This is an archive of the discontinued LLVM Phabricator instance.

Differential D124466

[CUDA][HIP] Fix linkage of __clang_gpu_used_external
ClosedPublic

Authored by yaxunl on Apr 26 2022, 10:10 AM.

Details

Reviewers
tra
Commits
rG57a210e5b705: [CUDA][HIP] Fix linkage of __clang_gpu_used_external
Summary

Different TU's may have this globl var. appending linkage can
only be used with lld recognized special variables.

Change it to internal linkage.

Diff Detail

Event Timeline

yaxunl created this revision.Apr 26 2022, 10:10 AM
Herald added a project: Restricted Project. · View Herald TranscriptApr 26 2022, 10:10 AM
Herald added a subscriber: mattd. · View Herald Transcript
yaxunl requested review of this revision.Apr 26 2022, 10:10 AM
tra accepted this revision.Apr 26 2022, 11:43 AM
This revision is now accepted and ready to land.Apr 26 2022, 11:43 AM
Harbormaster completed remote builds in B161427: Diff 425255.Apr 26 2022, 11:58 AM
Closed by commit rG57a210e5b705: [CUDA][HIP] Fix linkage of __clang_gpu_used_external (authored by yaxunl). · Explain WhyApr 26 2022, 5:44 PM
This revision was automatically updated to reflect the committed changes.
yaxunl added a commit: rG57a210e5b705: [CUDA][HIP] Fix linkage of __clang_gpu_used_external.
Herald added a project: Restricted Project. · View Herald TranscriptApr 26 2022, 5:44 PM

Revision Contents

PathSize
clang/
lib/
CodeGen/
2 lines
test/
CodeGenCUDA/
2 lines

Diff 425371

clang/lib/CodeGen/CodeGenModule.cpp

  • This file is larger than 256 KB, so syntax highlighting is disabled by default.
Show First 20 Lines • Show All 592 Lines • ▼ Show 20 Lines for (auto D : Context.CUDAExternalDeviceDeclODRUsedByHost) {
GD = GlobalDecl(D); GD = GlobalDecl(D);
UsedArray.push_back(llvm::ConstantExpr::getPointerBitCastOrAddrSpaceCast( UsedArray.push_back(llvm::ConstantExpr::getPointerBitCastOrAddrSpaceCast(
GetAddrOfGlobal(GD), Int8PtrTy)); GetAddrOfGlobal(GD), Int8PtrTy));
} }
llvm::ArrayType *ATy = llvm::ArrayType::get(Int8PtrTy, UsedArray.size()); llvm::ArrayType *ATy = llvm::ArrayType::get(Int8PtrTy, UsedArray.size());
auto *GV = new llvm::GlobalVariable( auto *GV = new llvm::GlobalVariable(
getModule(), ATy, false, llvm::GlobalValue::AppendingLinkage, getModule(), ATy, false, llvm::GlobalValue::InternalLinkage,
llvm::ConstantArray::get(ATy, UsedArray), "__clang_gpu_used_external"); llvm::ConstantArray::get(ATy, UsedArray), "__clang_gpu_used_external");
addCompilerUsedGlobal(GV); addCompilerUsedGlobal(GV);
} }
emitLLVMUsed(); emitLLVMUsed();
if (SanStats) if (SanStats)
SanStats->finish(); SanStats->finish();
▲ Show 20 Lines • Show All 6,211 Lines • Show Last 20 Lines

clang/test/CodeGenCUDA/host-used-extern.cu

// RUN: %clang_cc1 -triple amdgcn-amd-amdhsa -fcuda-is-device -x hip %s \ // RUN: %clang_cc1 -triple amdgcn-amd-amdhsa -fcuda-is-device -x hip %s \
// RUN: -fgpu-rdc -std=c++11 -emit-llvm -o - -target-cpu gfx906 | FileCheck %s // RUN: -fgpu-rdc -std=c++11 -emit-llvm -o - -target-cpu gfx906 | FileCheck %s
// RUN: %clang_cc1 -triple amdgcn-amd-amdhsa -fcuda-is-device -x hip %s \ // RUN: %clang_cc1 -triple amdgcn-amd-amdhsa -fcuda-is-device -x hip %s \
// RUN: -fgpu-rdc -std=c++11 -emit-llvm -o - -target-cpu gfx906 \ // RUN: -fgpu-rdc -std=c++11 -emit-llvm -o - -target-cpu gfx906 \
// RUN: | FileCheck -check-prefix=NEG %s // RUN: | FileCheck -check-prefix=NEG %s
// RUN: %clang_cc1 -triple amdgcn-amd-amdhsa -fcuda-is-device -x hip %s \ // RUN: %clang_cc1 -triple amdgcn-amd-amdhsa -fcuda-is-device -x hip %s \
// RUN: -std=c++11 -emit-llvm -o - -target-cpu gfx906 \ // RUN: -std=c++11 -emit-llvm -o - -target-cpu gfx906 \
// RUN: | FileCheck -check-prefixes=NEG,NORDC %s // RUN: | FileCheck -check-prefixes=NEG,NORDC %s
#include "Inputs/cuda.h" #include "Inputs/cuda.h"
// CHECK-LABEL: @__clang_gpu_used_external = appending {{.*}}global // CHECK-LABEL: @__clang_gpu_used_external = internal {{.*}}global
// CHECK-DAG: @_Z7kernel1v // CHECK-DAG: @_Z7kernel1v
// CHECK-DAG: @_Z7kernel4v // CHECK-DAG: @_Z7kernel4v
// CHECK-DAG: @var1 // CHECK-DAG: @var1
// CHECK-LABEL: @llvm.compiler.used = {{.*}} @__clang_gpu_used_external // CHECK-LABEL: @llvm.compiler.used = {{.*}} @__clang_gpu_used_external
// NEG-NOT: @__clang_gpu_used_external = {{.*}} @_Z7kernel2v // NEG-NOT: @__clang_gpu_used_external = {{.*}} @_Z7kernel2v
// NEG-NOT: @__clang_gpu_used_external = {{.*}} @_Z7kernel3v // NEG-NOT: @__clang_gpu_used_external = {{.*}} @_Z7kernel3v
// NEG-NOT: @__clang_gpu_used_external = {{.*}} @var2 // NEG-NOT: @__clang_gpu_used_external = {{.*}} @var2
Show All 29 Lines