[Clang][SPIR-V] Fix convergence tokens for dtor #133469
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Destructor calls were emitted without convergence intrinsics when building for SPIR-V, which means invalid IR since we mixed controlled and non-controlled convergence.
llvmbot
added
clang
|
@llvm/pr-subscribers-hlsl @llvm/pr-subscribers-clang Author: Nathan Gauër (Keenuts) ChangesDestructor calls were emitted without convergence Full diff: https://github.com/llvm/llvm-project/pull/133469.diff 2 Files Affected:
diff --git a/clang/lib/CodeGen/CGDeclCXX.cpp b/clang/lib/CodeGen/CGDeclCXX.cpp
index 1ad34ae61f96a..b142aaf3fe0ee 100644
--- a/clang/lib/CodeGen/CGDeclCXX.cpp
+++ b/clang/lib/CodeGen/CGDeclCXX.cpp
@@ -1152,7 +1152,7 @@ void CodeGenFunction::GenerateCXXGlobalCleanUpFunc(
llvm::Constant *Arg;
std::tie(CalleeTy, Callee, Arg) = DtorsOrStermFinalizers[e - i - 1];
- llvm::CallInst *CI = nullptr;
+ llvm::CallBase *CI = nullptr;
if (Arg == nullptr) {
assert(
CGM.getCXXABI().useSinitAndSterm() &&
@@ -1164,6 +1164,9 @@ void CodeGenFunction::GenerateCXXGlobalCleanUpFunc(
// Make sure the call and the callee agree on calling convention.
if (llvm::Function *F = dyn_cast<llvm::Function>(Callee))
CI->setCallingConv(F->getCallingConv());
+
+ if (CGM.shouldEmitConvergenceTokens() && CI->isConvergent())
+ CI = addConvergenceControlToken(CI);
}
}
diff --git a/clang/test/CodeGenHLSL/GlobalDestructors.hlsl b/clang/test/CodeGenHLSL/GlobalDestructors.hlsl
index f98318601134b..9f90971bafd05 100644
--- a/clang/test/CodeGenHLSL/GlobalDestructors.hlsl
+++ b/clang/test/CodeGenHLSL/GlobalDestructors.hlsl
@@ -1,5 +1,6 @@
-// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.0-compute -emit-llvm -disable-llvm-passes %s -o - | FileCheck %s --check-prefixes=CS,NOINLINE,CHECK
-// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.3-library -emit-llvm -disable-llvm-passes %s -o - | FileCheck %s --check-prefixes=LIB,NOINLINE,CHECK
+// RUN: %clang_cc1 -triple spirv-unknown-vulkan1.3-compute -emit-llvm -disable-llvm-passes %s -o - | FileCheck %s --check-prefixes=CS,NOINLINE-SPIRV,CHECK
+// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.0-compute -emit-llvm -disable-llvm-passes %s -o - | FileCheck %s --check-prefixes=CS,NOINLINE-DXIL,CHECK
+// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.3-library -emit-llvm -disable-llvm-passes %s -o - | FileCheck %s --check-prefixes=LIB,NOINLINE-DXIL,CHECK
// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.0-compute -emit-llvm -O0 %s -o - | FileCheck %s --check-prefixes=INLINE,CHECK
// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.3-library -emit-llvm -O0 %s -o - | FileCheck %s --check-prefixes=INLINE,CHECK
@@ -57,11 +58,19 @@ void main(unsigned GI : SV_GroupIndex) {
// CHECK: define void @main()
// CHECK-NEXT: entry:
// Verify destructor is emitted
-// NOINLINE-NEXT: call void @_GLOBAL__sub_I_GlobalDestructors.hlsl()
-// NOINLINE-NEXT: %0 = call i32 @llvm.dx.flattened.thread.id.in.group()
-// NOINLINE-NEXT: call void @_Z4mainj(i32 %0)
-// NOINLINE-NEXT: call void @_GLOBAL__D_a()
-// NOINLINE-NEXT: ret void
+// 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: 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: ret void
+
// Verify inlining leaves only calls to "llvm." intrinsics
// INLINE-NOT: call {{[^@]*}} @{{[^l][^l][^v][^m][^\.]}}
// INLINE: ret void
@@ -69,10 +78,17 @@ void main(unsigned GI : SV_GroupIndex) {
// This is really just a sanity check I needed for myself to verify that
// function scope static variables also get destroyed properly.
-// NOINLINE: define internal void @_GLOBAL__D_a() [[IntAttr:\#[0-9]+]]
-// NOINLINE-NEXT: entry:
-// NOINLINE-NEXT: call void @_ZN4TailD1Ev(ptr @_ZZ3WagvE1T)
-// NOINLINE-NEXT: call void @_ZN6PupperD1Ev(ptr @GlobalPup)
-// NOINLINE-NEXT: ret void
+// 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-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 @_ZZ3WagvE1T) [ "convergencectrl"(token %0) ]
+// NOINLINE-SPIRV-NEXT: call spir_func void @_ZN6PupperD1Ev(ptr @GlobalPup) [ "convergencectrl"(token %0) ]
+// NOINLINE-SPIRV-NEXT: ret void
// NOINLINE: attributes [[IntAttr]] = {{.*}} alwaysinline
|
|
@llvm/pr-subscribers-clang-codegen Author: Nathan Gauër (Keenuts) ChangesDestructor calls were emitted without convergence Full diff: https://github.com/llvm/llvm-project/pull/133469.diff 2 Files Affected:
diff --git a/clang/lib/CodeGen/CGDeclCXX.cpp b/clang/lib/CodeGen/CGDeclCXX.cpp
index 1ad34ae61f96a..b142aaf3fe0ee 100644
--- a/clang/lib/CodeGen/CGDeclCXX.cpp
+++ b/clang/lib/CodeGen/CGDeclCXX.cpp
@@ -1152,7 +1152,7 @@ void CodeGenFunction::GenerateCXXGlobalCleanUpFunc(
llvm::Constant *Arg;
std::tie(CalleeTy, Callee, Arg) = DtorsOrStermFinalizers[e - i - 1];
- llvm::CallInst *CI = nullptr;
+ llvm::CallBase *CI = nullptr;
if (Arg == nullptr) {
assert(
CGM.getCXXABI().useSinitAndSterm() &&
@@ -1164,6 +1164,9 @@ void CodeGenFunction::GenerateCXXGlobalCleanUpFunc(
// Make sure the call and the callee agree on calling convention.
if (llvm::Function *F = dyn_cast<llvm::Function>(Callee))
CI->setCallingConv(F->getCallingConv());
+
+ if (CGM.shouldEmitConvergenceTokens() && CI->isConvergent())
+ CI = addConvergenceControlToken(CI);
}
}
diff --git a/clang/test/CodeGenHLSL/GlobalDestructors.hlsl b/clang/test/CodeGenHLSL/GlobalDestructors.hlsl
index f98318601134b..9f90971bafd05 100644
--- a/clang/test/CodeGenHLSL/GlobalDestructors.hlsl
+++ b/clang/test/CodeGenHLSL/GlobalDestructors.hlsl
@@ -1,5 +1,6 @@
-// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.0-compute -emit-llvm -disable-llvm-passes %s -o - | FileCheck %s --check-prefixes=CS,NOINLINE,CHECK
-// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.3-library -emit-llvm -disable-llvm-passes %s -o - | FileCheck %s --check-prefixes=LIB,NOINLINE,CHECK
+// RUN: %clang_cc1 -triple spirv-unknown-vulkan1.3-compute -emit-llvm -disable-llvm-passes %s -o - | FileCheck %s --check-prefixes=CS,NOINLINE-SPIRV,CHECK
+// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.0-compute -emit-llvm -disable-llvm-passes %s -o - | FileCheck %s --check-prefixes=CS,NOINLINE-DXIL,CHECK
+// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.3-library -emit-llvm -disable-llvm-passes %s -o - | FileCheck %s --check-prefixes=LIB,NOINLINE-DXIL,CHECK
// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.0-compute -emit-llvm -O0 %s -o - | FileCheck %s --check-prefixes=INLINE,CHECK
// RUN: %clang_cc1 -triple dxil-pc-shadermodel6.3-library -emit-llvm -O0 %s -o - | FileCheck %s --check-prefixes=INLINE,CHECK
@@ -57,11 +58,19 @@ void main(unsigned GI : SV_GroupIndex) {
// CHECK: define void @main()
// CHECK-NEXT: entry:
// Verify destructor is emitted
-// NOINLINE-NEXT: call void @_GLOBAL__sub_I_GlobalDestructors.hlsl()
-// NOINLINE-NEXT: %0 = call i32 @llvm.dx.flattened.thread.id.in.group()
-// NOINLINE-NEXT: call void @_Z4mainj(i32 %0)
-// NOINLINE-NEXT: call void @_GLOBAL__D_a()
-// NOINLINE-NEXT: ret void
+// 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: 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: ret void
+
// Verify inlining leaves only calls to "llvm." intrinsics
// INLINE-NOT: call {{[^@]*}} @{{[^l][^l][^v][^m][^\.]}}
// INLINE: ret void
@@ -69,10 +78,17 @@ void main(unsigned GI : SV_GroupIndex) {
// This is really just a sanity check I needed for myself to verify that
// function scope static variables also get destroyed properly.
-// NOINLINE: define internal void @_GLOBAL__D_a() [[IntAttr:\#[0-9]+]]
-// NOINLINE-NEXT: entry:
-// NOINLINE-NEXT: call void @_ZN4TailD1Ev(ptr @_ZZ3WagvE1T)
-// NOINLINE-NEXT: call void @_ZN6PupperD1Ev(ptr @GlobalPup)
-// NOINLINE-NEXT: ret void
+// 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-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 @_ZZ3WagvE1T) [ "convergencectrl"(token %0) ]
+// NOINLINE-SPIRV-NEXT: call spir_func void @_ZN6PupperD1Ev(ptr @GlobalPup) [ "convergencectrl"(token %0) ]
+// NOINLINE-SPIRV-NEXT: ret void
// NOINLINE: attributes [[IntAttr]] = {{.*}} alwaysinline
|
s-perron
reviewed
Mar 28, 2025
s-perron
approved these changes
Mar 31, 2025
Ankur-0429
pushed a commit
to Ankur-0429/llvm-project
that referenced
this pull request
Apr 2, 2025
Destructor calls were emitted without convergence intrinsics when building for SPIR-V, which means invalid IR since we mixed controlled and non-controlled convergence.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Destructor calls were emitted without convergence
intrinsics when building for SPIR-V, which means invalid IR since we mixed controlled and non-controlled convergence.