From 0fd3c68c1a753f947fd91d5aad3a366234fb6f3d Mon Sep 17 00:00:00 2001
From: Alexey Bataev <a.bataev@hotmail.com>
Date: Tue, 2 Apr 2019 19:44:46 +0000
Subject: [PATCH] [OPENMP]Add codegen for private vars with allocate clause.

Added codegen/test for the privatized variables with the allocate
clause.

llvm-svn: 357514
---
 clang/lib/CodeGen/CGOpenMPRuntime.cpp          |  8 ++------
 clang/test/OpenMP/parallel_private_codegen.cpp | 20 ++++++++++++++++++--
 2 files changed, 20 insertions(+), 8 deletions(-)

diff --git a/clang/lib/CodeGen/CGOpenMPRuntime.cpp b/clang/lib/CodeGen/CGOpenMPRuntime.cpp
index 8e23dcd..65c5c1c 100644
--- a/clang/lib/CodeGen/CGOpenMPRuntime.cpp
+++ b/clang/lib/CodeGen/CGOpenMPRuntime.cpp
@@ -9761,13 +9761,9 @@ Address CGOpenMPRuntime::getAddressOfLocalVariable(CodeGenFunction &CGF,
     return Address::invalid();
   const auto *AA = CVD->getAttr<OMPAllocateDeclAttr>();
   // Use the default allocation.
-  if (AA->getAllocatorType() == OMPAllocateDeclAttr::OMPDefaultMemAlloc)
+  if (AA->getAllocatorType() == OMPAllocateDeclAttr::OMPDefaultMemAlloc &&
+      !AA->getAllocator())
     return Address::invalid();
-  auto &Elem = OpenMPLocThreadIDMap.FindAndConstruct(CGF.CurFn);
-  if (!Elem.second.ServiceInsertPt)
-    setLocThreadIdInsertPt(CGF);
-  CGBuilderTy::InsertPointGuard IPG(CGF.Builder);
-  CGF.Builder.SetInsertPoint(Elem.second.ServiceInsertPt);
   llvm::Value *Size;
   CharUnits Align = CGM.getContext().getDeclAlign(CVD);
   if (CVD->getType()->isVariablyModifiedType()) {
diff --git a/clang/test/OpenMP/parallel_private_codegen.cpp b/clang/test/OpenMP/parallel_private_codegen.cpp
index 33a5b4c..7271c3f 100644
--- a/clang/test/OpenMP/parallel_private_codegen.cpp
+++ b/clang/test/OpenMP/parallel_private_codegen.cpp
@@ -13,6 +13,16 @@
 // expected-no-diagnostics
 #ifndef HEADER
 #define HEADER
+typedef void **omp_allocator_handle_t;
+extern const omp_allocator_handle_t omp_default_mem_alloc;
+extern const omp_allocator_handle_t omp_large_cap_mem_alloc;
+extern const omp_allocator_handle_t omp_const_mem_alloc;
+extern const omp_allocator_handle_t omp_high_bw_mem_alloc;
+extern const omp_allocator_handle_t omp_low_lat_mem_alloc;
+extern const omp_allocator_handle_t omp_cgroup_mem_alloc;
+extern const omp_allocator_handle_t omp_pteam_mem_alloc;
+extern const omp_allocator_handle_t omp_thread_mem_alloc;
+
 template <class T>
 struct S {
   T f;
@@ -54,7 +64,7 @@ template<typename T>
 struct SST {
   T a;
   SST() : a(T()) {
-#pragma omp parallel private(a)
+#pragma omp parallel private(a) allocate(omp_large_cap_mem_alloc:a)
 #ifdef LAMBDA
     [&]() {
       [&]() {
@@ -343,12 +353,18 @@ int main() {
 // CHECK: ret
 
 // CHECK: define internal void [[SST_MICROTASK]](i{{[0-9]+}}* noalias [[GTID_ADDR:%.+]], i{{[0-9]+}}* noalias %{{.+}}, [[SST_TY]]* %{{.+}})
-// CHECK: [[A_PRIV:%.+]] = alloca i{{[0-9]+}},
+// CHECK: [[GTID_ADDR_PTR:%.+]] = alloca i32*,
+// CHECK: [[GTID_ADDR:%.+]] = load i32*, i32** [[GTID_ADDR_PTR]],
+// CHECK: [[GTID:%.+]] = load i32, i32* [[GTID_ADDR]],
+// CHECK: [[LARGE_CAP_ALLOC:%.+]] = load i8**, i8*** @omp_large_cap_mem_alloc,
+// CHECK: [[A_VOID_PTR:%.+]] = call i8* @__kmpc_alloc(i32 [[GTID]], i64 4, i8** [[LARGE_CAP_ALLOC]])
+// CHECK: [[A_PRIV:%.+]] = bitcast i8* [[A_VOID_PTR]] to i32*
 // CHECK: store i{{[0-9]+}}* [[A_PRIV]], i{{[0-9]+}}** [[REF:%.+]],
 // CHECK-NEXT: [[A_PRIV:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[REF]],
 // CHECK-NEXT: [[A_VAL:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[A_PRIV]],
 // CHECK-NEXT: [[INC:%.+]] = add nsw i{{[0-9]+}} [[A_VAL]], 1
 // CHECK-NEXT: store i{{[0-9]+}} [[INC]], i{{[0-9]+}}* [[A_PRIV]],
+// CHECK-NEXT: call void @__kmpc_free(i32 [[GTID]], i8* [[A_VOID_PTR]], i8** [[LARGE_CAP_ALLOC]])
 // CHECK-NEXT: ret void
 
 #endif
-- 
2.7.4