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -20,6 +21,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return12u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -33,6 +35,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return13u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -46,6 +49,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return14u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -59,6 +63,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return21u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -72,6 +77,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return22u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -85,6 +91,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return23u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -98,6 +105,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return24u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -111,6 +119,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return31u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -124,6 +133,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return32u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -137,6 +147,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return33u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -150,6 +161,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return34u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -163,6 +175,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return41u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -176,6 +189,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return42u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -189,6 +203,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return43u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -202,6 +217,7 @@
// CHECK-LABEL: define hidden noundef i32 @_Z8Return44u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
diff --git a/clang/test/CodeGenHLSL/matrix-member-one-based-accessor-scalar-store.hlsl b/clang/test/CodeGenHLSL/matrix-member-one-based-accessor-scalar-store.hlsl
index 9ea292e..868514b 100644
--- a/clang/test/CodeGenHLSL/matrix-member-one-based-accessor-scalar-store.hlsl
+++ b/clang/test/CodeGenHLSL/matrix-member-one-based-accessor-scalar-store.hlsl
@@ -7,6 +7,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat11Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -23,6 +24,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat12Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -40,6 +42,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat13Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -57,6 +60,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat14Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -74,6 +78,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat21Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -91,6 +96,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat22Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -108,6 +114,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat23Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -125,6 +132,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat24Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -142,6 +150,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat31Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -159,6 +168,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat32Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -176,6 +186,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat33Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -193,6 +204,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat34Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -210,6 +222,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat41Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -227,6 +240,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat42Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -244,6 +258,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat43Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -261,6 +276,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat44Ru11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
diff --git a/clang/test/CodeGenHLSL/matrix-member-one-based-swizzle-load.hlsl b/clang/test/CodeGenHLSL/matrix-member-one-based-swizzle-load.hlsl
index 6567d49..de6b78dc 100644
--- a/clang/test/CodeGenHLSL/matrix-member-one-based-swizzle-load.hlsl
+++ b/clang/test/CodeGenHLSL/matrix-member-one-based-swizzle-load.hlsl
@@ -6,6 +6,7 @@
// CHECK-LABEL: define hidden noundef <4 x i32> @_Z17ReturnOnesSwizzleu11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -19,6 +20,7 @@
// CHECK-LABEL: define hidden noundef <4 x i32> @_Z18ReturnOnesSwizzle2u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -32,6 +34,7 @@
// CHECK-LABEL: define hidden noundef <4 x i32> @_Z17ReturnTwosSwizzleu11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -45,6 +48,7 @@
// CHECK-LABEL: define hidden noundef <4 x i32> @_Z18ReturnTwosSwizzle2u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -58,6 +62,7 @@
// CHECK-LABEL: define hidden noundef <4 x i32> @_Z19ReturnThreesSwizzleu11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -71,6 +76,7 @@
// CHECK-LABEL: define hidden noundef <4 x i32> @_Z20ReturnThreesSwizzle2u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -84,6 +90,7 @@
// CHECK-LABEL: define hidden noundef <4 x i32> @_Z18ReturnFoursSwizzleu11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
@@ -97,6 +104,7 @@
// CHECK-LABEL: define hidden noundef <4 x i32> @_Z19ReturnFoursSwizzle2u11matrix_typeILm4ELm4EiE(
// CHECK-SAME: <16 x i32> noundef [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x i32>], align 4
// CHECK-NEXT: store <16 x i32> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x i32>, ptr [[A_ADDR]], align 4
diff --git a/clang/test/CodeGenHLSL/matrix-member-one-based-swizzle-store.hlsl b/clang/test/CodeGenHLSL/matrix-member-one-based-swizzle-store.hlsl
index d6a3744..b30f62e 100644
--- a/clang/test/CodeGenHLSL/matrix-member-one-based-swizzle-store.hlsl
+++ b/clang/test/CodeGenHLSL/matrix-member-one-based-swizzle-store.hlsl
@@ -6,6 +6,7 @@
// CHECK-LABEL: define hidden void @_Z19OnesSwizzleToScalarRu11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -34,6 +35,7 @@
// CHECK-LABEL: define hidden void @_Z19OnesSwizzleToVectorRu11matrix_typeILm4ELm4EiEDv4_i(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], <4 x i32> noundef [[V:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <4 x i32>, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -60,6 +62,7 @@
// CHECK-LABEL: define hidden void @_Z19TwosSwizzleToScalarRu11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -89,6 +92,7 @@
// CHECK-LABEL: define hidden void @_Z19TwosSwizzleToVectorRu11matrix_typeILm4ELm4EiEDv4_i(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], <4 x i32> noundef [[V:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <4 x i32>, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -116,6 +120,7 @@
// CHECK-LABEL: define hidden void @_Z21ThreesSwizzleToScalarRu11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -145,6 +150,7 @@
// CHECK-LABEL: define hidden void @_Z21ThreesSwizzleToVectorRu11matrix_typeILm4ELm4EiEDv4_i(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], <4 x i32> noundef [[V:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <4 x i32>, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -172,6 +178,7 @@
// CHECK-LABEL: define hidden void @_Z20FoursSwizzleToScalarRu11matrix_typeILm4ELm4EiEi(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], i32 noundef [[I:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[I_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -201,6 +208,7 @@
// CHECK-LABEL: define hidden void @_Z20FoursSwizzleToVectorRu11matrix_typeILm4ELm4EiEDv4_i(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], <4 x i32> noundef [[V:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <4 x i32>, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
diff --git a/clang/test/CodeGenHLSL/matrix-member-zero-based-accessor-scalar-load.hlsl b/clang/test/CodeGenHLSL/matrix-member-zero-based-accessor-scalar-load.hlsl
index def0b88..e076b88 100644
--- a/clang/test/CodeGenHLSL/matrix-member-zero-based-accessor-scalar-load.hlsl
+++ b/clang/test/CodeGenHLSL/matrix-member-zero-based-accessor-scalar-load.hlsl
@@ -7,6 +7,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return00u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
@@ -20,6 +21,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return01u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
@@ -33,6 +35,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return02u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
@@ -46,6 +49,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return03u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
@@ -59,6 +63,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return10u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
@@ -72,6 +77,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return11u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
@@ -85,6 +91,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return12u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
@@ -98,6 +105,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return13u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
@@ -111,6 +119,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return20u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
@@ -124,6 +133,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return21u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
@@ -137,6 +147,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return22u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
@@ -150,6 +161,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return23u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
@@ -163,6 +175,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return30u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
@@ -176,6 +189,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return31u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
@@ -189,6 +203,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return32u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
@@ -202,6 +217,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z8Return33u11matrix_typeILm4ELm4EfE(
// CHECK-SAME: <16 x float> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x float>], align 4
// CHECK-NEXT: store <16 x float> [[A]], ptr [[A_ADDR]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x float>, ptr [[A_ADDR]], align 4
diff --git a/clang/test/CodeGenHLSL/matrix-member-zero-based-accessor-scalar-store.hlsl b/clang/test/CodeGenHLSL/matrix-member-zero-based-accessor-scalar-store.hlsl
index ab5b69c..ba220b6 100644
--- a/clang/test/CodeGenHLSL/matrix-member-zero-based-accessor-scalar-store.hlsl
+++ b/clang/test/CodeGenHLSL/matrix-member-zero-based-accessor-scalar-store.hlsl
@@ -7,6 +7,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat00Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -23,6 +24,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat01Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -40,6 +42,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat02Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -57,6 +60,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat03Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -74,6 +78,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat10Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -91,6 +96,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat11Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -108,6 +114,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat12Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -125,6 +132,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat13Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -142,6 +150,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat20Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -159,6 +168,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat21Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -176,6 +186,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat22Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -193,6 +204,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat23Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -210,6 +222,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat30Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -227,6 +240,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat31Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -244,6 +258,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat32Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -261,6 +276,7 @@
// CHECK-LABEL: define hidden void @_Z18StoreScalarAtMat33Ru11matrix_typeILm4ELm4EfEf(
// CHECK-SAME: ptr noalias noundef nonnull align 4 dereferenceable(64) [[A:%.*]], float noundef nofpclass(nan inf) [[F:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[F_ADDR:%.*]] = alloca float, align 4
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
diff --git a/clang/test/CodeGenHLSL/matrix-member-zero-based-swizzle-load.hlsl b/clang/test/CodeGenHLSL/matrix-member-zero-based-swizzle-load.hlsl
index d9daf6e..6c0953a 100644
--- a/clang/test/CodeGenHLSL/matrix-member-zero-based-swizzle-load.hlsl
+++ b/clang/test/CodeGenHLSL/matrix-member-zero-based-swizzle-load.hlsl
@@ -6,6 +6,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <4 x half> @_Z18ReturnZerosSwizzleu11matrix_typeILm4ELm4EDhE(
// CHECK-SAME: <16 x half> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x half>], align 2
// CHECK-NEXT: store <16 x half> [[A]], ptr [[A_ADDR]], align 2
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x half>, ptr [[A_ADDR]], align 2
@@ -19,6 +20,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <4 x half> @_Z19ReturnZerosSwizzle2u11matrix_typeILm4ELm4EDhE(
// CHECK-SAME: <16 x half> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x half>], align 2
// CHECK-NEXT: store <16 x half> [[A]], ptr [[A_ADDR]], align 2
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x half>, ptr [[A_ADDR]], align 2
@@ -32,6 +34,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <4 x half> @_Z17ReturnOnesSwizzleu11matrix_typeILm4ELm4EDhE(
// CHECK-SAME: <16 x half> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x half>], align 2
// CHECK-NEXT: store <16 x half> [[A]], ptr [[A_ADDR]], align 2
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x half>, ptr [[A_ADDR]], align 2
@@ -45,6 +48,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <4 x half> @_Z18ReturnOnesSwizzle2u11matrix_typeILm4ELm4EDhE(
// CHECK-SAME: <16 x half> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x half>], align 2
// CHECK-NEXT: store <16 x half> [[A]], ptr [[A_ADDR]], align 2
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x half>, ptr [[A_ADDR]], align 2
@@ -58,6 +62,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <4 x half> @_Z17ReturnTwosSwizzleu11matrix_typeILm4ELm4EDhE(
// CHECK-SAME: <16 x half> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x half>], align 2
// CHECK-NEXT: store <16 x half> [[A]], ptr [[A_ADDR]], align 2
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x half>, ptr [[A_ADDR]], align 2
@@ -71,6 +76,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <4 x half> @_Z18ReturnTwosSwizzle2u11matrix_typeILm4ELm4EDhE(
// CHECK-SAME: <16 x half> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x half>], align 2
// CHECK-NEXT: store <16 x half> [[A]], ptr [[A_ADDR]], align 2
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x half>, ptr [[A_ADDR]], align 2
@@ -84,6 +90,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <4 x half> @_Z19ReturnThreesSwizzleu11matrix_typeILm4ELm4EDhE(
// CHECK-SAME: <16 x half> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x half>], align 2
// CHECK-NEXT: store <16 x half> [[A]], ptr [[A_ADDR]], align 2
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x half>, ptr [[A_ADDR]], align 2
@@ -97,6 +104,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <4 x half> @_Z20ReturnThreesSwizzle2u11matrix_typeILm4ELm4EDhE(
// CHECK-SAME: <16 x half> noundef nofpclass(nan inf) [[A:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca [4 x <4 x half>], align 2
// CHECK-NEXT: store <16 x half> [[A]], ptr [[A_ADDR]], align 2
// CHECK-NEXT: [[TMP0:%.*]] = load <16 x half>, ptr [[A_ADDR]], align 2
diff --git a/clang/test/CodeGenHLSL/matrix-member-zero-based-swizzle-store.hlsl b/clang/test/CodeGenHLSL/matrix-member-zero-based-swizzle-store.hlsl
index 001f470..011cadc 100644
--- a/clang/test/CodeGenHLSL/matrix-member-zero-based-swizzle-store.hlsl
+++ b/clang/test/CodeGenHLSL/matrix-member-zero-based-swizzle-store.hlsl
@@ -6,6 +6,7 @@
// CHECK-LABEL: define hidden void @_Z20ZerosSwizzleToScalarRu11matrix_typeILm4ELm4EdEd(
// CHECK-SAME: ptr noalias noundef nonnull align 8 dereferenceable(128) [[A:%.*]], double noundef nofpclass(nan inf) [[D:%.*]]) #[[ATTR0:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[D_ADDR:%.*]] = alloca double, align 8
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -34,6 +35,7 @@
// CHECK-LABEL: define hidden void @_Z20ZerosSwizzleToVectorRu11matrix_typeILm4ELm4EdEDv4_d(
// CHECK-SAME: ptr noalias noundef nonnull align 8 dereferenceable(128) [[A:%.*]], <4 x double> noundef nofpclass(nan inf) [[V:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <4 x double>, align 8
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -60,6 +62,7 @@
// CHECK-LABEL: define hidden void @_Z19OnesSwizzleToScalarRu11matrix_typeILm4ELm4EdEd(
// CHECK-SAME: ptr noalias noundef nonnull align 8 dereferenceable(128) [[A:%.*]], double noundef nofpclass(nan inf) [[D:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[D_ADDR:%.*]] = alloca double, align 8
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -89,6 +92,7 @@
// CHECK-LABEL: define hidden void @_Z19OnesSwizzleToVectorRu11matrix_typeILm4ELm4EdEDv4_d(
// CHECK-SAME: ptr noalias noundef nonnull align 8 dereferenceable(128) [[A:%.*]], <4 x double> noundef nofpclass(nan inf) [[V:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <4 x double>, align 8
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -116,6 +120,7 @@
// CHECK-LABEL: define hidden void @_Z19TwosSwizzleToScalarRu11matrix_typeILm4ELm4EdEd(
// CHECK-SAME: ptr noalias noundef nonnull align 8 dereferenceable(128) [[A:%.*]], double noundef nofpclass(nan inf) [[D:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[D_ADDR:%.*]] = alloca double, align 8
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -145,6 +150,7 @@
// CHECK-LABEL: define hidden void @_Z19TwosSwizzleToVectorRu11matrix_typeILm4ELm4EdEDv4_d(
// CHECK-SAME: ptr noalias noundef nonnull align 8 dereferenceable(128) [[A:%.*]], <4 x double> noundef nofpclass(nan inf) [[V:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <4 x double>, align 8
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -172,6 +178,7 @@
// CHECK-LABEL: define hidden void @_Z21ThreesSwizzleToScalarRu11matrix_typeILm4ELm4EdEd(
// CHECK-SAME: ptr noalias noundef nonnull align 8 dereferenceable(128) [[A:%.*]], double noundef nofpclass(nan inf) [[D:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[D_ADDR:%.*]] = alloca double, align 8
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
@@ -201,6 +208,7 @@
// CHECK-LABEL: define hidden void @_Z21ThreesSwizzleToVectorRu11matrix_typeILm4ELm4EdEDv4_d(
// CHECK-SAME: ptr noalias noundef nonnull align 8 dereferenceable(128) [[A:%.*]], <4 x double> noundef nofpclass(nan inf) [[V:%.*]]) #[[ATTR0]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[A_ADDR:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[V_ADDR:%.*]] = alloca <4 x double>, align 8
// CHECK-NEXT: store ptr [[A]], ptr [[A_ADDR]], align 4
diff --git a/clang/test/CodeGenHLSL/resources/ByteAddressBuffers-constructors.hlsl b/clang/test/CodeGenHLSL/resources/ByteAddressBuffers-constructors.hlsl
index 4ac71d0..4d0e79f 100644
--- a/clang/test/CodeGenHLSL/resources/ByteAddressBuffers-constructors.hlsl
+++ b/clang/test/CodeGenHLSL/resources/ByteAddressBuffers-constructors.hlsl
@@ -33,6 +33,7 @@
// Buf1 initialization part 1 - global init function that calls ByteAddressBuffer::__createFromBinding
// CHECK: define internal void @__cxx_global_var_init()
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: call void @hlsl::ByteAddressBuffer::__createFromBinding(unsigned int, unsigned int, int, unsigned int, char const*)
// CHECK-SAME: (ptr {{.*}} @Buf1, i32 noundef 1, i32 noundef 2, i32 noundef 1, i32 noundef 0, ptr noundef @[[Buf1Str]])
@@ -50,6 +51,7 @@
// Buf2 initialization part 1 - global init function that calls RWByteAddressBuffer::__createFromImplicitBinding
// CHECK: define internal void @__cxx_global_var_init.1()
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: call void @hlsl::RWByteAddressBuffer::__createFromImplicitBinding(unsigned int, unsigned int, int, unsigned int, char const*)
// CHECK-SAME: (ptr {{.*}} @Buf2, i32 noundef 0, i32 noundef 0, i32 noundef 1, i32 noundef 0, ptr noundef @[[Buf2Str]])
@@ -68,6 +70,7 @@
// RasterizerOrderedByteAddressBuffer C1 default constructor
// CHECK: define void @foo()
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: %Buf3 = alloca %"class.hlsl::RasterizerOrderedByteAddressBuffer", align 4
// CHECK-NEXT: call void @hlsl::RasterizerOrderedByteAddressBuffer::RasterizerOrderedByteAddressBuffer()(ptr {{.*}} %Buf3)
@@ -86,5 +89,6 @@
// Module initialization
// CHECK: define internal void @_GLOBAL__sub_I_ByteAddressBuffers_constructors.hlsl()
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: call void @__cxx_global_var_init()
// CHECK-NEXT: call void @__cxx_global_var_init.1()
diff --git a/clang/test/CodeGenHLSL/resources/ByteAddressBuffers-methods.hlsl b/clang/test/CodeGenHLSL/resources/ByteAddressBuffers-methods.hlsl
index fa04d82..ac5cedd 100644
--- a/clang/test/CodeGenHLSL/resources/ByteAddressBuffers-methods.hlsl
+++ b/clang/test/CodeGenHLSL/resources/ByteAddressBuffers-methods.hlsl
@@ -65,9 +65,9 @@
// CHECK: define {{.*}} float @TestLoadWithStatus()()
// CHECK: call {{.*}} i32 @hlsl::ByteAddressBuffer::Load(unsigned int, unsigned int&)(ptr {{.*}} @Buf, i32 noundef 0, ptr {{.*}} %tmp)
-// CHECK: call {{.*}} <4 x i32> @hlsl::RWByteAddressBuffer::Load4(unsigned int, unsigned int&)(ptr {{.*}} @RWBuf, i32 noundef 4, ptr {{.*}} %tmp1)
-// CHECK: call {{.*}} float @float hlsl::ByteAddressBuffer::Load<float>(unsigned int, unsigned int&)(ptr {{.*}} @Buf, i32 noundef 20, ptr {{.*}} %tmp4)
-// CHECK: call {{.*}} <4 x float> @float vector[4] hlsl::RWByteAddressBuffer::Load<float vector[4]>(unsigned int, unsigned int&)(ptr {{.*}} @RWBuf, i32 noundef 24, ptr {{.*}} %tmp7)
+// CHECK: call {{.*}} <4 x i32> @hlsl::RWByteAddressBuffer::Load4(unsigned int, unsigned int&)(ptr {{.*}} @RWBuf, i32 noundef 4, ptr {{.*}} %[[TMP1:.*]])
+// CHECK: call {{.*}} float @float hlsl::ByteAddressBuffer::Load<float>(unsigned int, unsigned int&)(ptr {{.*}} @Buf, i32 noundef 20, ptr {{.*}} %[[TMP2:.*]])
+// CHECK: call {{.*}} <4 x float> @float vector[4] hlsl::RWByteAddressBuffer::Load<float vector[4]>(unsigned int, unsigned int&)(ptr {{.*}} @RWBuf, i32 noundef 24, ptr {{.*}} %[[TMP3:.*]])
// CHECK: add
// CHECK: ret float
diff --git a/clang/test/CodeGenHLSL/resources/CBufferMatrixSingleSubscriptSwizzle.hlsl b/clang/test/CodeGenHLSL/resources/CBufferMatrixSingleSubscriptSwizzle.hlsl
index fe74e98..ad6c352 100644
--- a/clang/test/CodeGenHLSL/resources/CBufferMatrixSingleSubscriptSwizzle.hlsl
+++ b/clang/test/CodeGenHLSL/resources/CBufferMatrixSingleSubscriptSwizzle.hlsl
@@ -9,6 +9,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <2 x float> @_Z15get_row_swizzlev(
// CHECK-SAME: ) #[[ATTR2:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[MATRIX_BUF_COPY:%.*]] = alloca [3 x <2 x float>], align 4
// CHECK-NEXT: [[CBUF_DEST:%.*]] = getelementptr inbounds [3 x <2 x float>], ptr [[MATRIX_BUF_COPY]], i32 0, i32 0
// CHECK-NEXT: [[CBUF_LOAD:%.*]] = load <2 x float>, ptr addrspace(2) @Mat, align 4
diff --git a/clang/test/CodeGenHLSL/resources/MatrixElement_cbuffer.hlsl b/clang/test/CodeGenHLSL/resources/MatrixElement_cbuffer.hlsl
index 0b4dc5b..bcb8b4b 100644
--- a/clang/test/CodeGenHLSL/resources/MatrixElement_cbuffer.hlsl
+++ b/clang/test/CodeGenHLSL/resources/MatrixElement_cbuffer.hlsl
@@ -14,6 +14,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) float @_Z23getCBufferScalarElementv(
// CHECK-SAME: ) #[[ATTR2:[0-9]+]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// COL-CHECK-NEXT: [[MATRIX_BUF_COPY:%.*]] = alloca [2 x <3 x float>], align 4
// COL-CHECK-NEXT: [[CBUF_DEST:%.*]] = getelementptr inbounds [2 x <3 x float>], ptr [[MATRIX_BUF_COPY]], i32 0, i32 0
// COL-CHECK-NEXT: [[CBUF_LOAD:%.*]] = load <3 x float>, ptr addrspace(2) @Mat, align 4
@@ -47,6 +48,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <4 x float> @_Z23getCBufferSwizzleAccessv(
// CHECK-SAME: ) #[[ATTR2]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// COL-CHECK-NEXT: [[MATRIX_BUF_COPY:%.*]] = alloca [2 x <3 x float>], align 4
// COL-CHECK-NEXT: [[CBUF_DEST:%.*]] = getelementptr inbounds [2 x <3 x float>], ptr [[MATRIX_BUF_COPY]], i32 0, i32 0
// COL-CHECK-NEXT: [[CBUF_LOAD:%.*]] = load <3 x float>, ptr addrspace(2) @Mat, align 4
@@ -79,6 +81,7 @@
// CHECK-LABEL: define hidden noundef nofpclass(nan inf) <2 x float> @_Z22getZeroBasedSwizzleEltu11matrix_typeILm3ELm2EfE(
// CHECK-SAME: <6 x float> noundef nofpclass(nan inf) [[M:%.*]]) #[[ATTR2]] {
// CHECK-NEXT: [[ENTRY:.*:]]
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// COL-CHECK-NEXT: [[M_ADDR:%.*]] = alloca [2 x <3 x float>], align 4
// COL-CHECK-NEXT: store <6 x float> [[M]], ptr [[M_ADDR]], align 4
// COL-CHECK-NEXT: [[TMP0:%.*]] = load <6 x float>, ptr [[M_ADDR]], align 4
diff --git a/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-lib.hlsl b/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-lib.hlsl
index 12a644b..1815ab6 100644
--- a/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-lib.hlsl
+++ b/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-lib.hlsl
@@ -47,7 +47,7 @@
}
// CHECK: define void @TestAppend(float)(float {{.*}} %value)
-// CHECK: call void @hlsl::AppendStructuredBuffer<float>::Append(float)(ptr {{.*}} @ASB, float noundef nofpclass(nan inf) %0)
+// CHECK: call void @hlsl::AppendStructuredBuffer<float>::Append(float)(ptr {{.*}} @ASB, float noundef nofpclass(nan inf) %[[#]])
// CHECK: ret void
// CHECK: define {{.*}} void @hlsl::AppendStructuredBuffer<float>::Append(float)(ptr {{.*}} %this, float noundef nofpclass(nan inf) %value)
@@ -114,7 +114,7 @@
// CHECK: define noundef nofpclass(nan inf) float @TestLoadWithStatus()()
// CHECK: call {{.*}} float @hlsl::RWStructuredBuffer<float>::Load(unsigned int, unsigned int&)(ptr {{.*}} @RWSB1, i32 noundef 1, ptr {{.*}} %tmp)
-// CHECK: call {{.*}} float @hlsl::StructuredBuffer<float>::Load(unsigned int, unsigned int&)(ptr {{.*}} @SB1, i32 noundef 2, ptr {{.*}} %tmp1)
+// CHECK: call {{.*}} float @hlsl::StructuredBuffer<float>::Load(unsigned int, unsigned int&)(ptr {{.*}} @SB1, i32 noundef 2, ptr {{.*}} %[[TMP0:.*]])
// CHECK: add
// CHECK: ret float
diff --git a/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-ps.hlsl b/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-ps.hlsl
index 85ca449..83bcc6e 100644
--- a/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-ps.hlsl
+++ b/clang/test/CodeGenHLSL/resources/StructuredBuffers-methods-ps.hlsl
@@ -74,7 +74,7 @@
// CHECK: define {{.*}} float @TestLoadWithStatus()()
// CHECK: call {{.*}} float @hlsl::RasterizerOrderedStructuredBuffer<float>::Load(unsigned int, unsigned int&)(ptr {{.*}} @ROSB1, i32 noundef 10, ptr {{.*}} %tmp)
-// CHECK: call {{.*}} <2 x i32> @hlsl::RasterizerOrderedStructuredBuffer<int vector[2]>::Load(unsigned int, unsigned int&)(ptr {{.*}} @ROSB2, i32 noundef 20, ptr {{.*}} %tmp2)
+// CHECK: call {{.*}} <2 x i32> @hlsl::RasterizerOrderedStructuredBuffer<int vector[2]>::Load(unsigned int, unsigned int&)(ptr {{.*}} @ROSB2, i32 noundef 20, ptr {{.*}} %[[TMP0:.*]])
// CHECK: ret
// CHECK: define {{.*}} float @hlsl::RasterizerOrderedStructuredBuffer<float>::Load(unsigned int, unsigned int&)(ptr {{.*}} %this, i32 noundef %Index, ptr {{.*}} %Status)
diff --git a/clang/test/CodeGenHLSL/resources/Texture2D-Mips.hlsl b/clang/test/CodeGenHLSL/resources/Texture2D-Mips.hlsl
index b84df2b1..335ebe1 100644
--- a/clang/test/CodeGenHLSL/resources/Texture2D-Mips.hlsl
+++ b/clang/test/CodeGenHLSL/resources/Texture2D-Mips.hlsl
@@ -2,7 +2,7 @@
Texture2D<float4> t;
-// CHECK: define internal {{.*}} <4 x float> @test_mips(float vector[2])(<2 x float> {{.*}} %loc) #1 {
+// CHECK: define internal {{.*}} <4 x float> @test_mips(float vector[2])(<2 x float> {{.*}} %loc)
// CHECK: entry:
// CHECK: %[[LOC_ADDR:.*]] = alloca <2 x float>
// CHECK: %[[REF_TMP:.*]] = alloca %"struct.hlsl::Texture2D<>::mips_slice_type"
diff --git a/clang/test/CodeGenHLSL/resources/TypedBuffers-constructor.hlsl b/clang/test/CodeGenHLSL/resources/TypedBuffers-constructor.hlsl
index 1ec9f0f..d027d76 100644
--- a/clang/test/CodeGenHLSL/resources/TypedBuffers-constructor.hlsl
+++ b/clang/test/CodeGenHLSL/resources/TypedBuffers-constructor.hlsl
@@ -33,6 +33,7 @@
// Buf1 initialization part 1 - global init function that calls RWBuffer<float>::__createFromBinding
// CHECK: define internal void @__cxx_global_var_init()
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: call void @hlsl::RWBuffer<float>::__createFromBinding(unsigned int, unsigned int, int, unsigned int, char const*)
// CHECK-SAME: (ptr {{.*}} sret(%"class.hlsl::RWBuffer") align 4 @Buf1, i32 noundef 5, i32 noundef 3, i32 noundef 1, i32 noundef 0, ptr noundef @[[Buf1Str]])
@@ -50,6 +51,7 @@
// Buf2 initialization part 1 - global init function that RWBuffer<float>::__createFromImplicitBinding
// CHECK: define internal void @__cxx_global_var_init.1()
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: call void @hlsl::Buffer<double>::__createFromImplicitBinding(unsigned int, unsigned int, int, unsigned int, char const*)
// CHECK-SAME: (ptr {{.*}} @Buf2, i32 noundef 0, i32 noundef 0, i32 noundef 1, i32 noundef 0, ptr noundef @[[Buf2Str]])
@@ -67,6 +69,7 @@
// Buf3 initialization part 1 - local variable declared in function foo() is initialized by RWBuffer<int> C1 default constructor
// CHECK: define void @foo()
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: %Buf3 = alloca %"class.hlsl::RWBuffer.0", align 4
// CHECK-NEXT: call void @hlsl::RWBuffer<int>::RWBuffer()(ptr {{.*}} %Buf3)
@@ -82,5 +85,6 @@
// Module initialization
// CHECK: define internal void @_GLOBAL__sub_I_TypedBuffers_constructor.hlsl()
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: call void @__cxx_global_var_init()
// CHECK-NEXT: call void @__cxx_global_var_init.1()
diff --git a/clang/test/CodeGenHLSL/resources/TypedBuffers-methods.hlsl b/clang/test/CodeGenHLSL/resources/TypedBuffers-methods.hlsl
index bdf7f4d..2c92c51 100644
--- a/clang/test/CodeGenHLSL/resources/TypedBuffers-methods.hlsl
+++ b/clang/test/CodeGenHLSL/resources/TypedBuffers-methods.hlsl
@@ -48,7 +48,7 @@
// CHECK: define noundef nofpclass(nan inf) float @TestLoadWithStatus()()
// CHECK: call {{.*}} float @hlsl::Buffer<float>::Load(unsigned int, unsigned int&)(ptr {{.*}} @Buf, i32 noundef 1, ptr {{.*}} %tmp)
-// CHECK: call {{.*}} <4 x i32> @hlsl::RWBuffer<unsigned int vector[4]>::Load(unsigned int, unsigned int&)(ptr {{.*}} @RWBuf, i32 noundef 2, ptr {{.*}} %tmp1)
+// CHECK: call {{.*}} <4 x i32> @hlsl::RWBuffer<unsigned int vector[4]>::Load(unsigned int, unsigned int&)(ptr {{.*}} @RWBuf, i32 noundef 2, ptr {{.*}} %[[TMP0:.*]])
// CHECK: add
// CHECK: ret float
diff --git a/clang/test/CodeGenHLSL/resources/cbuffer.hlsl b/clang/test/CodeGenHLSL/resources/cbuffer.hlsl
index 335e68e..ef9c63a 100644
--- a/clang/test/CodeGenHLSL/resources/cbuffer.hlsl
+++ b/clang/test/CodeGenHLSL/resources/cbuffer.hlsl
@@ -421,6 +421,7 @@
// CHECK: define internal void @_GLOBAL__sub_I_cbuffer.hlsl()
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: call void @_init_buffer_CBScalars.cb()
// CHECK-NEXT: call void @_init_buffer_CBVectors.cb()
// CHECK-NEXT: call void @_init_buffer_CBArrays.cb()
diff --git a/clang/test/CodeGenHLSL/resources/cbuffer_with_packoffset.hlsl b/clang/test/CodeGenHLSL/resources/cbuffer_with_packoffset.hlsl
index 48463e8..818786c 100644
--- a/clang/test/CodeGenHLSL/resources/cbuffer_with_packoffset.hlsl
+++ b/clang/test/CodeGenHLSL/resources/cbuffer_with_packoffset.hlsl
@@ -50,6 +50,7 @@
}
// CHECK: define internal void @_GLOBAL__sub_I_cbuffer_with_packoffset.hlsl()
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: call void @_init_buffer_CB.cb()
[numthreads(4,1,1)]
diff --git a/clang/test/CodeGenHLSL/resources/res-array-global-subarray-many.hlsl b/clang/test/CodeGenHLSL/resources/res-array-global-subarray-many.hlsl
index 036feec..5264c43 100644
--- a/clang/test/CodeGenHLSL/resources/res-array-global-subarray-many.hlsl
+++ b/clang/test/CodeGenHLSL/resources/res-array-global-subarray-many.hlsl
@@ -9,6 +9,7 @@
// CHECK: define {{.*}} float @foo(hlsl::RWBuffer<float> [3][2])
// CHECK-SAME: (ptr noundef byval([3 x [2 x %"class.hlsl::RWBuffer"]]) align 4 %Arr)
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
float foo(RWBuffer<float> Arr[3][2]) {
// CHECK-NEXT: %[[Arr_1_Ptr:.*]] = getelementptr inbounds [3 x [2 x %"class.hlsl::RWBuffer"]], ptr %Arr, i32 0, i32 1
// CHECK-NEXT: %[[Arr_1_0_Ptr:.*]] = getelementptr inbounds [2 x %"class.hlsl::RWBuffer"], ptr %[[Arr_1_Ptr]], i32 0, i32 0
@@ -20,7 +21,8 @@
// CHECK: define internal void @main(unsigned int)(i32 noundef %GI)
// CHECK-NEXT: entry:
-// CHECK-NEXT: %[[GI_alloca:.*]] = alloca i32, align 4
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
+// CHECK-NEXT: %[[GI_alloca:.*]] = alloca i32, align 4
// CHECK-NEXT: %Sub = alloca [3 x [2 x %"class.hlsl::RWBuffer"]], align 4
// CHECK-NEXT: %[[Tmp0:.*]] = alloca [3 x [2 x %"class.hlsl::RWBuffer"]], align 4
// CHECK-NEXT: %a = alloca float, align 4
@@ -66,6 +68,18 @@
// After this Tmp0 values are copied to %Sub using the standard array loop initializaion
// (generated from ArrayInitLoopExpr AST node)
+// CHECK: br label %arrayinit.body
+// CHECK: arrayinit.body:
+// CHECK-NEXT: %{{.*}} = phi i32 [ 0, %entry ], [ %arrayinit.next{{.*}}, %arrayinit.end ]
+// CHECK-NEXT: %[[#CV_LOOP:]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token %[[#C_ENTRY]]) ]
+// CHECK: br label %arrayinit.body2
+// CHECK: arrayinit.body2:
+// CHECK-NEXT: %{{.*}} = phi i32 [ 0, %arrayinit.body ], [ %arrayinit.next, %arrayinit.body2 ]
+// CHECK-NEXT: %[[#CV_INNER_LOOP:]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token %[[#CV_LOOP]]) ]
+// CHECK: call void @hlsl::RWBuffer<float>::RWBuffer(hlsl::RWBuffer<float> const&){{.*}} [ "convergencectrl"(token %[[#CV_INNER_LOOP]]) ]
+// CHECK: br i1 %arrayinit.done, label %arrayinit.end, label %arrayinit.body2
+// CHECK: arrayinit.end:
+// CHECK: br i1 %arrayinit.done{{.*}}, label %arrayinit.end{{.*}}, label %arrayinit.body
RWBuffer<float> Sub[3][2] = A[4][1];
// CHECK: %[[Ptr_Sub_2:.*]] = getelementptr inbounds [3 x [2 x %"class.hlsl::RWBuffer"]], ptr %Sub, i32 0, i32 2
diff --git a/clang/test/CodeGenHLSL/resources/res-array-global-subarray-one.hlsl b/clang/test/CodeGenHLSL/resources/res-array-global-subarray-one.hlsl
index bbd48b7..8f41020c 100644
--- a/clang/test/CodeGenHLSL/resources/res-array-global-subarray-one.hlsl
+++ b/clang/test/CodeGenHLSL/resources/res-array-global-subarray-one.hlsl
@@ -12,7 +12,8 @@
// CHECK: define internal void @main(unsigned int)(i32 noundef %GI)
// CHECK-NEXT: entry:
-// CHECK-NEXT: %[[GI_alloca:.*]] = alloca i32, align 4
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
+// CHECK-NEXT: %[[GI_alloca:.*]] = alloca i32, align 4
// CHECK-NEXT: %Sub = alloca [2 x %"class.hlsl::RWBuffer"], align 4
// CHECK-NEXT: %[[Tmp0:.*]] = alloca [2 x %"class.hlsl::RWBuffer"], align 4
// CHECK-NEXT: %a = alloca float, align 4
@@ -34,6 +35,12 @@
// CHECK-SAME: i32 noundef 10, i32 noundef 2, i32 noundef 8, i32 noundef 7, ptr noundef @[[BufA]])
// After this Tmp0 values are copied to %Sub using the standard array loop initializaion
// (generated from ArrayInitLoopExpr AST node)
+// CHECK: br label %arrayinit.body
+// CHECK: arrayinit.body:
+// CHECK-NEXT: %{{.*}} = phi i32 [ 0, %entry ], [ %arrayinit.next, %arrayinit.body ]
+// CHECK-NEXT: %[[#CV_LOOP:]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token %[[#C_ENTRY]]) ]
+// CHECK: call void @hlsl::RWBuffer<float>::RWBuffer(hlsl::RWBuffer<float> const&){{.*}} [ "convergencectrl"(token %[[#CV_LOOP]]) ]
+// CHECK: br i1 %arrayinit.done, label %arrayinit.end, label %arrayinit.body
RWBuffer<float> Sub[2] = A[3];
// CHECK: %[[Ptr_Sub_1:.*]] = getelementptr inbounds [2 x %"class.hlsl::RWBuffer"], ptr %Sub, i32 0, i32 1
diff --git a/clang/test/CodeGenHLSL/resources/res-array-local-multi-dim.hlsl b/clang/test/CodeGenHLSL/resources/res-array-local-multi-dim.hlsl
index 92dba21..d3ad02c 100644
--- a/clang/test/CodeGenHLSL/resources/res-array-local-multi-dim.hlsl
+++ b/clang/test/CodeGenHLSL/resources/res-array-local-multi-dim.hlsl
@@ -16,6 +16,7 @@
// CHECK: define {{.*}} float @_Z3fooA2_A2_N4hlsl8RWBufferIfEE(ptr noundef byval([2 x [2 x %"class.hlsl::RWBuffer"]]) align 4 %Arr)
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
float foo(RWBuffer<float> Arr[2][2]) {
// CHECK-NEXT: %[[Arr_1_Ptr:.*]] = getelementptr inbounds [2 x [2 x %"class.hlsl::RWBuffer"]], ptr %Arr, i32 0, i32 1
// CHECK-NEXT: %[[Arr_1_1_Ptr:.*]] = getelementptr inbounds [2 x %"class.hlsl::RWBuffer"], ptr %[[Arr_1_Ptr]], i32 0, i32 1
@@ -27,6 +28,7 @@
// CHECK: define internal void @_Z4mainv()
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
[numthreads(4,1,1)]
void main() {
// CHECK: %L = alloca [2 x [2 x %"class.hlsl::RWBuffer"]], align 4
diff --git a/clang/test/CodeGenHLSL/resources/res-array-local1.hlsl b/clang/test/CodeGenHLSL/resources/res-array-local1.hlsl
index 9e31f4f..bb8fb39 100644
--- a/clang/test/CodeGenHLSL/resources/res-array-local1.hlsl
+++ b/clang/test/CodeGenHLSL/resources/res-array-local1.hlsl
@@ -14,6 +14,7 @@
// CHECK: define internal void @_Z4mainv()
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
[numthreads(4,1,1)]
void main() {
// CHECK-NEXT: %First = alloca [3 x %"class.hlsl::RWBuffer"], align 4
@@ -35,6 +36,7 @@
// CHECK-NEXT: br label %[[ArrayInitLoop:.*]]
// CHECK: [[ArrayInitLoop]]:
// CHECK-NEXT: %[[ArrayCurPtr:.*]] = phi ptr [ %[[ArrayBeginPtr]], %entry ], [ %[[ArrayNextPtr:.*]], %[[ArrayInitLoop]] ]
+// CHECK-NEXT: %[[#C_LOOP:]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token %[[#C_ENTRY]]) ]
// CHECK-NEXT: call void @_ZN4hlsl8RWBufferIfEC1Ev(ptr {{.*}} %[[ArrayCurPtr]])
// CHECK-NEXT: %[[ArrayNextPtr]] = getelementptr inbounds %"class.hlsl::RWBuffer", ptr %[[ArrayCurPtr]], i32 1
// CHECK-NEXT: %[[ArrayInitDone:.*]] = icmp eq ptr %[[ArrayNextPtr]], %[[ArrayEndPtr]]
diff --git a/clang/test/CodeGenHLSL/resources/res-array-local2.hlsl b/clang/test/CodeGenHLSL/resources/res-array-local2.hlsl
index 4a02b9f..82fbdc1 100644
--- a/clang/test/CodeGenHLSL/resources/res-array-local2.hlsl
+++ b/clang/test/CodeGenHLSL/resources/res-array-local2.hlsl
@@ -13,6 +13,7 @@
// CHECK: define {{.*}} float @_Z3fooA3_N4hlsl8RWBufferIfEE(ptr noundef byval([3 x %"class.hlsl::RWBuffer"]) align 4 %LocalA)
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
float foo(RWBuffer<float> LocalA[3]) {
// CHECK-NEXT: %[[LocalA_2_Ptr:.*]] = getelementptr inbounds [3 x %"class.hlsl::RWBuffer"], ptr %LocalA, i32 0, i32 2
// CHECK-NEXT: %[[BufPtr:.*]] = call {{.*}} ptr @_ZN4hlsl8RWBufferIfEixEj(ptr {{.*}} %[[LocalA_2_Ptr]], i32 noundef 0)
@@ -23,6 +24,7 @@
// CHECK: define internal void @_Z4mainv()
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
[numthreads(4,1,1)]
void main() {
// Check that the `main` function calls `foo` with a local copy of the array
diff --git a/clang/test/CodeGenHLSL/resources/res-array-local3.hlsl b/clang/test/CodeGenHLSL/resources/res-array-local3.hlsl
index 21ca3f4..c21afba 100644
--- a/clang/test/CodeGenHLSL/resources/res-array-local3.hlsl
+++ b/clang/test/CodeGenHLSL/resources/res-array-local3.hlsl
@@ -13,6 +13,7 @@
// CHECK: define {{.*}} @_Z6SomeFnA2_N4hlsl8RWBufferIiEEji(
// CHECK-SAME: ptr noundef byval([2 x %"class.hlsl::RWBuffer"]) align 4 %B, i32 noundef %Idx, i32 noundef %Val0)
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: %[[Idx_addr:.*]] = alloca i32, align 4
// CHECK-NEXT: %[[Val0_addr:.*]] = alloca i32, align 4
// CHECK-NEXT: store i32 %Idx, ptr %[[Idx_addr]], align 4
@@ -35,6 +36,7 @@
// CHECK: define {{.*}} void @_Z4mainj(i32 noundef %GI)
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: %[[GI_addr:.*]] = alloca i32, align 4
[numthreads(4,1,1)]
void main(uint GI : SV_GroupIndex) {
diff --git a/clang/test/CodeGenHLSL/static-local-ctor.hlsl b/clang/test/CodeGenHLSL/static-local-ctor.hlsl
index fd92e41..d763779 100644
--- a/clang/test/CodeGenHLSL/static-local-ctor.hlsl
+++ b/clang/test/CodeGenHLSL/static-local-ctor.hlsl
@@ -19,6 +19,7 @@
// CHECK-NOT: _Init_thread_epoch
// CHECK: define internal void @main
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[Tmp0:%.*]] = alloca %"class.hlsl::RWBuffer"
// CHECK-NEXT: [[Tmp1:%.*]] = alloca %"class.hlsl::RWBuffer"
// CHECK-NEXT: [[Tmp2:%.*]] = load i8, ptr @guard variable for main()::mybuf
diff --git a/clang/test/CodeGenHLSL/this-assignment-overload.hlsl b/clang/test/CodeGenHLSL/this-assignment-overload.hlsl
index a2df307..843cab1 100644
--- a/clang/test/CodeGenHLSL/this-assignment-overload.hlsl
+++ b/clang/test/CodeGenHLSL/this-assignment-overload.hlsl
@@ -27,6 +27,7 @@
// This test makes a probably safe assumption that HLSL 202x includes operator overloading for assignment operators.
// CHECK: define linkonce_odr hidden noundef i32 @_ZN4Pair8getFirstEv(ptr noundef nonnull align 1 dereferenceable(8) %this) #0 align 2 {
// CHECK-NEXT:entry:
+// CHECK-NEXT:%[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT:%this.addr = alloca ptr, align 4
// CHECK-NEXT:%Another = alloca %struct.Pair, align 1
// CHECK-NEXT:%agg.tmp = alloca %struct.Pair, align 1
@@ -39,11 +40,12 @@
// CHECK-NEXT:call void @llvm.memcpy.p0.p0.i32(ptr align 1 %agg.tmp, ptr align 1 %Another, i32 8, i1 false)
// CHECK-NEXT:call void @_ZN4PairaSES_(ptr noundef nonnull align 1 dereferenceable(8) %this1, ptr noundef byval(%struct.Pair) align 1 %agg.tmp)
// CHECK-NEXT:%First2 = getelementptr inbounds nuw %struct.Pair, ptr %this1, i32 0, i32 0
-// CHECK-NEXT:%0 = load i32, ptr %First2, align 1
-// CHECK-NEXT:ret i32 %0
+// CHECK-NEXT:%[[#FIRST:]] = load i32, ptr %First2, align 1
+// CHECK-NEXT:ret i32 %[[#FIRST]]
// CHECK: define linkonce_odr hidden noundef i32 @_ZN4Pair9getSecondEv(ptr noundef nonnull align 1 dereferenceable(8) %this) #0 align 2 {
// CHECK-NEXT:entry:
+// CHECK-NEXT:%[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT:%this.addr = alloca ptr, align 4
// CHECK-NEXT:%agg.tmp = alloca %struct.Pair, align 1
// CHECK-NEXT:store ptr %this, ptr %this.addr, align 4
@@ -51,5 +53,5 @@
// CHECK-NEXT:call void @llvm.memset.p0.i32(ptr align 1 %agg.tmp, i8 0, i32 8, i1 false)
// CHECK-NEXT:call void @_ZN4PairaSES_(ptr noundef nonnull align 1 dereferenceable(8) %this1, ptr noundef byval(%struct.Pair) align 1 %agg.tmp)
// CHECK-NEXT:%Second = getelementptr inbounds nuw %struct.Pair, ptr %this1, i32 0, i32 1
-// CHECK-NEXT:%0 = load i32, ptr %Second, align 1
-// CHECK-NEXT:ret i32 %0
+// CHECK-NEXT:%[[#SECOND:]] = load i32, ptr %Second, align 1
+// CHECK-NEXT:ret i32 %[[#SECOND]]
diff --git a/clang/test/CodeGenHLSL/this-assignment.hlsl b/clang/test/CodeGenHLSL/this-assignment.hlsl
index efccc96..c79f94d5 100644
--- a/clang/test/CodeGenHLSL/this-assignment.hlsl
+++ b/clang/test/CodeGenHLSL/this-assignment.hlsl
@@ -35,6 +35,7 @@
// This tests reference like implicit this in HLSL
// CHECK-LABEL: define {{.*}}getFirst
// CHECK-NEXT:entry:
+// CHECK-NEXT:%[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT:%this.addr = alloca ptr, align 4
// CHECK-NEXT:%Another = alloca %struct.Pair, align 1
// CHECK-NEXT:store ptr %this, ptr %this.addr, align 4
@@ -45,6 +46,7 @@
// CHECK-LABEL: define {{.*}}getSecond
// CHECK-NEXT:entry:
+// CHECK-NEXT:%[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT:%this.addr = alloca ptr, align 4
// CHECK-NEXT:%ref.tmp = alloca %struct.Pair, align 1
// CHECK-NEXT:store ptr %this, ptr %this.addr, align 4
@@ -55,6 +57,7 @@
// CHECK-LABEL: define {{.*}}DoSilly
// CHECK-NEXT:entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: [[ThisPtrAddr:%.*]] = alloca ptr
// CHECK-NEXT: store ptr {{.*}}, ptr [[ThisPtrAddr]]
// CHECK-NEXT: [[ThisPtr:%.*]] = load ptr, ptr [[ThisPtrAddr]]
diff --git a/clang/test/CodeGenHLSL/this-reference.hlsl b/clang/test/CodeGenHLSL/this-reference.hlsl
index edf1abf..e19d0a4 100644
--- a/clang/test/CodeGenHLSL/this-reference.hlsl
+++ b/clang/test/CodeGenHLSL/this-reference.hlsl
@@ -21,10 +21,11 @@
}
// This tests reference like `this` in HLSL
- // CHECK: %call = call noundef i32 @_ZN4Pair8getFirstEv(ptr noundef nonnull align 1 dereferenceable(8) %Vals)
+ // CHECK: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
+ // CHECK: %[[CALL1:.*]] = call noundef i32 @_ZN4Pair8getFirstEv(ptr noundef nonnull align 1 dereferenceable(8) %Vals) #{{[0-9]+}} [ "convergencectrl"(token %[[#C_ENTRY]]) ]
// CHECK-NEXT: %First = getelementptr inbounds nuw %struct.Pair, ptr %Vals, i32 0, i32 0
- // CHECK-NEXT: store i32 %call, ptr %First, align 1
- // CHECK-NEXT: %call1 = call reassoc nnan ninf nsz arcp afn noundef nofpclass(nan inf) float @_ZN4Pair9getSecondEv(ptr noundef nonnull align 1 dereferenceable(8) %Vals)
+ // CHECK-NEXT: store i32 %[[CALL1]], ptr %First, align 1
+ // CHECK-NEXT: %[[CALL2:.*]] = call reassoc nnan ninf nsz arcp afn noundef nofpclass(nan inf) float @_ZN4Pair9getSecondEv(ptr noundef nonnull align 1 dereferenceable(8) %Vals) #{{[0-9]+}} [ "convergencectrl"(token %[[#C_ENTRY]]) ]
// CHECK-NEXT: %Second = getelementptr inbounds nuw %struct.Pair, ptr %Vals, i32 0, i32 1
// CHECK: [[Pair:![0-9]+]] = distinct !DICompositeType(tag: DW_TAG_structure_type, name: "Pair"
diff --git a/clang/test/SemaHLSL/Resources/static_resources.hlsl b/clang/test/SemaHLSL/Resources/static_resources.hlsl
index 0ea64a3..a539582 100644
--- a/clang/test/SemaHLSL/Resources/static_resources.hlsl
+++ b/clang/test/SemaHLSL/Resources/static_resources.hlsl
@@ -15,6 +15,7 @@
// startup (register 1, space 5).
// CHECK: define internal void @__cxx_global_var_init{{.*}}
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: call void @hlsl::RWBuffer<float>::__createFromBinding(unsigned int, unsigned int, int, unsigned int, char const*)
// CHECK-SAME: (ptr {{.*}} @One, i32 noundef 1, i32 noundef 5, i32 noundef 1, i32 noundef 0, ptr noundef [[ONE_STR]])
@@ -22,6 +23,7 @@
// startup (register 2, space 4).
// CHECK: define internal void @__cxx_global_var_init{{.*}}
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: call void @hlsl::RWStructuredBuffer<int>::__createFromBindingWithImplicitCounter(unsigned int, unsigned int, int, unsigned int, char const*, unsigned int)
// CHECK-SAME: (ptr {{.*}} @OneWithCounter, i32 noundef 2, i32 noundef 4, i32 noundef 1, i32 noundef 0, ptr noundef [[ONEWITHCOUNTER_STR]], i32 noundef 0)
@@ -35,6 +37,7 @@
// constructor and not from binding. It will initalize the handle to poison.
// CHECK: define internal void @__cxx_global_var_init{{.*}}
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: call void @hlsl::RWBuffer<float>::RWBuffer()(ptr {{.*}} @StaticOne)
// Check that StaticArray elements are initialized on startup with the default
@@ -42,9 +45,11 @@
// elements and call the default constructor for each one, setting the handle to poison.
// CHECK: define internal void @__cxx_global_var_init{{.*}}
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: br label %arrayctor.loop
// CHECK: arrayctor.loop: ; preds = %arrayctor.loop, %entry
// CHECK-NEXT: %arrayctor.cur = phi ptr [ @StaticArray, %entry ], [ %arrayctor.next, %arrayctor.loop ]
+// CHECK-NEXT: %[[#CV_LOOP:]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token %[[#C_ENTRY]]) ]
// CHECK-NEXT: call void @hlsl::RWBuffer<float>::RWBuffer()(ptr {{.*}} %arrayctor.cur)
// CHECK-NEXT: %arrayctor.next = getelementptr inbounds %"class.hlsl::RWBuffer", ptr %arrayctor.cur, i32 1
// CHECK-NEXT: %arrayctor.done = icmp eq ptr %arrayctor.next, getelementptr inbounds nuw (i8, ptr @StaticArray, i32 8)
@@ -58,6 +63,7 @@
// constructor and not from binding. It will initalize the handle to poison.
// CHECK: define internal void @__cxx_global_var_init{{.*}}
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: call void @hlsl::RWStructuredBuffer<int>::RWStructuredBuffer()(ptr {{.*}} @StaticOneWithCounter)
// No other global initialization routines should be present.
@@ -67,6 +73,7 @@
void main() {
// CHECK: define internal void @main()()
// CHECK-NEXT: entry:
+// CHECK-NEXT: %[[#C_ENTRY:]] = call token @llvm.experimental.convergence.entry()
// CHECK-NEXT: %[[TMP0:.*]] = alloca %"class.hlsl::RWBuffer"
static RWBuffer<float> StaticLocal;
diff --git a/llvm/test/Transforms/IndVarSimplify/convergent-controlled-loop.ll b/llvm/test/Transforms/IndVarSimplify/convergent-controlled-loop.ll
new file mode 100644
index 0000000..c023a8e
--- /dev/null
+++ b/llvm/test/Transforms/IndVarSimplify/convergent-controlled-loop.ll
@@ -0,0 +1,70 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt < %s -passes='loop(indvars)' -S | FileCheck %s
+
+; IndVarSimplify tries to simplify or remove loop induction variables and exit
+; conditions that it can prove are redundant. This test ensures that the pass
+; preserves the exit condition of a loop whose body contains a convergent
+; operation guarded by a convergence.loop token.
+;
+; This is one of three sibling tests (LoopRotate, IndVarSimplify,
+; SimpleLoopUnswitch) that each guard against a pass moving or removing the
+; exit condition of a loop with a convergent operation, which can change the
+; set of active threads inside the loop. See https://godbolt.org/z/K1a8xW1TE
+; for the problematic pass pipeline and
+; https://github.com/llvm/llvm-project/issues/180621 for full context.
+
+define void @convergent_loop(i32 %tid, ptr %array) #0 {
+; CHECK-LABEL: @convergent_loop(
+; CHECK-NEXT: entry:
+; CHECK-NEXT: [[TOK_ENTRY:%.*]] = call token @llvm.experimental.convergence.entry()
+; CHECK-NEXT: br label [[FOR_COND:%.*]]
+; CHECK: for.cond:
+; CHECK-NEXT: [[I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC:%.*]], [[FOR_BODY:%.*]] ]
+; CHECK-NEXT: [[TOK_LOOP:%.*]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[TOK_ENTRY]]) ]
+; CHECK-NEXT: [[CMP:%.*]] = icmp samesign ult i32 [[I]], 8
+; CHECK-NEXT: br i1 [[CMP]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT:%.*]]
+; CHECK: for.body:
+; CHECK-NEXT: [[CMP1:%.*]] = icmp eq i32 [[I]], [[TID:%.*]]
+; CHECK-NEXT: [[INC]] = add nuw nsw i32 [[I]], 1
+; CHECK-NEXT: br i1 [[CMP1]], label [[IF_THEN:%.*]], label [[FOR_COND]]
+; CHECK: if.then:
+; CHECK-NEXT: [[WAVE:%.*]] = call i32 @llvm.dx.wave.reduce.umax.i32(i32 [[TID]]) [ "convergencectrl"(token [[TOK_LOOP]]) ]
+; CHECK-NEXT: [[GEP:%.*]] = getelementptr inbounds i32, ptr [[ARRAY:%.*]], i32 [[TID]]
+; CHECK-NEXT: store i32 [[WAVE]], ptr [[GEP]], align 4
+; CHECK-NEXT: br label [[FOR_END:%.*]]
+; CHECK: for.end.loopexit:
+; CHECK-NEXT: br label [[FOR_END]]
+; CHECK: for.end:
+; CHECK-NEXT: ret void
+;
+entry:
+ %tok.entry = call token @llvm.experimental.convergence.entry()
+ br label %for.cond
+
+for.cond:
+ %i = phi i32 [ 0, %entry ], [ %inc, %for.body ]
+ %tok.loop = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token %tok.entry) ]
+ %cmp = icmp ult i32 %i, 8
+ br i1 %cmp, label %for.body, label %for.end
+
+for.body:
+ %cmp1 = icmp eq i32 %i, %tid
+ %inc = add nuw i32 %i, 1
+ br i1 %cmp1, label %if.then, label %for.cond
+
+if.then:
+ %wave = call i32 @llvm.dx.wave.reduce.umax.i32(i32 %tid) [ "convergencectrl"(token %tok.loop) ]
+ %gep = getelementptr inbounds i32, ptr %array, i32 %tid
+ store i32 %wave, ptr %gep, align 4
+ br label %for.end
+
+for.end:
+ ret void
+}
+
+declare token @llvm.experimental.convergence.entry() #1
+declare token @llvm.experimental.convergence.loop() #1
+declare i32 @llvm.dx.wave.reduce.umax.i32(i32) #1
+
+attributes #0 = { convergent mustprogress nounwind "no-trapping-math"="true" }
+attributes #1 = { convergent }
diff --git a/llvm/test/Transforms/LoopRotate/convergent-controlled.ll b/llvm/test/Transforms/LoopRotate/convergent-controlled.ll
new file mode 100644
index 0000000..bef7617
--- /dev/null
+++ b/llvm/test/Transforms/LoopRotate/convergent-controlled.ll
@@ -0,0 +1,72 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt < %s -passes=loop-rotate -S | FileCheck %s
+
+; LoopRotate restructures a loop so the exit test appears at the end of the
+; loop body rather than the beginning. This test ensures the pass does not
+; rotate a loop whose header contains a convergence.loop token that guards a
+; convergent operation. Rotating the header would place the exit test before
+; the convergence token, which breaks the convergence contract and enables
+; later passes to collapse the loop into straight-line code.
+;
+; This is one of three sibling tests (LoopRotate, IndVarSimplify,
+; SimpleLoopUnswitch) that each guard against a pass moving or removing the
+; exit condition of a loop with a convergent operation, which can change the
+; set of active threads inside the loop. See https://godbolt.org/z/K1a8xW1TE
+; for the problematic pass pipeline and
+; https://github.com/llvm/llvm-project/issues/180621 for full context.
+
+define void @convergent_loop(i32 %tid, ptr %array) #0 {
+; CHECK-LABEL: @convergent_loop(
+; CHECK-NEXT: entry:
+; CHECK-NEXT: [[TOK_ENTRY:%.*]] = call token @llvm.experimental.convergence.entry()
+; CHECK-NEXT: br label [[FOR_COND:%.*]]
+; CHECK: for.cond:
+; CHECK-NEXT: [[I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC:%.*]], [[FOR_BODY:%.*]] ]
+; CHECK-NEXT: [[TOK_LOOP:%.*]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[TOK_ENTRY]]) ]
+; CHECK-NEXT: [[CMP:%.*]] = icmp ult i32 [[I]], 8
+; CHECK-NEXT: br i1 [[CMP]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT:%.*]]
+; CHECK: for.body:
+; CHECK-NEXT: [[CMP1:%.*]] = icmp eq i32 [[I]], [[TID:%.*]]
+; CHECK-NEXT: [[INC]] = add nuw i32 [[I]], 1
+; CHECK-NEXT: br i1 [[CMP1]], label [[IF_THEN:%.*]], label [[FOR_COND]]
+; CHECK: if.then:
+; CHECK-NEXT: [[WAVE:%.*]] = call i32 @llvm.dx.wave.reduce.umax.i32(i32 [[TID]]) [ "convergencectrl"(token [[TOK_LOOP]]) ]
+; CHECK-NEXT: [[GEP:%.*]] = getelementptr inbounds i32, ptr [[ARRAY:%.*]], i32 [[TID]]
+; CHECK-NEXT: store i32 [[WAVE]], ptr [[GEP]], align 4
+; CHECK-NEXT: br label [[FOR_END:%.*]]
+; CHECK: for.end.loopexit:
+; CHECK-NEXT: br label [[FOR_END]]
+; CHECK: for.end:
+; CHECK-NEXT: ret void
+;
+entry:
+ %tok.entry = call token @llvm.experimental.convergence.entry()
+ br label %for.cond
+
+for.cond:
+ %i = phi i32 [ 0, %entry ], [ %inc, %for.body ]
+ %tok.loop = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token %tok.entry) ]
+ %cmp = icmp ult i32 %i, 8
+ br i1 %cmp, label %for.body, label %for.end
+
+for.body:
+ %cmp1 = icmp eq i32 %i, %tid
+ %inc = add nuw i32 %i, 1
+ br i1 %cmp1, label %if.then, label %for.cond
+
+if.then:
+ %wave = call i32 @llvm.dx.wave.reduce.umax.i32(i32 %tid) [ "convergencectrl"(token %tok.loop) ]
+ %gep = getelementptr inbounds i32, ptr %array, i32 %tid
+ store i32 %wave, ptr %gep, align 4
+ br label %for.end
+
+for.end:
+ ret void
+}
+
+declare token @llvm.experimental.convergence.entry() #1
+declare token @llvm.experimental.convergence.loop() #1
+declare i32 @llvm.dx.wave.reduce.umax.i32(i32) #1
+
+attributes #0 = { convergent mustprogress nounwind "no-trapping-math"="true" }
+attributes #1 = { convergent }
diff --git a/llvm/test/Transforms/SimpleLoopUnswitch/convergent-controlled.ll b/llvm/test/Transforms/SimpleLoopUnswitch/convergent-controlled.ll
new file mode 100644
index 0000000..e764275
--- /dev/null
+++ b/llvm/test/Transforms/SimpleLoopUnswitch/convergent-controlled.ll
@@ -0,0 +1,70 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt < %s -passes='loop(simple-loop-unswitch)' -S | FileCheck %s
+
+; SimpleLoopUnswitch hoists loop-invariant branch conditions out of the loop by
+; duplicating the loop body under each branch. This test ensures the pass does
+; not hoist the exit condition of a loop whose body contains a convergent
+; operation guarded by a convergence.loop token.
+;
+; This is one of three sibling tests (LoopRotate, IndVarSimplify,
+; SimpleLoopUnswitch) that each guard against a pass moving or removing the
+; exit condition of a loop with a convergent operation, which can change the
+; set of active threads inside the loop. See https://godbolt.org/z/K1a8xW1TE
+; for the problematic pass pipeline and
+; https://github.com/llvm/llvm-project/issues/180621 for full context.
+
+define void @convergent_loop(i32 %tid, ptr %array) #0 {
+; CHECK-LABEL: @convergent_loop(
+; CHECK-NEXT: entry:
+; CHECK-NEXT: [[TOK_ENTRY:%.*]] = call token @llvm.experimental.convergence.entry()
+; CHECK-NEXT: br label [[FOR_COND:%.*]]
+; CHECK: for.cond:
+; CHECK-NEXT: [[I:%.*]] = phi i32 [ 0, [[ENTRY:%.*]] ], [ [[INC:%.*]], [[FOR_BODY:%.*]] ]
+; CHECK-NEXT: [[TOK_LOOP:%.*]] = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token [[TOK_ENTRY]]) ]
+; CHECK-NEXT: [[CMP:%.*]] = icmp ult i32 [[I]], 8
+; CHECK-NEXT: br i1 [[CMP]], label [[FOR_BODY]], label [[FOR_END_LOOPEXIT:%.*]]
+; CHECK: for.body:
+; CHECK-NEXT: [[CMP1:%.*]] = icmp eq i32 [[I]], [[TID:%.*]]
+; CHECK-NEXT: [[INC]] = add nuw i32 [[I]], 1
+; CHECK-NEXT: br i1 [[CMP1]], label [[IF_THEN:%.*]], label [[FOR_COND]]
+; CHECK: if.then:
+; CHECK-NEXT: [[WAVE:%.*]] = call i32 @llvm.dx.wave.reduce.umax.i32(i32 [[TID]]) [ "convergencectrl"(token [[TOK_LOOP]]) ]
+; CHECK-NEXT: [[GEP:%.*]] = getelementptr inbounds i32, ptr [[ARRAY:%.*]], i32 [[TID]]
+; CHECK-NEXT: store i32 [[WAVE]], ptr [[GEP]], align 4
+; CHECK-NEXT: br label [[FOR_END:%.*]]
+; CHECK: for.end.loopexit:
+; CHECK-NEXT: br label [[FOR_END]]
+; CHECK: for.end:
+; CHECK-NEXT: ret void
+;
+entry:
+ %tok.entry = call token @llvm.experimental.convergence.entry()
+ br label %for.cond
+
+for.cond:
+ %i = phi i32 [ 0, %entry ], [ %inc, %for.body ]
+ %tok.loop = call token @llvm.experimental.convergence.loop() [ "convergencectrl"(token %tok.entry) ]
+ %cmp = icmp ult i32 %i, 8
+ br i1 %cmp, label %for.body, label %for.end
+
+for.body:
+ %cmp1 = icmp eq i32 %i, %tid
+ %inc = add nuw i32 %i, 1
+ br i1 %cmp1, label %if.then, label %for.cond
+
+if.then:
+ %wave = call i32 @llvm.dx.wave.reduce.umax.i32(i32 %tid) [ "convergencectrl"(token %tok.loop) ]
+ %gep = getelementptr inbounds i32, ptr %array, i32 %tid
+ store i32 %wave, ptr %gep, align 4
+ br label %for.end
+
+for.end:
+ ret void
+}
+
+declare token @llvm.experimental.convergence.entry() #1
+declare token @llvm.experimental.convergence.loop() #1
+declare i32 @llvm.dx.wave.reduce.umax.i32(i32) #1
+
+attributes #0 = { convergent mustprogress nounwind "no-trapping-math"="true" }
+attributes #1 = { convergent }
