[OPENMP50]Codegen for lastprivate conditional list items.

Added codegen support for lastprivate conditional. According to the
standard, if  when the conditional modifier appears on the clause, if an
assignment to a list item is encountered in the construct then the
original list item is assigned the value that is assigned to the new
list item in the sequentially last iteration or lexically last section
in which such an assignment is encountered.
We look for the assignment operations and check if the left side
references lastprivate conditional variable. Then the next code is
emitted:
if (last_iv_a <= iv) {
  last_iv_a = iv;
  last_a = lp_a;
}

At the end the implicit barrier is generated to wait for the end of all
threads and then in the check for the last iteration the private copy is
assigned the last value.

if (last_iter) {
  lp_a = last_a; // <--- new code
  a = lp_a;      // <--- store of private value to the original  variable.
}

2 files changed, 86 insertions, 15 deletions

diff --git a/clang/test/OpenMP/for_lastprivate_codegen.cpp b/clang/test/OpenMP/for_lastprivate_codegen.cpp
index b7c82c2f303..ecaf81ae469 100644
--- a/clang/test/OpenMP/for_lastprivate_codegen.cpp
+++ b/clang/test/OpenMP/for_lastprivate_codegen.cpp
@@ -3,17 +3,33 @@
 // RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-apple-darwin10 -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
 // RUN: %clang_cc1 -verify -fopenmp -x c++ -std=c++11 -DLAMBDA -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck -check-prefix=LAMBDA %s
 // RUN: %clang_cc1 -verify -fopenmp -x c++ -fblocks -DBLOCKS -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck -check-prefix=BLOCKS %s
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=50 -DOMP5 -x c++ -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck %s -check-prefix=CHECK -check-prefix=OMP50
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=50 -DOMP5 -x c++ -std=c++11 -triple x86_64-apple-darwin10 -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp -fopenmp-version=50 -DOMP5 -x c++ -triple x86_64-apple-darwin10 -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s -check-prefix=CHECK -check-prefix=OMP50
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=50 -DOMP5 -x c++ -std=c++11 -DLAMBDA -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck -check-prefix=LAMBDA %s
+// RUN: %clang_cc1 -verify -fopenmp -fopenmp-version=50 -DOMP5 -x c++ -fblocks -DBLOCKS -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck -check-prefix=BLOCKS %s
 
 // RUN: %clang_cc1 -verify -fopenmp-simd -x c++ -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck --check-prefix SIMD-ONLY0 %s
 // RUN: %clang_cc1 -fopenmp-simd -x c++ -std=c++11 -triple x86_64-apple-darwin10 -emit-pch -o %t %s
 // RUN: %clang_cc1 -fopenmp-simd -x c++ -triple x86_64-apple-darwin10 -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck --check-prefix SIMD-ONLY0 %s
 // RUN: %clang_cc1 -verify -fopenmp-simd -x c++ -std=c++11 -DLAMBDA -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck --check-prefix SIMD-ONLY0 %s
 // RUN: %clang_cc1 -verify -fopenmp-simd -x c++ -fblocks -DBLOCKS -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck --check-prefix SIMD-ONLY0 %s
+// RUN: %clang_cc1 -verify -fopenmp-simd -fopenmp-version=50 -DOMP5 -x c++ -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck --check-prefix SIMD-ONLY0 %s
+// RUN: %clang_cc1 -fopenmp-simd -fopenmp-version=50 -DOMP5 -x c++ -std=c++11 -triple x86_64-apple-darwin10 -emit-pch -o %t %s
+// RUN: %clang_cc1 -fopenmp-simd -fopenmp-version=50 -DOMP5 -x c++ -triple x86_64-apple-darwin10 -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck --check-prefix SIMD-ONLY0 %s
+// RUN: %clang_cc1 -verify -fopenmp-simd -fopenmp-version=50 -DOMP5 -x c++ -std=c++11 -DLAMBDA -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck --check-prefix SIMD-ONLY0 %s
+// RUN: %clang_cc1 -verify -fopenmp-simd -fopenmp-version=50 -DOMP5 -x c++ -fblocks -DBLOCKS -triple x86_64-apple-darwin10 -emit-llvm %s -o - | FileCheck --check-prefix SIMD-ONLY0 %s
 // SIMD-ONLY0-NOT: {{__kmpc|__tgt}}
 // expected-no-diagnostics
 #ifndef HEADER
 #define HEADER
 
+#ifdef OMP5
+#define CONDITIONAL conditional :
+#else
+#define CONDITIONAL
+#endif //OMP5
+
 enum omp_allocator_handle_t {
   omp_null_allocator = 0,
   omp_default_mem_alloc = 1,
@@ -160,6 +176,10 @@ char cnt;
 // CHECK-DAG: [[X:@.+]] = global double 0.0
 // CHECK-DAG: [[F:@.+]] = global float 0.0
 // CHECK-DAG: [[CNT:@.+]] = global i8 0
+// OMP50-DAG: [[IV_REF:@.+]] = {{.*}}common global i32* null
+// OMP50-DAG: [[LAST_IV_F:@.+]] = {{.*}}common global i32 0
+// OMP50-DAG: [[LAST_F:@.+]] = {{.*}}common global float 0.000000e+00,
+
 template <typename T>
 T tmain() {
   S<T> test;
@@ -477,9 +497,10 @@ int main() {
     A::x++;
   }
 #pragma omp parallel
-#pragma omp for allocate(omp_const_mem_alloc :cnt) lastprivate(cnt)
+#pragma omp for allocate(omp_const_mem_alloc :cnt) lastprivate(cnt) lastprivate(CONDITIONAL f)
   for (cnt = 0; cnt < 2; ++cnt) {
     A::x++;
+    f = 0;
   }
   return tmain<int>();
 #endif
@@ -642,6 +663,9 @@ int main() {
 
 // CHECK: [[GTID_REF:%.+]] = load i{{[0-9]+}}*, i{{[0-9]+}}** [[GTID_ADDR_REF]]
 // CHECK: [[GTID:%.+]] = load i{{[0-9]+}}, i{{[0-9]+}}* [[GTID_REF]]
+// OMP50: [[LOCAL_IV_REF:%.+]] = call i8* @__kmpc_threadprivate_cached(%struct.ident_t* @{{.+}}, i32 [[GTID]], i8* bitcast (i32** [[IV_REF]] to i8*), i64 8, i8*** @{{.+}})
+// OMP50: [[BC:%.+]] = bitcast i8* [[LOCAL_IV_REF]] to i32**
+// OMP50: store i32* %{{.+}}, i32** [[BC]],
 // CHECK: [[CNT_PRIV:%.+]] = call i8* @__kmpc_alloc(i32 [[GTID]], i64 1, i8* inttoptr (i64 3 to i8*))
 // CHECK: call {{.+}} @__kmpc_for_static_init_4(%{{.+}}* @{{.+}}, i32 [[GTID]], i32 34, i32* [[IS_LAST_ADDR:%.+]], i32* [[OMP_LB:%[^,]+]], i32* [[OMP_UB:%[^,]+]], i32* [[OMP_ST:%[^,]+]], i32 1, i32 1)
 // UB = min(UB, GlobalUB)
@@ -653,11 +677,30 @@ int main() {
 // CHECK-NEXT: [[LB:%.+]] = load i32, i32* [[OMP_LB]]
 // CHECK-NEXT: store i32 [[LB]], i32* [[OMP_IV:[^,]+]]
 // <Skip loop body>
+// CHECK: store float 0.000000e+00, float* [[F_PRIV:%.+]],
+// OMP50: [[LOCAL_IV_REF:%.+]] = call i8* @__kmpc_threadprivate_cached(%struct.ident_t* @{{.+}}, i32 [[GTID]], i8* bitcast (i32** [[IV_REF]] to i8*), i64 8, i8*** @{{.+}})
+// OMP50: [[BC:%.+]] = bitcast i8* [[LOCAL_IV_REF]] to i32**
+// OMP50: [[IV_ADDR:%.+]] = load i32*, i32** [[BC]],
+// OMP50: [[IV:%.+]] = load i32, i32* [[IV_ADDR]],
+// OMP50: call void @__kmpc_critical(%struct.ident_t* @{{.+}}, i32 [[GTID]], [8 x i32]* [[F_REGION:@.+]])
+// OMP50: [[LAST_IV:%.+]] = load i32, i32* [[LAST_IV_F]],
+// OMP50: [[CMP:%.+]] = icmp sle i32 [[LAST_IV]], [[IV]]
+// OMP50: br i1 [[CMP]], label %[[LP_THEN:.+]], label %[[LP_DONE:[^,]+]]
+
+// OMP50: [[LP_THEN]]:
+// OMP50: store i32 [[IV]], i32* [[LAST_IV_F]],
+// OMP50: [[F_VAL:%.+]] = load float, float* [[F_PRIV]],
+// OMP50: store float [[F_VAL]], float* [[LAST_F]],
+// OMP50: br label %[[LP_DONE]]
+
+// OMP50: [[LP_DONE]]:
+// OMP50: call void @__kmpc_end_critical(%struct.ident_t* @{{.+}}, i32 [[GTID]], [8 x i32]* [[F_REGION]])
 // CHECK: call void @__kmpc_for_static_fini(%{{.+}}* @{{.+}}, i32 [[GTID]])
 
 // Check for final copying of private values back to original vars.
 // CHECK: [[IS_LAST_VAL:%.+]] = load i32, i32* [[IS_LAST_ADDR]],
 // CHECK: [[IS_LAST_ITER:%.+]] = icmp ne i32 [[IS_LAST_VAL]], 0
+// OMP50-NEXT: call void @__kmpc_barrier(%{{.+}}* [[IMPLICIT_BARRIER_LOC]], i{{[0-9]+}} [[GTID]])
 // CHECK: br i1 [[IS_LAST_ITER:%.+]], label %[[LAST_THEN:.+]], label %[[LAST_DONE:.+]]
 // CHECK: [[LAST_THEN]]
 
@@ -666,6 +709,10 @@ int main() {
 // original cnt=private_cnt;
 // CHECK: [[CNT_VAL:%.+]] = load i8, i8* [[CNT_PRIV]],
 // CHECK: store i8 [[CNT_VAL]], i8* [[CNT]],
+// OMP50: [[F_VAL:%.+]] = load float, float* [[LAST_F]],
+// OMP50: store float [[F_VAL]], float* [[F_PRIV]],
+// CHECK: [[F_VAL:%.+]] = load float, float* [[F_PRIV]],
+// CHECK: store float [[F_VAL]], float* [[F]],
 
 // CHECK-NEXT: br label %[[LAST_DONE]]
 // CHECK: [[LAST_DONE]]
diff --git a/clang/test/OpenMP/simd_codegen.cpp b/clang/test/OpenMP/simd_codegen.cpp
index aac76ad127d..cb53bb1aa38 100644
--- a/clang/test/OpenMP/simd_codegen.cpp
+++ b/clang/test/OpenMP/simd_codegen.cpp
@@ -1,15 +1,15 @@
-// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple x86_64-unknown-unknown -emit-llvm %s -fexceptions -fcxx-exceptions -o - | FileCheck %s
-// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -emit-pch -o %t %s
-// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -debug-info-kind=limited -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
-// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -fopenmp -fexceptions -fcxx-exceptions -debug-info-kind=line-tables-only -x c++ -emit-llvm %s -o - | FileCheck %s --check-prefix=TERM_DEBUG
-// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple x86_64-unknown-unknown -emit-llvm %s -fexceptions -fcxx-exceptions -o - -fopenmp-version=50 -DOMP5 | FileCheck %s --check-prefix=CHECK --check-prefix=OMP50
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple x86_64-unknown-unknown -emit-llvm %s -fexceptions -fcxx-exceptions -o - -fopenmp-version=45 | FileCheck %s
+// RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -emit-pch -o %t %s -fopenmp-version=45
+// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -debug-info-kind=limited -std=c++11 -include-pch %t -verify %s -emit-llvm -o - -fopenmp-version=45 | FileCheck %s
+// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -fopenmp -fexceptions -fcxx-exceptions -debug-info-kind=line-tables-only -x c++ -emit-llvm %s -o - -fopenmp-version=45 | FileCheck %s --check-prefix=TERM_DEBUG
+// RUN: %clang_cc1 -verify -fopenmp -x c++ -triple x86_64-unknown-unknown -emit-llvm %s -fexceptions -fcxx-exceptions -o - -fopenmp-version=50 -DOMP5 | FileCheck %s --check-prefix=CHECK --check-prefix=OMP50 --check-prefix=OMP50RT
 // RUN: %clang_cc1 -fopenmp -x c++ -std=c++11 -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -emit-pch -o %t %s -fopenmp-version=50 -DOMP5
-// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -debug-info-kind=limited -std=c++11 -include-pch %t -verify %s -emit-llvm -o - -fopenmp-version=50 -DOMP5 | FileCheck %s --check-prefix=CHECK --check-prefix=OMP50
+// RUN: %clang_cc1 -fopenmp -x c++ -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -debug-info-kind=limited -std=c++11 -include-pch %t -verify %s -emit-llvm -o - -fopenmp-version=50 -DOMP5 | FileCheck %s --check-prefix=CHECK --check-prefix=OMP50 --check-prefix=OMP50RT
 
-// RUN: %clang_cc1 -verify -fopenmp-simd -x c++ -triple x86_64-unknown-unknown -emit-llvm %s -fexceptions -fcxx-exceptions -o - | FileCheck %s
-// RUN: %clang_cc1 -fopenmp-simd -x c++ -std=c++11 -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -emit-pch -o %t %s
-// RUN: %clang_cc1 -fopenmp-simd -x c++ -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -debug-info-kind=limited -std=c++11 -include-pch %t -verify %s -emit-llvm -o - | FileCheck %s
-// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -fopenmp-simd -fexceptions -fcxx-exceptions -debug-info-kind=line-tables-only -x c++ -emit-llvm %s -o - | FileCheck --check-prefix=TERM_DEBUG %s
+// RUN: %clang_cc1 -verify -fopenmp-simd -x c++ -triple x86_64-unknown-unknown -emit-llvm %s -fexceptions -fcxx-exceptions -o - -fopenmp-version=45 | FileCheck %s
+// RUN: %clang_cc1 -fopenmp-simd -x c++ -std=c++11 -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -emit-pch -o %t %s -fopenmp-version=45
+// RUN: %clang_cc1 -fopenmp-simd -x c++ -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -debug-info-kind=limited -std=c++11 -include-pch %t -verify %s -emit-llvm -o - -fopenmp-version=45 | FileCheck %s
+// RUN: %clang_cc1 -verify -triple x86_64-apple-darwin10 -fopenmp-simd -fexceptions -fcxx-exceptions -debug-info-kind=line-tables-only -x c++ -emit-llvm %s -o - -fopenmp-version=45 | FileCheck --check-prefix=TERM_DEBUG %s
 // RUN: %clang_cc1 -verify -fopenmp-simd -x c++ -triple x86_64-unknown-unknown -emit-llvm %s -fexceptions -fcxx-exceptions -o - -fopenmp-version=50 -DOMP5 | FileCheck %s --check-prefix=CHECK --check-prefix=OMP50
 // RUN: %clang_cc1 -fopenmp-simd -x c++ -std=c++11 -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -emit-pch -o %t %s -fopenmp-version=50 -DOMP5
 // RUN: %clang_cc1 -fopenmp-simd -x c++ -triple x86_64-unknown-unknown -fexceptions -fcxx-exceptions -debug-info-kind=limited -std=c++11 -include-pch %t -verify %s -emit-llvm -o - -fopenmp-version=50 -DOMP5 | FileCheck %s --check-prefix=CHECK --check-prefix=OMP50
@@ -17,7 +17,14 @@
  #ifndef HEADER
  #define HEADER
 
+#ifdef OMP5
+#define CONDITIONAL conditional :
+#else
+#define CONDITIONAL
+#endif //OMP5
 // CHECK: [[SS_TY:%.+]] = type { i32 }
+// OMP50-DAG: [[LAST_IV:@.+]] = {{.*}}common global i64 0
+// OMP50-DAG: [[LAST_A:@.+]] = {{.*}}common global i32 0
 
 long long get_val() { return 0; }
 double *g_ptr;
@@ -214,7 +221,7 @@ void simple(float *a, float *b, float *c, float *d) {
   int A;
   // CHECK: store i32 -1, i32* [[A:%.+]],
   A = -1;
-  #pragma omp simd lastprivate(A)
+  #pragma omp simd lastprivate(CONDITIONAL A)
 // CHECK: store i64 0, i64* [[OMP_IV7:%[^,]+]]
 // CHECK: br label %[[SIMD_LOOP7_COND:[^,]+]]
 // CHECK: [[SIMD_LOOP7_COND]]:
@@ -229,15 +236,32 @@ void simple(float *a, float *b, float *c, float *d) {
 // CHECK-NEXT: [[LC_IT_2:%.+]] = add nsw i64 -10, [[LC_IT_1]]
 // CHECK-NEXT: store i64 [[LC_IT_2]], i64* [[LC:%[^,]+]],{{.+}}!llvm.access.group
 // CHECK-NEXT: [[LC_VAL:%.+]] = load i64, i64* [[LC]]{{.+}}!llvm.access.group
-// CHECK-NEXT: [[CONV:%.+]] = trunc i64 [[LC_VAL]] to i32
-// CHECK-NEXT: store i32 [[CONV]], i32* [[A_PRIV:%[^,]+]],{{.+}}!llvm.access.group
-    A = i;
+// CHECK-NEXT: [[A_VAL:%.+]] = load i32, i32* [[A_PRIV:%[^,]+]],{{.+}}!llvm.access.group
+// CHECK-NEXT: [[CAST:%.+]] = sext i32 [[A_VAL]] to i64
+// CHECK-NEXT: [[ADD:%.+]] = add nsw i64 [[CAST]], [[LC_VAL]]
+// CHECK-NEXT: [[CONV:%.+]] = trunc i64 [[ADD]] to i32
+// CHECK-NEXT: store i32 [[CONV]], i32* [[A_PRIV]],{{.+}}!llvm.access.group
+// OMP50-NEXT: [[IV:%.+]] = load i64, i64* [[OMP_IV7]],{{.+}}!llvm.access.group
+// OMP50RT:    call void @__kmpc_critical(%struct.ident_t* {{.+}}, i32 [[GTID:%.+]], [8 x i32]* [[A_REGION:@.+]]),{{.+}}!llvm.access.group
+// OMP50-NEXT: [[LAST_IV_VAL:%.+]] = load i64, i64* [[LAST_IV]],{{.+}}!llvm.access.group
+// OMP50-NEXT: [[CMP:%.+]] = icmp sle i64 [[LAST_IV_VAL]], [[IV]]
+// OMP50-NEXT: br i1 [[CMP]], label %[[LP_THEN:.+]], label %[[LP_DONE:[^,]+]]
+// OMP50:      [[LP_THEN]]:
+// OMP50-NEXT: store i64 [[IV]], i64* [[LAST_IV]],{{.+}}!llvm.access.group
+// OMP50-NEXT: [[A_VAL:%.+]] = load i32, i32* [[A_PRIV]],{{.+}}!llvm.access.group
+// OMP50-NEXT: store i32 [[A_VAL]], i32* [[LAST_A]],{{.+}}!llvm.access.group
+// OMP50-NEXT: br label %[[LP_DONE]]
+// OMP50:      [[LP_DONE]]:
+// OMP50RT-NEXT: call void @__kmpc_end_critical(%struct.ident_t* {{.+}}, i32 [[GTID]], [8 x i32]* [[A_REGION]]),{{.+}}!llvm.access.group
+    A += i;
 // CHECK: [[IV7_2:%.+]] = load i64, i64* [[OMP_IV7]]{{.*}}!llvm.access.group
 // CHECK-NEXT: [[ADD7_2:%.+]] = add nsw i64 [[IV7_2]], 1
 // CHECK-NEXT: store i64 [[ADD7_2]], i64* [[OMP_IV7]]{{.*}}!llvm.access.group
   }
 // CHECK: [[SIMPLE_LOOP7_END]]:
 // CHECK-NEXT: store i64 11, i64*
+// OMP50-NEXT: [[LAST_A_VAL:%.+]] = load i32, i32* [[LAST_A]],
+// OMP50-NEXT: store i32 [[LAST_A_VAL]], i32* [[A_PRIV]],
 // CHECK-NEXT: [[A_PRIV_VAL:%.+]] = load i32, i32* [[A_PRIV]],
 // CHECK-NEXT: store i32 [[A_PRIV_VAL]], i32* [[A]],
   int R;