[Clang] Fix C library wrappers for offloading (#99716)

Summary:
This block of code wraps around the standard C library includes.
However, the order C library includes are presented is actually
important. If they are visible before the `libc++` headers then it will
cause errors. This patch simply moves the logic to just before it is
normally done. A more optimal solution would be to put this in the
toolchain, however doing it correctly would require knowing the
offloading kind and that would require rewriting the function signature
in all 30 or so ToolChains.

Author: jhuber6

clang/lib/Driver/ToolChains/Clang.cpp

@@ -1077,33 +1077,6 @@ void Clang::AddPreprocessingOptions(Compilation &C, const JobAction &JA,
   if (JA.isOffloading(Action::OFK_HIP))
     getToolChain().AddHIPIncludeArgs(Args, CmdArgs);
 
-  // If we are compiling for a GPU target we want to override the system headers
-  // with ones created by the 'libc' project if present.
-  if (!Args.hasArg(options::OPT_nostdinc) &&
-      !Args.hasArg(options::OPT_nogpuinc) &&
-      !Args.hasArg(options::OPT_nobuiltininc)) {
-    // Without an offloading language we will include these headers directly.
-    // Offloading languages will instead only use the declarations stored in
-    // the resource directory at clang/lib/Headers/llvm_libc_wrappers.
-    if ((getToolChain().getTriple().isNVPTX() ||
-         getToolChain().getTriple().isAMDGCN()) &&
-        C.getActiveOffloadKinds() == Action::OFK_None) {
-      SmallString<128> P(llvm::sys::path::parent_path(D.Dir));
-      llvm::sys::path::append(P, "include");
-      llvm::sys::path::append(P, getToolChain().getTripleString());
-      CmdArgs.push_back("-internal-isystem");
-      CmdArgs.push_back(Args.MakeArgString(P));
-    } else if (C.getActiveOffloadKinds() == Action::OFK_OpenMP) {
-      // TODO: CUDA / HIP include their own headers for some common functions
-      // implemented here. We'll need to clean those up so they do not conflict.
-      SmallString<128> P(D.ResourceDir);
-      llvm::sys::path::append(P, "include");
-      llvm::sys::path::append(P, "llvm_libc_wrappers");
-      CmdArgs.push_back("-internal-isystem");
-      CmdArgs.push_back(Args.MakeArgString(P));
-    }
-  }
-
   // If we are offloading to a target via OpenMP we need to include the
   // openmp_wrappers folder which contains alternative system headers.
   if (JA.isDeviceOffloading(Action::OFK_OpenMP) &&
@@ -1276,6 +1249,35 @@ void Clang::AddPreprocessingOptions(Compilation &C, const JobAction &JA,
         });
   }
 
+  // If we are compiling for a GPU target we want to override the system headers
+  // with ones created by the 'libc' project if present.
+  // TODO: This should be moved to `AddClangSystemIncludeArgs` by passing the
+  //       OffloadKind as an argument.
+  if (!Args.hasArg(options::OPT_nostdinc) &&
+      !Args.hasArg(options::OPT_nogpuinc) &&
+      !Args.hasArg(options::OPT_nobuiltininc)) {
+    // Without an offloading language we will include these headers directly.
+    // Offloading languages will instead only use the declarations stored in
+    // the resource directory at clang/lib/Headers/llvm_libc_wrappers.
+    if ((getToolChain().getTriple().isNVPTX() ||
+         getToolChain().getTriple().isAMDGCN()) &&
+        C.getActiveOffloadKinds() == Action::OFK_None) {
+      SmallString<128> P(llvm::sys::path::parent_path(D.Dir));
+      llvm::sys::path::append(P, "include");
+      llvm::sys::path::append(P, getToolChain().getTripleString());
+      CmdArgs.push_back("-internal-isystem");
+      CmdArgs.push_back(Args.MakeArgString(P));
+    } else if (C.getActiveOffloadKinds() == Action::OFK_OpenMP) {
+      // TODO: CUDA / HIP include their own headers for some common functions
+      // implemented here. We'll need to clean those up so they do not conflict.
+      SmallString<128> P(D.ResourceDir);
+      llvm::sys::path::append(P, "include");
+      llvm::sys::path::append(P, "llvm_libc_wrappers");
+      CmdArgs.push_back("-internal-isystem");
+      CmdArgs.push_back(Args.MakeArgString(P));
+    }
+  }
+
   // Add system include arguments for all targets but IAMCU.
   if (!IsIAMCU)
     forAllAssociatedToolChains(C, JA, getToolChain(),

clang/test/Driver/gpu-libc-headers.c

@@ -4,15 +4,15 @@
 // RUN:   %clang -### --target=x86_64-unknown-linux-gnu -fopenmp=libomp --sysroot=./ \
 // RUN:     -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target=nvptx64-nvidia-cuda --offload-arch=sm_70  \
 // RUN:     -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-HEADERS
-// CHECK-HEADERS: "-cc1"{{.*}}"-internal-isystem" "{{.*}}include{{.*}}llvm_libc_wrappers"{{.*}}"-isysroot" "./"
-// CHECK-HEADERS: "-cc1"{{.*}}"-internal-isystem" "{{.*}}include{{.*}}llvm_libc_wrappers"{{.*}}"-isysroot" "./"
+// CHECK-HEADERS: "-cc1"{{.*}}"-isysroot" "./"{{.*}}"-internal-isystem" "{{.*}}include{{.*}}llvm_libc_wrappers"
+// CHECK-HEADERS: "-cc1"{{.*}}"-isysroot" "./"{{.*}}"-internal-isystem" "{{.*}}include{{.*}}llvm_libc_wrappers"
 
 // RUN:   %clang -### --target=amdgcn-amd-amdhsa -mcpu=gfx90a --sysroot=./ \
 // RUN:     -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-HEADERS-AMDGPU
 // RUN:   %clang -### --target=nvptx64-nvidia-cuda -march=sm_89 --sysroot=./ \
 // RUN:     -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-HEADERS-NVPTX
-// CHECK-HEADERS-AMDGPU: "-cc1"{{.*}}"-internal-isystem" "{{.*}}include{{.*}}amdgcn-amd-amdhsa"{{.*}}"-isysroot" "./"
-// CHECK-HEADERS-NVPTX: "-cc1"{{.*}}"-internal-isystem" "{{.*}}include{{.*}}nvptx64-nvidia-cuda"{{.*}}"-isysroot" "./"
+// CHECK-HEADERS-AMDGPU: "-cc1"{{.*}}"-isysroot" "./"{{.*}}"-internal-isystem" "{{.*}}include{{.*}}amdgcn-amd-amdhsa"
+// CHECK-HEADERS-NVPTX: "-cc1"{{.*}}"-isysroot" "./"{{.*}}"-internal-isystem" "{{.*}}include{{.*}}nvptx64-nvidia-cuda"
 
 // RUN:   %clang -### --target=amdgcn-amd-amdhsa -mcpu=gfx1030 -nogpulib \
 // RUN:     -nogpuinc %s 2>&1 | FileCheck %s --check-prefix=CHECK-HEADERS-DISABLED