[OpenCL][Sema] Fix builtin rewriting

[mantognini]

Summary

This patch ensures built-in functions are rewritten using the proper
parent declaration. Tests are added to ensure the functionality works
also with C++ for OpenCL.

[Anastasia]

I would say in the interest of saving testing time running one of those test in C++ mode should be enough.

clang/lib/Sema/SemaExpr.cpp

5411

I think these builtins were added before we could specify address spaces in tablegen file. So I think we should actually change these builtins to add the address spaces on pointers such that this code wouldn't be called at all. But your fix is still needed for general cases.

If you either write FIXME next to the OpenCL builtins in Builtins.td to add the address spaces or create a bugzilla bug describing this problem I am happy to go ahead with this patch.

[mantognini]

Address comments: reduce testing & add FIXME.

[mantognini]

I added the FIXME and reduced the amount of testing. Let me know if it looks alright.

[Anastasia]

LGTM! Thanks!

clang/include/clang/Basic/Builtins.def

1516

-> FIXME:

[mantognini]

Addressed missing colon in FIXME comment.

[mantognini]

I'll commit this along with D64083 in a few days in case John has any feedback.

clang/include/clang/Basic/Builtins.def

 BUILTIN(__builtin_coro_free, "v*v*", "n")
 
 BUILTIN(__builtin_coro_id, "v*Iiv*v*v*", "n")
 BUILTIN(__builtin_coro_alloc, "b", "n")
 BUILTIN(__builtin_coro_begin, "v*v*", "n")
 BUILTIN(__builtin_coro_end, "bv*Ib", "n")
 BUILTIN(__builtin_coro_suspend, "cIb", "n")
 BUILTIN(__builtin_coro_param, "bv*v*", "n")
+
 // OpenCL v2.0 s6.13.16, s9.17.3.5 - Pipe functions.
 // We need the generic prototype, since the packet type could be anything.
 LANGBUILTIN(read_pipe, "i.", "tn", OCLC20_LANG)
 LANGBUILTIN(write_pipe, "i.", "tn", OCLC20_LANG)
 
 LANGBUILTIN(reserve_read_pipe, "i.", "tn", OCLC20_LANG)
 LANGBUILTIN(reserve_write_pipe, "i.", "tn", OCLC20_LANG)
 

 // Custom builtin check allows to perform special check of passed block arguments.
 LANGBUILTIN(enqueue_kernel, "i.", "tn", OCLC20_LANG)
 LANGBUILTIN(get_kernel_work_group_size, "Ui.", "tn", OCLC20_LANG)
 LANGBUILTIN(get_kernel_preferred_work_group_size_multiple, "Ui.", "tn", OCLC20_LANG)
 LANGBUILTIN(get_kernel_max_sub_group_size_for_ndrange, "Ui.", "tn", OCLC20_LANG)
 LANGBUILTIN(get_kernel_sub_group_count_for_ndrange, "Ui.", "tn", OCLC20_LANG)
 
 // OpenCL v2.0 s6.13.9 - Address space qualifier functions.
+// FIXME: Pointer parameters of OpenCL builtins should have their address space
+// requirement defined.
 LANGBUILTIN(to_global, "v*v*", "tn", OCLC20_LANG)
 LANGBUILTIN(to_local, "v*v*", "tn", OCLC20_LANG)
 LANGBUILTIN(to_private, "v*v*", "tn", OCLC20_LANG)
 
 // OpenCL half load/store builtin
 LANGBUILTIN(__builtin_store_half, "vdh*", "n", ALL_OCLC_LANGUAGES)
 LANGBUILTIN(__builtin_store_halff, "vfh*", "n", ALL_OCLC_LANGUAGES)
 LANGBUILTIN(__builtin_load_half, "dhC*", "nc", ALL_OCLC_LANGUAGES)

clang/lib/Sema/SemaExpr.cpp

 /// as the call.
 ///
 /// \returns nullptr If this builtin is not a candidate for a rewrite i.e.
 ///                  it does not contain any pointer arguments without
 ///                  an address space qualifer.  Otherwise the rewritten
 ///                  FunctionDecl is returned.
 /// TODO: Handle pointer return types.
 static FunctionDecl *rewriteBuiltinFunctionDecl(Sema *Sema, ASTContext &Context,
-                                                const FunctionDecl *FDecl,
+                                                FunctionDecl *FDecl,
                                                 MultiExprArg ArgExprs) {
 
   QualType DeclType = FDecl->getType();
   const FunctionProtoType *FT = dyn_cast<FunctionProtoType>(DeclType);
 
   if (!Context.BuiltinInfo.hasPtrArgsOrResult(FDecl->getBuiltinID()) ||
       !FT || FT->isVariadic() || ArgExprs.size() != FT->getNumParams())
     return nullptr;
static FunctionDecl *rewriteBuiltinFunctionDecl(Sema *Sema, ASTContext &Context,
   }
 
   if (!NeedsNewDecl)
     return nullptr;
 
   FunctionProtoType::ExtProtoInfo EPI;
   QualType OverloadTy = Context.getFunctionType(FT->getReturnType(),
                                                 OverloadParams, EPI);
-  DeclContext *Parent = Context.getTranslationUnitDecl();
+  DeclContext *Parent = FDecl->getParent();
   FunctionDecl *OverloadDecl = FunctionDecl::Create(Context, Parent,
                                                     FDecl->getLocation(),
                                                     FDecl->getLocation(),
                                                     FDecl->getIdentifier(),
                                                     OverloadTy,
                                                     /*TInfo=*/nullptr,
                                                     SC_Extern, false,
                                                     /*hasPrototype=*/true);

clang/test/CodeGenOpenCL/builtins.cl

-// RUN: %clang_cc1 %s -finclude-default-header -cl-std=CL2.0 -O0 -emit-llvm -o - -triple "spir-unknown-unknown" | FileCheck %s
+// RUN: %clang_cc1 %s -finclude-default-header -cl-std=c++ -fblocks -O0 -emit-llvm -o - -triple "spir-unknown-unknown" | FileCheck %s
 
 void testBranchingOnEnqueueKernel(queue_t default_queue, unsigned flags, ndrange_t ndrange) {
     // Ensure `enqueue_kernel` can be branched upon.
 
     if (enqueue_kernel(default_queue, flags, ndrange, ^(void) {}))
         (void)0;
     // CHECK: [[P:%[0-9]+]] = call i32 @__enqueue_kernel
     // CHECK-NEXT: [[Q:%[a-z0-9]+]] = icmp ne i32 [[P]], 0

clang/test/CodeGenOpenCL/pipe_builtin.cl

-// RUN: %clang_cc1 -emit-llvm -cl-ext=+cl_khr_subgroups -O0 -cl-std=CL2.0 -o - %s | FileCheck %s
+// RUN: %clang_cc1 -emit-llvm -cl-ext=+cl_khr_subgroups -O0 -cl-std=c++ -o - %s | FileCheck %s
 
 // CHECK-DAG: %opencl.pipe_ro_t = type opaque
 // CHECK-DAG: %opencl.pipe_wo_t = type opaque
 // CHECK-DAG: %opencl.reserve_id_t = type opaque
 
 #pragma OPENCL EXTENSION cl_khr_subgroups : enable
 
 void test1(read_only pipe int p, global int *ptr) {

clang/test/CodeGenOpenCL/to_addr_builtin.cl

-// RUN: %clang_cc1 -triple spir-unknown-unknown -emit-llvm -O0 -cl-std=CL2.0 -o - %s | FileCheck %s
+// RUN: %clang_cc1 -triple spir-unknown-unknown -emit-llvm -O0 -cl-std=c++ -o - %s | FileCheck %s
 
 // CHECK: %[[A:.*]] = type { float, float, float }
 typedef struct {
   float x,y,z;
 } A;
 typedef private A *PA;
 typedef global A *GA;