Commit 8d78ac26 authored by Huber, Joseph's avatar Huber, Joseph Committed by Tom Stellard
Browse files

[OpenMP]Fix PR51349: Remove AlwaysInline for if regions.

After D94315 we add the `NoInline` attribute to the outlined function to handle
data environments in the OpenMP if clause. This conflicted with the `AlwaysInline`
attribute added to the outlined function. for better performance in D106799.
The data environments should ideally not require NoInline, but for now this
fixes PR51349.

Reviewed By: mikerice

Differential Revision: https://reviews.llvm.org/D107649

(cherry picked from commit 41a6b50c)
parent d811546f
Loading
Loading
Loading
Loading
+2 −1
Original line number Diff line number Diff line
@@ -2120,11 +2120,12 @@ void CGOpenMPRuntime::emitParallelCall(CodeGenFunction &CGF, SourceLocation Loc,
    OutlinedFnArgs.append(CapturedVars.begin(), CapturedVars.end());
    // Ensure we do not inline the function. This is trivially true for the ones
    // passed to __kmpc_fork_call but the ones calles in serialized regions
    // passed to __kmpc_fork_call but the ones called in serialized regions
    // could be inlined. This is not a perfect but it is closer to the invariant
    // we want, namely, every data environment starts with a new function.
    // TODO: We should pass the if condition to the runtime function and do the
    //       handling there. Much cleaner code.
    OutlinedFn->removeFnAttr(llvm::Attribute::AlwaysInline);
    OutlinedFn->addFnAttr(llvm::Attribute::NoInline);
    RT.emitOutlinedFunctionCall(CGF, Loc, OutlinedFn, OutlinedFnArgs);
+38 −0
Original line number Diff line number Diff line
// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py UTC_ARGS: --function-signature --check-attributes --include-generated-funcs
// RUN: %clang_cc1 -x c++ -O1 -fopenmp-version=45 -disable-llvm-optzns -verify -fopenmp -triple x86_64-unknown-linux -emit-llvm %s -o - | FileCheck %s --check-prefix=CHECK
// expected-no-diagnostics

#ifndef HEADER
#define HEADER

void foo() {
#pragma omp parallel if(0)
  ;
}

#endif
// CHECK: Function Attrs: mustprogress nounwind
// CHECK-LABEL: define {{[^@]+}}@_Z3foov
// CHECK-SAME: () #[[ATTR0:[0-9]+]] {
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[DOTTHREADID_TEMP_:%.*]] = alloca i32, align 4
// CHECK-NEXT:    [[DOTBOUND_ZERO_ADDR:%.*]] = alloca i32, align 4
// CHECK-NEXT:    store i32 0, i32* [[DOTBOUND_ZERO_ADDR]], align 4
// CHECK-NEXT:    [[TMP0:%.*]] = call i32 @__kmpc_global_thread_num(%struct.ident_t* @[[GLOB1:[0-9]+]])
// CHECK-NEXT:    call void @__kmpc_serialized_parallel(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
// CHECK-NEXT:    store i32 [[TMP0]], i32* [[DOTTHREADID_TEMP_]], align 4, !tbaa [[TBAA3:![0-9]+]]
// CHECK-NEXT:    call void @.omp_outlined.(i32* [[DOTTHREADID_TEMP_]], i32* [[DOTBOUND_ZERO_ADDR]]) #[[ATTR2:[0-9]+]]
// CHECK-NEXT:    call void @__kmpc_end_serialized_parallel(%struct.ident_t* @[[GLOB1]], i32 [[TMP0]])
// CHECK-NEXT:    ret void
//
//
// CHECK: Function Attrs: noinline norecurse nounwind
// CHECK-LABEL: define {{[^@]+}}@.omp_outlined.
// CHECK-SAME: (i32* noalias [[DOTGLOBAL_TID_:%.*]], i32* noalias [[DOTBOUND_TID_:%.*]]) #[[ATTR1:[0-9]+]] {
// CHECK-NEXT:  entry:
// CHECK-NEXT:    [[DOTGLOBAL_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK-NEXT:    [[DOTBOUND_TID__ADDR:%.*]] = alloca i32*, align 8
// CHECK-NEXT:    store i32* [[DOTGLOBAL_TID_]], i32** [[DOTGLOBAL_TID__ADDR]], align 8, !tbaa [[TBAA7:![0-9]+]]
// CHECK-NEXT:    store i32* [[DOTBOUND_TID_]], i32** [[DOTBOUND_TID__ADDR]], align 8, !tbaa [[TBAA7]]
// CHECK-NEXT:    ret void
//