[OpenMP] Change AAKernelInfo to ignore non-kernels

[jhuber6]

Summary

Currently, AAKernelInfo will fail on an assertion if we attempt to run
it on a kernel without the init / deinit runtime calls. However, this
occurs for global constructors on the device. This will cause OpenMPOpt
to crash whenever global constructors are present. This patch removes
this assertion and just gives up instead.

[jdoerfert]

LG

llvm/lib/Transforms/IPO/OpenMPOpt.cpp

InitRFI.foreachUse(
         Fn);
     DeinitRFI.foreachUse(
         [&](Use &U, Function &) {
           StoreCallBase(U, DeinitRFI, KernelDeinitCB);
           return false;
         },
         Fn);
 
-    assert((KernelInitCB && KernelDeinitCB) &&
-           "Kernel without __kmpc_target_init or __kmpc_target_deinit!");
+    // Ignore kernels without initializers such as global constructors.
+    if (!KernelInitCB || !KernelDeinitCB) {
+      indicateOptimisticFixpoint();
+      return;
+    }
 
     // For kernels we might need to initialize/finalize the IsSPMD state and
     // we need to register a simplification callback so that the Attributor
     // knows the constant arguments to __kmpc_target_init and
     // __kmpc_target_deinit might actually change.
 
     Attributor::SimplifictionCallbackTy StateMachineSimplifyCB =
         [&](const IRPosition &IRP, const AbstractAttribute *AA,

llvm/test/Transforms/OpenMP/global_constructor.ll

+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --function-signature --check-attributes --include-generated-funcs
+; RUN: opt -S -passes=openmp-opt < %s | FileCheck %s
+
+%struct.ident_t = type { i32, i32, i32, i32, i8* }
+
+@0 = private unnamed_addr constant [23 x i8] c";unknown;unknown;0;0;;\00", align 1
+@1 = private unnamed_addr constant %struct.ident_t { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* @0, i32 0, i32 0) }, align 8
+@_ZL6Device = internal global double 0.000000e+00, align 8
+@__omp_offloading_fd02_85283c04_main_l11_exec_mode = weak constant i8 0
+
+define weak void @__omp_offloading_fd02_85283c04_main_l11(double* nonnull align 8 dereferenceable(8) %X) local_unnamed_addr {
+entry:
+  %0 = tail call i32 @__kmpc_target_init(%struct.ident_t* nonnull @1, i1 true, i1 false, i1 false) #0
+  %exec_user_code = icmp eq i32 %0, -1
+  br i1 %exec_user_code, label %user_code.entry, label %common.ret
+
+common.ret:
+  ret void
+
+user_code.entry:
+  %1 = load double, double* @_ZL6Device, align 8, !tbaa !11
+  %2 = tail call i32 @__kmpc_get_hardware_thread_id_in_block() #0
+  %3 = icmp eq i32 %2, 0
+  br i1 %3, label %region.guarded, label %region.barrier
+
+region.guarded:
+  store double %1, double* %X, align 8, !tbaa !11
+  br label %region.barrier
+
+region.barrier:
+  tail call void @__kmpc_barrier_simple_spmd(%struct.ident_t* nonnull @1, i32 %2)
+  tail call void @__kmpc_target_deinit(%struct.ident_t* nonnull @1, i1 true, i1 false) #0
+  br label %common.ret
+}
+
+declare i32 @__kmpc_target_init(%struct.ident_t*, i1, i1, i1) local_unnamed_addr
+
+declare void @__kmpc_target_deinit(%struct.ident_t*, i1, i1) local_unnamed_addr
+
+define internal void @__omp_offloading__fd02_85283c04_Device_l6_ctor() {
+entry:
+  %call.i = tail call double @__nv_log(double noundef 2.000000e+00) #1
+  %call.i2 = tail call double @__nv_log(double noundef 2.000000e+00) #1
+  %div = fdiv double %call.i, %call.i2
+  store double %div, double* @_ZL6Device, align 8, !tbaa !11
+  ret void
+}
+
+declare double @__nv_log(double)
+
+declare i32 @__kmpc_get_hardware_thread_id_in_block()
+
+declare void @__kmpc_barrier_simple_spmd(%struct.ident_t*, i32)
+
+attributes #0 = { nounwind }
+attributes #1 = { convergent nounwind }
+
+!omp_offload.info = !{!0, !1, !2}
+!nvvm.annotations = !{!3, !4}
+!llvm.module.flags = !{!5, !6, !7, !8, !9}
+!llvm.ident = !{!10}
+
+!0 = !{i32 0, i32 64770, i32 -2060960764, !"__omp_offloading__fd02_85283c04_Device_l6_ctor", i32 6, i32 1}
+!1 = !{i32 0, i32 64770, i32 -2060960764, !"main", i32 11, i32 2}
+!2 = !{i32 1, !"_ZL6Device", i32 0, i32 0}
+!3 = !{void ()* @__omp_offloading__fd02_85283c04_Device_l6_ctor, !"kernel", i32 1}
+!4 = !{void (double*)* @__omp_offloading_fd02_85283c04_main_l11, !"kernel", i32 1}
+!5 = !{i32 1, !"wchar_size", i32 4}
+!6 = !{i32 7, !"openmp", i32 50}
+!7 = !{i32 7, !"openmp-device", i32 50}
+!8 = !{i32 7, !"PIC Level", i32 2}
+!9 = !{i32 7, !"frame-pointer", i32 2}
+!10 = !{!"clang version 14.0.0"}
+!11 = !{!12, !12, i64 0}
+!12 = !{!"double", !13, i64 0}
+!13 = !{!"omnipotent char", !14, i64 0}
+!14 = !{!"Simple C++ TBAA"}
+; CHECK-LABEL: define {{[^@]+}}@__omp_offloading_fd02_85283c04_main_l11
+; CHECK-SAME: (double* nonnull align 8 dereferenceable(8) [[X:%.*]]) local_unnamed_addr {
+; CHECK-NEXT:  entry:
+; CHECK-NEXT:    [[TMP0:%.*]] = tail call i32 @__kmpc_target_init(%struct.ident_t* nonnull @[[GLOB1:[0-9]+]], i1 true, i1 false, i1 false) #[[ATTR1:[0-9]+]]
+; CHECK-NEXT:    [[EXEC_USER_CODE:%.*]] = icmp eq i32 [[TMP0]], -1
+; CHECK-NEXT:    br i1 [[EXEC_USER_CODE]], label [[USER_CODE_ENTRY:%.*]], label [[COMMON_RET:%.*]]
+; CHECK:       common.ret:
+; CHECK-NEXT:    ret void
+; CHECK:       user_code.entry:
+; CHECK-NEXT:    [[TMP1:%.*]] = load double, double* @_ZL6Device, align 8, !tbaa [[TBAA11:![0-9]+]]
+; CHECK-NEXT:    [[TMP2:%.*]] = tail call i32 @__kmpc_get_hardware_thread_id_in_block() #[[ATTR1]]
+; CHECK-NEXT:    [[TMP3:%.*]] = icmp eq i32 [[TMP2]], 0
+; CHECK-NEXT:    br i1 [[TMP3]], label [[REGION_GUARDED:%.*]], label [[REGION_BARRIER:%.*]]
+; CHECK:       region.guarded:
+; CHECK-NEXT:    store double [[TMP1]], double* [[X]], align 8, !tbaa [[TBAA11]]
+; CHECK-NEXT:    br label [[REGION_BARRIER]]
+; CHECK:       region.barrier:
+; CHECK-NEXT:    tail call void @__kmpc_barrier_simple_spmd(%struct.ident_t* nonnull @[[GLOB1]], i32 [[TMP2]]) #[[ATTR1]]
+; CHECK-NEXT:    tail call void @__kmpc_target_deinit(%struct.ident_t* nonnull @[[GLOB1]], i1 true, i1 false) #[[ATTR1]]
+; CHECK-NEXT:    br label [[COMMON_RET]]
+;