[clang] [Clang] Disable RTTI for offloading at the frontend level (PR #127082)
Joseph Huber via cfe-commits
cfe-commits at lists.llvm.org
Thu Feb 13 07:47:45 PST 2025
https://github.com/jhuber6 created https://github.com/llvm/llvm-project/pull/127082
Summary:
Currently we force this to be off in `-cc1`, but I feel like this should
be something the driver informs the compiler of. It's not
*inconcievable* that RTTI might 'work' on the GPU some day, but mostly I
just don't like having these kinds of checks hard-coded on a triple +
language.
>From b1497c836fd67a03810dd5806b2e8ea43acefd86 Mon Sep 17 00:00:00 2001
From: Joseph Huber <huberjn at outlook.com>
Date: Thu, 13 Feb 2025 09:27:24 -0600
Subject: [PATCH] [Clang] Disable RTTI for offloading at the frontend level
Summary:
Currently we force this to be off in `-cc1`, but I feel like this should
be something the driver informs the compiler of. It's not
*inconcievable* that RTTI might 'work' on the GPU some day, but mostly I
just don't like having these kinds of checks hard-coded on a triple +
language.
---
clang/lib/CodeGen/CodeGenModule.h | 5 +----
clang/lib/Driver/ToolChains/Clang.cpp | 6 ++++++
clang/test/CodeGenCUDA/device-vtable.cu | 2 +-
clang/test/CodeGenCUDA/record-layout.cu | 2 +-
clang/test/Headers/hip-header.hip | 14 +++++++-------
.../OpenMP/nvptx_target_pure_deleted_codegen.cpp | 6 +++---
6 files changed, 19 insertions(+), 16 deletions(-)
diff --git a/clang/lib/CodeGen/CodeGenModule.h b/clang/lib/CodeGen/CodeGenModule.h
index c6f6fd5b9a7bd..2d7835fcfc4e9 100644
--- a/clang/lib/CodeGen/CodeGenModule.h
+++ b/clang/lib/CodeGen/CodeGenModule.h
@@ -1065,10 +1065,7 @@ class CodeGenModule : public CodeGenTypeCache {
// Return whether RTTI information should be emitted for this target.
bool shouldEmitRTTI(bool ForEH = false) {
- return (ForEH || getLangOpts().RTTI) && !getLangOpts().CUDAIsDevice &&
- !(getLangOpts().OpenMP && getLangOpts().OpenMPIsTargetDevice &&
- (getTriple().isNVPTX() || getTriple().isAMDGPU() ||
- getTriple().isSPIRV()));
+ return (ForEH || getLangOpts().RTTI);
}
/// Get the address of the RTTI descriptor for the given type.
diff --git a/clang/lib/Driver/ToolChains/Clang.cpp b/clang/lib/Driver/ToolChains/Clang.cpp
index 55ec3db0ee994..fb85c113d898b 100644
--- a/clang/lib/Driver/ToolChains/Clang.cpp
+++ b/clang/lib/Driver/ToolChains/Clang.cpp
@@ -7115,6 +7115,12 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA,
CmdArgs.push_back("-fgpu-exclude-wrong-side-overloads");
CmdArgs.push_back("-fgpu-defer-diag");
}
+ }
+
+ // The offloading devices do not support RTTI.
+ if (IsCudaDevice || IsHIPDevice || (IsOpenMPDevice && Triple.isNVPTX() &&
+ Triple.isAMDGCN() && Triple.isSPIRV())) {
+ CmdArgs.push_back("-fno-rtti");
}
// Forward --no-offloadlib to -cc1.
diff --git a/clang/test/CodeGenCUDA/device-vtable.cu b/clang/test/CodeGenCUDA/device-vtable.cu
index fa4ca0c8a4708..b1ab51dd74087 100644
--- a/clang/test/CodeGenCUDA/device-vtable.cu
+++ b/clang/test/CodeGenCUDA/device-vtable.cu
@@ -8,7 +8,7 @@
// RUN: %clang_cc1 -triple x86_64-unknown-linux-gnu -emit-llvm -o - %s \
// RUN: | FileCheck %s -check-prefix=CHECK-HOST -check-prefix=CHECK-BOTH
-// RUN: %clang_cc1 -triple nvptx64-nvidia-cuda -fcuda-is-device -emit-llvm -o - %s \
+// RUN: %clang_cc1 -triple nvptx64-nvidia-cuda -fcuda-is-device -fno-rtti -emit-llvm -o - %s \
// RUN: | FileCheck %s -check-prefix=CHECK-DEVICE -check-prefix=CHECK-BOTH
#include "Inputs/cuda.h"
diff --git a/clang/test/CodeGenCUDA/record-layout.cu b/clang/test/CodeGenCUDA/record-layout.cu
index 847a81d88d280..0fc1318fc79cc 100644
--- a/clang/test/CodeGenCUDA/record-layout.cu
+++ b/clang/test/CodeGenCUDA/record-layout.cu
@@ -3,7 +3,7 @@
// RUN: cat %t | FileCheck --check-prefixes=CHECK,HOST %s
// RUN: %clang_cc1 -fcuda-is-device -triple amdgcn-amd-amdhsa -target-cpu gfx1100 \
// RUN: -emit-llvm -fdump-record-layouts -aux-triple x86_64-pc-windows-msvc \
-// RUN: -o %t -xhip %s | FileCheck %s --check-prefix=AST
+// RUN: -fno-rtti -o %t -xhip %s | FileCheck %s --check-prefix=AST
// RUN: cat %t | FileCheck --check-prefixes=CHECK,DEV %s
#include "Inputs/cuda.h"
diff --git a/clang/test/Headers/hip-header.hip b/clang/test/Headers/hip-header.hip
index 146a43b643dba..b99b2ff39357b 100644
--- a/clang/test/Headers/hip-header.hip
+++ b/clang/test/Headers/hip-header.hip
@@ -3,47 +3,47 @@
// RUN: -internal-isystem %S/../../lib/Headers/cuda_wrappers \
// RUN: -internal-isystem %S/Inputs/include \
// RUN: -triple amdgcn-amd-amdhsa -aux-triple x86_64-unknown-unknown \
-// RUN: -target-cpu gfx906 -emit-llvm %s -fcuda-is-device -o - \
+// RUN: -target-cpu gfx906 -emit-llvm %s -fno-rtti -fcuda-is-device -o - \
// RUN: -D__HIPCC_RTC__ | FileCheck -check-prefixes=CHECK,NOMALLOC %s
// RUN: %clang_cc1 -include __clang_hip_runtime_wrapper.h \
// RUN: -internal-isystem %S/../../lib/Headers/cuda_wrappers \
// RUN: -internal-isystem %S/Inputs/include \
// RUN: -include cmath \
// RUN: -triple amdgcn-amd-amdhsa -aux-triple x86_64-unknown-unknown \
-// RUN: -target-cpu gfx906 -emit-llvm %s -fcuda-is-device -o - \
+// RUN: -target-cpu gfx906 -emit-llvm %s -fno-rtti -fcuda-is-device -o - \
// RUN: -D__HIPCC_RTC__ | FileCheck %s -check-prefixes=AMD_BOOL_RETURN
// RUN: %clang_cc1 -include __clang_hip_runtime_wrapper.h \
// RUN: -internal-isystem %S/../../lib/Headers/cuda_wrappers \
// RUN: -internal-isystem %S/Inputs/include \
// RUN: -include cmath \
// RUN: -triple amdgcn-amd-amdhsa -aux-triple x86_64-unknown-unknown \
-// RUN: -target-cpu gfx906 -emit-llvm %s -fcuda-is-device -o - \
+// RUN: -target-cpu gfx906 -emit-llvm %s -fno-rtti -fcuda-is-device -o - \
// RUN: -D__HIPCC_RTC__ -DUSE_ISNAN_WITH_INT_RETURN | FileCheck %s -check-prefixes=AMD_INT_RETURN
// RUN: %clang_cc1 -include __clang_hip_runtime_wrapper.h \
// RUN: -internal-isystem %S/../../lib/Headers/cuda_wrappers \
// RUN: -internal-isystem %S/Inputs/include \
// RUN: -triple amdgcn-amd-amdhsa -aux-triple x86_64-unknown-unknown \
-// RUN: -target-cpu gfx906 -emit-llvm %s -fcuda-is-device -o - \
+// RUN: -target-cpu gfx906 -emit-llvm %s -fno-rtti -fcuda-is-device -o - \
// RUN: -D__HIPCC_RTC__ -std=c++14 | FileCheck -check-prefixes=CHECK,CXX14 %s
// RUN: %clang_cc1 -include __clang_hip_runtime_wrapper.h \
// RUN: -internal-isystem %S/../../lib/Headers/cuda_wrappers \
// RUN: -internal-isystem %S/Inputs/include \
// RUN: -triple amdgcn-amd-amdhsa -aux-triple x86_64-unknown-unknown \
-// RUN: -target-cpu gfx906 -emit-llvm %s -fcuda-is-device -o - \
+// RUN: -target-cpu gfx906 -emit-llvm %s -fno-rtti -fcuda-is-device -o - \
// RUN: -DHIP_VERSION_MAJOR=4 -DHIP_VERSION_MINOR=5 \
// RUN: -D__HIPCC_RTC__ -disable-llvm-passes | FileCheck -check-prefixes=MALLOC %s
// RUN: %clang_cc1 -include __clang_hip_runtime_wrapper.h \
// RUN: -internal-isystem %S/../../lib/Headers/cuda_wrappers \
// RUN: -internal-isystem %S/Inputs/include \
// RUN: -triple amdgcn-amd-amdhsa -aux-triple x86_64-unknown-unknown \
-// RUN: -target-cpu gfx906 -emit-llvm %s -fcuda-is-device -o - \
+// RUN: -target-cpu gfx906 -emit-llvm %s -fno-rtti -fcuda-is-device -o - \
// RUN: -DHIP_VERSION_MAJOR=4 -DHIP_VERSION_MINOR=5 \
// RUN: -disable-llvm-passes | FileCheck -check-prefixes=MALLOC %s
// RUN: %clang_cc1 -include __clang_hip_runtime_wrapper.h \
// RUN: -internal-isystem %S/../../lib/Headers/cuda_wrappers \
// RUN: -internal-isystem %S/Inputs/include \
// RUN: -triple amdgcn-amd-amdhsa -aux-triple x86_64-unknown-unknown \
-// RUN: -target-cpu gfx906 -emit-llvm %s -fcuda-is-device -o - \
+// RUN: -target-cpu gfx906 -emit-llvm %s -fno-rtti -fcuda-is-device -o - \
// RUN: -DHIP_VERSION_MAJOR=4 -DHIP_VERSION_MINOR=5 \
// RUN: -fsanitize=address -disable-llvm-passes -D__HIPCC_RTC__ \
// RUN: | FileCheck -check-prefixes=MALLOC-ASAN %s
diff --git a/clang/test/OpenMP/nvptx_target_pure_deleted_codegen.cpp b/clang/test/OpenMP/nvptx_target_pure_deleted_codegen.cpp
index 071a501b66398..cb1f563571fe3 100644
--- a/clang/test/OpenMP/nvptx_target_pure_deleted_codegen.cpp
+++ b/clang/test/OpenMP/nvptx_target_pure_deleted_codegen.cpp
@@ -1,9 +1,9 @@
// Test target codegen - host bc file has to be created first.
// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple powerpc64le-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm-bc %s -o %t-ppc-host.bc
-// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fopenmp-is-target-device -fvisibility=protected -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx64-unknown-unknown -fopenmp-targets=nvptx64-nvidia-cuda -emit-llvm %s -fno-rtti -fopenmp-is-target-device -fvisibility=protected -fopenmp-host-ir-file-path %t-ppc-host.bc -o - | FileCheck %s
// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple i386-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm-bc %s -o %t-x86-host.bc
-// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-target-device -fvisibility=protected -fopenmp-host-ir-file-path %t-x86-host.bc -o - -fno-rtti | FileCheck %s
-// RUN: %clang_cc1 -verify -fopenmp -fexceptions -fcxx-exceptions -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fopenmp-is-target-device -fvisibility=protected -fopenmp-host-ir-file-path %t-x86-host.bc -o - -fno-rtti | FileCheck %s
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fno-rtti -fopenmp-is-target-device -fvisibility=protected -fopenmp-host-ir-file-path %t-x86-host.bc -o - -fno-rtti | FileCheck %s
+// RUN: %clang_cc1 -verify -fopenmp -fexceptions -fcxx-exceptions -x c++ -triple nvptx-unknown-unknown -fopenmp-targets=nvptx-nvidia-cuda -emit-llvm %s -fno-rtti -fopenmp-is-target-device -fvisibility=protected -fopenmp-host-ir-file-path %t-x86-host.bc -o - -fno-rtti | FileCheck %s
// expected-no-diagnostics
#ifndef HEADER
More information about the cfe-commits
mailing list