[AMDGPU] Diagnose unaligned atomic for amdgpu

[yaxunl]

rebase and simplify code

[yaxunl]

ping

[tra]

clang/lib/Driver/ToolChains/Clang.cpp

7215 ↗

(On Diff #528937)

Should it be done from HIPAMDToolChain::addClangWarningOptions ?

That's where Darwin does similar propotion from a waring to an error.

[MaskRay]

clang/test/Driver/hip-options.hip

144

Prefer --target= to deprecated (since 3.x) -target

[yaxunl]

clang/lib/Driver/ToolChains/Clang.cpp

7215 ↗

(On Diff #528937)

will do. Since this option is shared between all AMDGPU toolchains, I will add it to AMDGPUToolChain::addClangWarningOptions and let the derived toolchains call it.

clang/test/Driver/hip-options.hip

144

will fix

[yaxunl]

revised by Artem's and Fangrui's comments

clang/lib/Driver/ToolChains/AMDGPU.h

protected:
   /// Get target ID, GPU arch, and target ID features if the target ID is
   /// specified and valid.
   ParsedTargetIDType
   getParsedTargetID(const llvm::opt::ArgList &DriverArgs) const;
 
   /// Get GPU arch from -mcpu without checking.
   StringRef getGPUArch(const llvm::opt::ArgList &DriverArgs) const;
 
+  /// Common warning options shared by AMDGPU HIP, OpenCL and OpenMP toolchains.
+  /// Language specific warning options should go to derived classes.
+  void addClangWarningOptions(llvm::opt::ArgStringList &CC1Args) const override;
 };
 
 class LLVM_LIBRARY_VISIBILITY ROCMToolChain : public AMDGPUToolChain {
 public:
   ROCMToolChain(const Driver &D, const llvm::Triple &Triple,
                 const llvm::opt::ArgList &Args);
   void
   addClangTargetOptions(const llvm::opt::ArgList &DriverArgs,

clang/lib/Driver/ToolChains/AMDGPU.cpp

void AMDGPUToolChain::addClangTargetOptions(
   // supported for the foreseeable future.
   if (!DriverArgs.hasArg(options::OPT_fvisibility_EQ,
                          options::OPT_fvisibility_ms_compat)) {
     CC1Args.push_back("-fvisibility=hidden");
     CC1Args.push_back("-fapply-global-visibility-to-externs");
   }
 }
 
+void AMDGPUToolChain::addClangWarningOptions(ArgStringList &CC1Args) const {
+  // AMDGPU does not support atomic lib call. Treat atomic alignment
+  // warnings as errors.
+  CC1Args.push_back("-Werror=atomic-alignment");
+}
+
 StringRef
 AMDGPUToolChain::getGPUArch(const llvm::opt::ArgList &DriverArgs) const {
   return getProcessorFromTargetID(
       getTriple(), DriverArgs.getLastArgValue(options::OPT_mcpu_EQ));
 }
 
 AMDGPUToolChain::ParsedTargetIDType
 AMDGPUToolChain::getParsedTargetID(const llvm::opt::ArgList &DriverArgs) const {

clang/lib/Driver/ToolChains/AMDGPUOpenMP.cpp

DAL->AddJoinedArg(nullptr, Opts.getOption(options::OPT_march_EQ),
                       BoundArch);
   }
 
   return DAL;
 }
 
 void AMDGPUOpenMPToolChain::addClangWarningOptions(
     ArgStringList &CC1Args) const {
+  AMDGPUToolChain::addClangWarningOptions(CC1Args);
   HostTC.addClangWarningOptions(CC1Args);
 }
 
 ToolChain::CXXStdlibType
 AMDGPUOpenMPToolChain::GetCXXStdlibType(const ArgList &Args) const {
   return HostTC.GetCXXStdlibType(Args);
 }
 

clang/lib/Driver/ToolChains/HIPAMD.cpp

 }
 
 Tool *HIPAMDToolChain::buildLinker() const {
   assert(getTriple().getArch() == llvm::Triple::amdgcn);
   return new tools::AMDGCN::Linker(*this);
 }
 
 void HIPAMDToolChain::addClangWarningOptions(ArgStringList &CC1Args) const {
+  AMDGPUToolChain::addClangWarningOptions(CC1Args);
   HostTC.addClangWarningOptions(CC1Args);
 }
 
 ToolChain::CXXStdlibType
 HIPAMDToolChain::GetCXXStdlibType(const ArgList &Args) const {
   return HostTC.GetCXXStdlibType(Args);
 }
 

clang/test/Driver/amdgpu-openmp-toolchain.c

 // RUN:   --rocm-device-lib-path=%S/Inputs/rocm/amdgcn/bitcode -fopenmp-new-driver %s  2>&1 | \
 // RUN: FileCheck %s --check-prefix=CHECK-LIB-DEVICE
 // CHECK-LIB-DEVICE: "-cc1" {{.*}}ocml.bc"{{.*}}ockl.bc"{{.*}}oclc_daz_opt_on.bc"{{.*}}oclc_unsafe_math_off.bc"{{.*}}oclc_finite_only_off.bc"{{.*}}oclc_correctly_rounded_sqrt_on.bc"{{.*}}oclc_wavefrontsize64_on.bc"{{.*}}oclc_isa_version_803.bc"
 
 // RUN: %clang -### -target x86_64-pc-linux-gnu -fopenmp --offload-arch=gfx803 -nogpulib \
 // RUN:   --rocm-device-lib-path=%S/Inputs/rocm/amdgcn/bitcode -fopenmp-new-driver %s  2>&1 | \
 // RUN: FileCheck %s --check-prefix=CHECK-LIB-DEVICE-NOGPULIB
 // CHECK-LIB-DEVICE-NOGPULIB-NOT: "-cc1" {{.*}}ocml.bc"{{.*}}ockl.bc"{{.*}}oclc_daz_opt_on.bc"{{.*}}oclc_unsafe_math_off.bc"{{.*}}oclc_finite_only_off.bc"{{.*}}oclc_correctly_rounded_sqrt_on.bc"{{.*}}oclc_wavefrontsize64_on.bc"{{.*}}oclc_isa_version_803.bc"
+
+// RUN:   %clang -### --target=x86_64-unknown-linux-gnu -emit-llvm -S -fopenmp -fopenmp-targets=amdgcn-amd-amdhsa -Xopenmp-target=amdgcn-amd-amdhsa -march=gfx803 -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-WARN-ATOMIC
+// CHECK-WARN-ATOMIC-NOT: "-cc1" "-triple" "x86_64-unknown-linux-gnu"{{.*}}"-Werror=atomic-alignment"
+// CHECK-WARN-ATOMIC: "-cc1" "-triple" "amdgcn-amd-amdhsa"{{.*}}"-Werror=atomic-alignment"

clang/test/Driver/amdgpu-toolchain-opencl.cl

 // CHECK_Ofast: "-cc1"{{.*}} "-Ofast"
 // CHECK_O_DEFAULT: "-cc1"{{.*}} "-O3"
 
 // CHK-INC: "-cc1" {{.*}}"-finclude-default-header" "-fdeclare-opencl-builtins" {{.*}}"-x" "cl"
 // CHK-INC-NOT: "-cc1" {{.*}}"-finclude-default-header" "-fdeclare-opencl-builtins" {{.*}}"-x" "cpp-output"
 
 // RUN: %clang -### --target=amdgcn-amd-amdhsa-opencl -x cl -emit-llvm -mcpu=fiji %s 2>&1 | FileCheck -check-prefix=CHK-LINK %s
 // CHK-LINK: ld.lld{{.*}} "--no-undefined" "-shared"
+
+// RUN: %clang -### --target=amdgcn-amd-amdhsa-opencl -x cl -c -emit-llvm -mcpu=fiji %s 2>&1 | FileCheck -check-prefix=CHECK-WARN-ATOMIC %s
+// CHECK-WARN-ATOMIC: "-cc1"{{.*}} "-Werror=atomic-alignment"

clang/test/Driver/hip-options.hip

 
 // RUN: %clang -### -nogpuinc -nogpulib -mamdgpu-ieee -mno-amdgpu-ieee -ffast-math \
 // RUN:   --cuda-gpu-arch=gfx906  %s 2>&1 | FileCheck -check-prefixes=IEEE-OFF %s
 // IEEE-OFF: clang{{.*}} "-triple" "amdgcn-amd-amdhsa" {{.*}} "-mno-amdgpu-ieee"
 
 // RUN: %clang -### -nogpuinc -nogpulib -mamdgpu-ieee -mno-amdgpu-ieee -ffast-math \
 // RUN:   --cuda-gpu-arch=gfx906  %s 2>&1 | FileCheck -check-prefixes=IEEE-OFF-NEG %s
 // IEEE-OFF-NEG-NOT: clang{{.*}} "-triple" "amdgcn-amd-amdhsa" {{.*}} "-mamdgpu-ieee"
+
+// Check -Werror=atomic-alignment is passed for amdpu by default.
+
+// RUN: %clang -### --target=x86_64-unknown-linux-gnu -nogpuinc -nogpulib \
+// RUN:   --cuda-gpu-arch=gfx906 %s 2>&1 | FileCheck -check-prefix=WARN-ATOMIC %s
+// WARN-ATOMIC: clang{{.*}} "-triple" "amdgcn-amd-amdhsa" {{.*}} "-Werror=atomic-alignment"
+// WARN-ATOMIC-NOT: clang{{.*}} "-triple" "x86_64-unknown-linux-gnu" {{.*}} "-Werror=atomic-alignment"