Files
clang-p2996/llvm/test/Transforms/Attributor/openmp_parallel.ll
Nikita Popov 304f1d59ca [IR] Switch everything to use memory attribute
This switches everything to use the memory attribute proposed in
https://discourse.llvm.org/t/rfc-unify-memory-effect-attributes/65579.
The old argmemonly, inaccessiblememonly and inaccessiblemem_or_argmemonly
attributes are dropped. The readnone, readonly and writeonly attributes
are restricted to parameters only.

The old attributes are auto-upgraded both in bitcode and IR.
The bitcode upgrade is a policy requirement that has to be retained
indefinitely. The IR upgrade is mainly there so it's not necessary
to update all tests using memory attributes in this patch, which
is already large enough. We could drop that part after migrating
tests, or retain it longer term, to make it easier to import IR
from older LLVM versions.

High-level Function/CallBase APIs like doesNotAccessMemory() or
setDoesNotAccessMemory() are mapped transparently to the memory
attribute. Code that directly manipulates attributes (e.g. via
AttributeList) on the other hand needs to switch to working with
the memory attribute instead.

Differential Revision: https://reviews.llvm.org/D135780
2022-11-04 10:21:38 +01:00

274 lines
19 KiB
LLVM

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --function-signature --check-attributes --check-globals
; RUN: opt -aa-pipeline=basic-aa -passes=attributor -attributor-manifest-internal -attributor-max-iterations-verify -attributor-annotate-decl-cs -attributor-max-iterations=3 -S < %s | FileCheck %s --check-prefixes=CHECK,TUNIT
; RUN: opt -aa-pipeline=basic-aa -passes=attributor-cgscc -attributor-manifest-internal -attributor-annotate-decl-cs -S < %s | FileCheck %s --check-prefixes=CHECK,CGSCC
target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
%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 514, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* @0, i32 0, i32 0) }, align 8
@2 = 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
; %a is write only, %b is read only, neither is captured or freed, or ...
; FIXME: %a and %b are *not* readnone!
;.
; CHECK: @[[GLOB0:[0-9]+]] = private unnamed_addr constant [23 x i8] c"
; CHECK: @[[GLOB1:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 514, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* @[[GLOB0]], i32 0, i32 0) }, align 8
; CHECK: @[[GLOB2:[0-9]+]] = private unnamed_addr constant [[STRUCT_IDENT_T:%.*]] { i32 0, i32 2, i32 0, i32 0, i8* getelementptr inbounds ([23 x i8], [23 x i8]* @[[GLOB0]], i32 0, i32 0) }, align 8
;.
define dso_local void @func(float* nocapture %a, float* %b, i32 %N) local_unnamed_addr #0 {
; TUNIT: Function Attrs: nounwind uwtable
; TUNIT-LABEL: define {{[^@]+}}@func
; TUNIT-SAME: (float* nocapture nofree writeonly [[A:%.*]], float* nocapture nofree readonly [[B:%.*]], i32 [[N:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] {
; TUNIT-NEXT: entry:
; TUNIT-NEXT: [[A_ADDR:%.*]] = alloca float*, align 8
; TUNIT-NEXT: [[B_ADDR:%.*]] = alloca float*, align 8
; TUNIT-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4
; TUNIT-NEXT: store float* [[A]], float** [[A_ADDR]], align 8
; TUNIT-NEXT: store float* [[B]], float** [[B_ADDR]], align 8
; TUNIT-NEXT: call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* noundef nonnull align 8 dereferenceable(24) @[[GLOB2]], i32 noundef 3, void (i32*, i32*, ...)* noundef bitcast (void (i32*, i32*, i32*, float**, float**)* @.omp_outlined. to void (i32*, i32*, ...)*), i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) undef, float** noalias nocapture nofree noundef nonnull readonly align 8 dereferenceable(8) [[A_ADDR]], float** noalias nocapture nofree noundef nonnull readonly align 8 dereferenceable(8) [[B_ADDR]])
; TUNIT-NEXT: ret void
;
; CGSCC: Function Attrs: nounwind uwtable
; CGSCC-LABEL: define {{[^@]+}}@func
; CGSCC-SAME: (float* nocapture nofree [[A:%.*]], float* nofree [[B:%.*]], i32 [[N:%.*]]) local_unnamed_addr #[[ATTR0:[0-9]+]] {
; CGSCC-NEXT: entry:
; CGSCC-NEXT: [[A_ADDR:%.*]] = alloca float*, align 8
; CGSCC-NEXT: [[B_ADDR:%.*]] = alloca float*, align 8
; CGSCC-NEXT: [[N_ADDR:%.*]] = alloca i32, align 4
; CGSCC-NEXT: store float* [[A]], float** [[A_ADDR]], align 8
; CGSCC-NEXT: store float* [[B]], float** [[B_ADDR]], align 8
; CGSCC-NEXT: store i32 199, i32* [[N_ADDR]], align 4
; CGSCC-NEXT: call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* noundef nonnull align 8 dereferenceable(24) @[[GLOB2]], i32 noundef 3, void (i32*, i32*, ...)* noundef bitcast (void (i32*, i32*, i32*, float**, float**)* @.omp_outlined. to void (i32*, i32*, ...)*), i32* nofree noundef nonnull readonly align 4 dereferenceable(4) [[N_ADDR]], float** nofree noundef nonnull readonly align 8 dereferenceable(8) [[A_ADDR]], float** nofree noundef nonnull readonly align 8 dereferenceable(8) [[B_ADDR]])
; CGSCC-NEXT: ret void
;
entry:
%a.addr = alloca float*, align 8
%b.addr = alloca float*, align 8
%N.addr = alloca i32, align 4
store float* %a, float** %a.addr, align 8
store float* %b, float** %b.addr, align 8
store i32 199, i32* %N.addr, align 4
call void (%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) @__kmpc_fork_call(%struct.ident_t* nonnull @2, i32 3, void (i32*, i32*, ...)* bitcast (void (i32*, i32*, i32*, float**, float**)* @.omp_outlined. to void (i32*, i32*, ...)*), i32* nonnull %N.addr, float** nonnull %a.addr, float** nonnull %b.addr)
ret void
}
; FIXME: %N should not be loaded but 199 should be used.
define internal void @.omp_outlined.(i32* noalias nocapture readonly %.global_tid., i32* noalias nocapture readnone %.bound_tid., i32* nocapture nonnull readonly align 4 dereferenceable(4) %N, float** nocapture nonnull readonly align 8 dereferenceable(8) %a, float** nocapture nonnull readonly align 8 dereferenceable(8) %b) #1 {
; TUNIT: Function Attrs: alwaysinline nofree norecurse nounwind uwtable
; TUNIT-LABEL: define {{[^@]+}}@.omp_outlined.
; TUNIT-SAME: (i32* noalias nocapture nofree readonly [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]], i32* noalias nocapture nofree nonnull readnone align 4 dereferenceable(4) [[N:%.*]], float** noalias nocapture nofree noundef nonnull readonly align 8 dereferenceable(8) [[A:%.*]], float** noalias nocapture nofree noundef nonnull readonly align 8 dereferenceable(8) [[B:%.*]]) #[[ATTR1:[0-9]+]] {
; TUNIT-NEXT: entry:
; TUNIT-NEXT: [[DOTOMP_LB:%.*]] = alloca i32, align 4
; TUNIT-NEXT: [[DOTOMP_UB:%.*]] = alloca i32, align 4
; TUNIT-NEXT: [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4
; TUNIT-NEXT: [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4
; TUNIT-NEXT: br label [[OMP_PRECOND_THEN:%.*]]
; TUNIT: omp.precond.then:
; TUNIT-NEXT: [[TMP0:%.*]] = bitcast i32* [[DOTOMP_LB]] to i8*
; TUNIT-NEXT: call void @llvm.lifetime.start.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP0]]) #[[ATTR3:[0-9]+]]
; TUNIT-NEXT: store i32 0, i32* [[DOTOMP_LB]], align 4
; TUNIT-NEXT: [[TMP1:%.*]] = bitcast i32* [[DOTOMP_UB]] to i8*
; TUNIT-NEXT: call void @llvm.lifetime.start.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP1]])
; TUNIT-NEXT: store i32 197, i32* [[DOTOMP_UB]], align 4
; TUNIT-NEXT: [[TMP2:%.*]] = bitcast i32* [[DOTOMP_STRIDE]] to i8*
; TUNIT-NEXT: call void @llvm.lifetime.start.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP2]])
; TUNIT-NEXT: store i32 1, i32* [[DOTOMP_STRIDE]], align 4
; TUNIT-NEXT: [[TMP3:%.*]] = bitcast i32* [[DOTOMP_IS_LAST]] to i8*
; TUNIT-NEXT: call void @llvm.lifetime.start.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP3]])
; TUNIT-NEXT: store i32 0, i32* [[DOTOMP_IS_LAST]], align 4
; TUNIT-NEXT: [[TMP4:%.*]] = load i32, i32* [[DOTGLOBAL_TID_]], align 4
; TUNIT-NEXT: call void @__kmpc_for_static_init_4(%struct.ident_t* noundef nonnull align 8 dereferenceable(24) @[[GLOB1]], i32 [[TMP4]], i32 noundef 34, i32* noundef nonnull align 4 dereferenceable(4) [[DOTOMP_IS_LAST]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTOMP_LB]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTOMP_UB]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTOMP_STRIDE]], i32 noundef 1, i32 noundef 1)
; TUNIT-NEXT: [[TMP5:%.*]] = load i32, i32* [[DOTOMP_UB]], align 4
; TUNIT-NEXT: [[CMP4:%.*]] = icmp sgt i32 [[TMP5]], 197
; TUNIT-NEXT: [[COND:%.*]] = select i1 [[CMP4]], i32 197, i32 [[TMP5]]
; TUNIT-NEXT: store i32 [[COND]], i32* [[DOTOMP_UB]], align 4
; TUNIT-NEXT: [[TMP6:%.*]] = load i32, i32* [[DOTOMP_LB]], align 4
; TUNIT-NEXT: [[CMP513:%.*]] = icmp sgt i32 [[TMP6]], [[COND]]
; TUNIT-NEXT: br i1 [[CMP513]], label [[OMP_LOOP_EXIT:%.*]], label [[OMP_INNER_FOR_BODY_LR_PH:%.*]]
; TUNIT: omp.inner.for.body.lr.ph:
; TUNIT-NEXT: [[TMP7:%.*]] = load float*, float** [[B]], align 8
; TUNIT-NEXT: [[TMP8:%.*]] = load float*, float** [[A]], align 8
; TUNIT-NEXT: [[TMP9:%.*]] = sext i32 [[TMP6]] to i64
; TUNIT-NEXT: [[TMP10:%.*]] = sext i32 [[COND]] to i64
; TUNIT-NEXT: br label [[OMP_INNER_FOR_BODY:%.*]]
; TUNIT: omp.inner.for.body:
; TUNIT-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], [[OMP_INNER_FOR_BODY]] ], [ [[TMP9]], [[OMP_INNER_FOR_BODY_LR_PH]] ]
; TUNIT-NEXT: [[INDVARS_IV_NEXT]] = add nsw i64 [[INDVARS_IV]], 1
; TUNIT-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds float, float* [[TMP7]], i64 [[INDVARS_IV_NEXT]]
; TUNIT-NEXT: [[TMP11:%.*]] = load float, float* [[ARRAYIDX]], align 4
; TUNIT-NEXT: [[CONV7:%.*]] = fadd float [[TMP11]], 1.000000e+00
; TUNIT-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds float, float* [[TMP8]], i64 [[INDVARS_IV_NEXT]]
; TUNIT-NEXT: store float [[CONV7]], float* [[ARRAYIDX9]], align 4
; TUNIT-NEXT: [[CMP5:%.*]] = icmp slt i64 [[INDVARS_IV]], [[TMP10]]
; TUNIT-NEXT: br i1 [[CMP5]], label [[OMP_INNER_FOR_BODY]], label [[OMP_LOOP_EXIT]]
; TUNIT: omp.loop.exit:
; TUNIT-NEXT: call void @__kmpc_for_static_fini(%struct.ident_t* noundef nonnull align 8 dereferenceable(24) @[[GLOB1]], i32 [[TMP4]])
; TUNIT-NEXT: call void @llvm.lifetime.end.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP3]])
; TUNIT-NEXT: call void @llvm.lifetime.end.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP2]])
; TUNIT-NEXT: call void @llvm.lifetime.end.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP1]])
; TUNIT-NEXT: call void @llvm.lifetime.end.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP0]])
; TUNIT-NEXT: br label [[OMP_PRECOND_END:%.*]]
; TUNIT: omp.precond.end:
; TUNIT-NEXT: ret void
;
; CGSCC: Function Attrs: alwaysinline nofree norecurse nounwind uwtable
; CGSCC-LABEL: define {{[^@]+}}@.omp_outlined.
; CGSCC-SAME: (i32* noalias nocapture nofree readonly [[DOTGLOBAL_TID_:%.*]], i32* noalias nocapture nofree readnone [[DOTBOUND_TID_:%.*]], i32* noalias nocapture nofree noundef nonnull readonly align 4 dereferenceable(4) [[N:%.*]], float** noalias nocapture nofree noundef nonnull readonly align 8 dereferenceable(8) [[A:%.*]], float** noalias nocapture nofree noundef nonnull readonly align 8 dereferenceable(8) [[B:%.*]]) #[[ATTR1:[0-9]+]] {
; CGSCC-NEXT: entry:
; CGSCC-NEXT: [[DOTOMP_LB:%.*]] = alloca i32, align 4
; CGSCC-NEXT: [[DOTOMP_UB:%.*]] = alloca i32, align 4
; CGSCC-NEXT: [[DOTOMP_STRIDE:%.*]] = alloca i32, align 4
; CGSCC-NEXT: [[DOTOMP_IS_LAST:%.*]] = alloca i32, align 4
; CGSCC-NEXT: [[TMP0:%.*]] = load i32, i32* [[N]], align 4
; CGSCC-NEXT: [[SUB2:%.*]] = add nsw i32 [[TMP0]], -2
; CGSCC-NEXT: [[CMP:%.*]] = icmp sgt i32 [[TMP0]], 1
; CGSCC-NEXT: br i1 [[CMP]], label [[OMP_PRECOND_THEN:%.*]], label [[OMP_PRECOND_END:%.*]]
; CGSCC: omp.precond.then:
; CGSCC-NEXT: [[TMP1:%.*]] = bitcast i32* [[DOTOMP_LB]] to i8*
; CGSCC-NEXT: call void @llvm.lifetime.start.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP1]]) #[[ATTR3:[0-9]+]]
; CGSCC-NEXT: store i32 0, i32* [[DOTOMP_LB]], align 4
; CGSCC-NEXT: [[TMP2:%.*]] = bitcast i32* [[DOTOMP_UB]] to i8*
; CGSCC-NEXT: call void @llvm.lifetime.start.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP2]])
; CGSCC-NEXT: store i32 [[SUB2]], i32* [[DOTOMP_UB]], align 4
; CGSCC-NEXT: [[TMP3:%.*]] = bitcast i32* [[DOTOMP_STRIDE]] to i8*
; CGSCC-NEXT: call void @llvm.lifetime.start.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP3]])
; CGSCC-NEXT: store i32 1, i32* [[DOTOMP_STRIDE]], align 4
; CGSCC-NEXT: [[TMP4:%.*]] = bitcast i32* [[DOTOMP_IS_LAST]] to i8*
; CGSCC-NEXT: call void @llvm.lifetime.start.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP4]])
; CGSCC-NEXT: store i32 0, i32* [[DOTOMP_IS_LAST]], align 4
; CGSCC-NEXT: [[TMP5:%.*]] = load i32, i32* [[DOTGLOBAL_TID_]], align 4
; CGSCC-NEXT: call void @__kmpc_for_static_init_4(%struct.ident_t* noundef nonnull align 8 dereferenceable(24) @[[GLOB1]], i32 [[TMP5]], i32 noundef 34, i32* noundef nonnull align 4 dereferenceable(4) [[DOTOMP_IS_LAST]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTOMP_LB]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTOMP_UB]], i32* noundef nonnull align 4 dereferenceable(4) [[DOTOMP_STRIDE]], i32 noundef 1, i32 noundef 1)
; CGSCC-NEXT: [[TMP6:%.*]] = load i32, i32* [[DOTOMP_UB]], align 4
; CGSCC-NEXT: [[CMP4:%.*]] = icmp sgt i32 [[TMP6]], [[SUB2]]
; CGSCC-NEXT: [[COND:%.*]] = select i1 [[CMP4]], i32 [[SUB2]], i32 [[TMP6]]
; CGSCC-NEXT: store i32 [[COND]], i32* [[DOTOMP_UB]], align 4
; CGSCC-NEXT: [[TMP7:%.*]] = load i32, i32* [[DOTOMP_LB]], align 4
; CGSCC-NEXT: [[CMP513:%.*]] = icmp sgt i32 [[TMP7]], [[COND]]
; CGSCC-NEXT: br i1 [[CMP513]], label [[OMP_LOOP_EXIT:%.*]], label [[OMP_INNER_FOR_BODY_LR_PH:%.*]]
; CGSCC: omp.inner.for.body.lr.ph:
; CGSCC-NEXT: [[TMP8:%.*]] = load float*, float** [[B]], align 8
; CGSCC-NEXT: [[TMP9:%.*]] = load float*, float** [[A]], align 8
; CGSCC-NEXT: [[TMP10:%.*]] = sext i32 [[TMP7]] to i64
; CGSCC-NEXT: [[TMP11:%.*]] = sext i32 [[COND]] to i64
; CGSCC-NEXT: br label [[OMP_INNER_FOR_BODY:%.*]]
; CGSCC: omp.inner.for.body:
; CGSCC-NEXT: [[INDVARS_IV:%.*]] = phi i64 [ [[INDVARS_IV_NEXT:%.*]], [[OMP_INNER_FOR_BODY]] ], [ [[TMP10]], [[OMP_INNER_FOR_BODY_LR_PH]] ]
; CGSCC-NEXT: [[INDVARS_IV_NEXT]] = add nsw i64 [[INDVARS_IV]], 1
; CGSCC-NEXT: [[ARRAYIDX:%.*]] = getelementptr inbounds float, float* [[TMP8]], i64 [[INDVARS_IV_NEXT]]
; CGSCC-NEXT: [[TMP12:%.*]] = load float, float* [[ARRAYIDX]], align 4
; CGSCC-NEXT: [[CONV7:%.*]] = fadd float [[TMP12]], 1.000000e+00
; CGSCC-NEXT: [[ARRAYIDX9:%.*]] = getelementptr inbounds float, float* [[TMP9]], i64 [[INDVARS_IV_NEXT]]
; CGSCC-NEXT: store float [[CONV7]], float* [[ARRAYIDX9]], align 4
; CGSCC-NEXT: [[CMP5:%.*]] = icmp slt i64 [[INDVARS_IV]], [[TMP11]]
; CGSCC-NEXT: br i1 [[CMP5]], label [[OMP_INNER_FOR_BODY]], label [[OMP_LOOP_EXIT]]
; CGSCC: omp.loop.exit:
; CGSCC-NEXT: call void @__kmpc_for_static_fini(%struct.ident_t* noundef nonnull align 8 dereferenceable(24) @[[GLOB1]], i32 [[TMP5]])
; CGSCC-NEXT: call void @llvm.lifetime.end.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP4]])
; CGSCC-NEXT: call void @llvm.lifetime.end.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP3]])
; CGSCC-NEXT: call void @llvm.lifetime.end.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP2]])
; CGSCC-NEXT: call void @llvm.lifetime.end.p0i8(i64 noundef 4, i8* nocapture nofree noundef nonnull align 4 dereferenceable(4) [[TMP1]])
; CGSCC-NEXT: br label [[OMP_PRECOND_END]]
; CGSCC: omp.precond.end:
; CGSCC-NEXT: ret void
;
entry:
%.omp.lb = alloca i32, align 4
%.omp.ub = alloca i32, align 4
%.omp.stride = alloca i32, align 4
%.omp.is_last = alloca i32, align 4
%0 = load i32, i32* %N, align 4
%sub2 = add nsw i32 %0, -2
%cmp = icmp sgt i32 %0, 1
br i1 %cmp, label %omp.precond.then, label %omp.precond.end
omp.precond.then: ; preds = %entry
%1 = bitcast i32* %.omp.lb to i8*
call void @llvm.lifetime.start.p0i8(i64 4, i8* nonnull %1) #3
store i32 0, i32* %.omp.lb, align 4
%2 = bitcast i32* %.omp.ub to i8*
call void @llvm.lifetime.start.p0i8(i64 4, i8* nonnull %2) #3
store i32 %sub2, i32* %.omp.ub, align 4
%3 = bitcast i32* %.omp.stride to i8*
call void @llvm.lifetime.start.p0i8(i64 4, i8* nonnull %3) #3
store i32 1, i32* %.omp.stride, align 4
%4 = bitcast i32* %.omp.is_last to i8*
call void @llvm.lifetime.start.p0i8(i64 4, i8* nonnull %4) #3
store i32 0, i32* %.omp.is_last, align 4
%5 = load i32, i32* %.global_tid., align 4
call void @__kmpc_for_static_init_4(%struct.ident_t* nonnull @1, i32 %5, i32 34, i32* nonnull %.omp.is_last, i32* nonnull %.omp.lb, i32* nonnull %.omp.ub, i32* nonnull %.omp.stride, i32 1, i32 1) #3
%6 = load i32, i32* %.omp.ub, align 4
%cmp4 = icmp sgt i32 %6, %sub2
%cond = select i1 %cmp4, i32 %sub2, i32 %6
store i32 %cond, i32* %.omp.ub, align 4
%7 = load i32, i32* %.omp.lb, align 4
%cmp513 = icmp sgt i32 %7, %cond
br i1 %cmp513, label %omp.loop.exit, label %omp.inner.for.body.lr.ph
omp.inner.for.body.lr.ph: ; preds = %omp.precond.then
%8 = load float*, float** %b, align 8
%9 = load float*, float** %a, align 8
%10 = sext i32 %7 to i64
%11 = sext i32 %cond to i64
br label %omp.inner.for.body
omp.inner.for.body: ; preds = %omp.inner.for.body, %omp.inner.for.body.lr.ph
%indvars.iv = phi i64 [ %indvars.iv.next, %omp.inner.for.body ], [ %10, %omp.inner.for.body.lr.ph ]
%indvars.iv.next = add nsw i64 %indvars.iv, 1
%arrayidx = getelementptr inbounds float, float* %8, i64 %indvars.iv.next
%12 = load float, float* %arrayidx, align 4
%conv7 = fadd float %12, 1.000000e+00
%arrayidx9 = getelementptr inbounds float, float* %9, i64 %indvars.iv.next
store float %conv7, float* %arrayidx9, align 4
%cmp5 = icmp slt i64 %indvars.iv, %11
br i1 %cmp5, label %omp.inner.for.body, label %omp.loop.exit
omp.loop.exit: ; preds = %omp.inner.for.body, %omp.precond.then
call void @__kmpc_for_static_fini(%struct.ident_t* nonnull @1, i32 %5)
call void @llvm.lifetime.end.p0i8(i64 4, i8* nonnull %4) #3
call void @llvm.lifetime.end.p0i8(i64 4, i8* nonnull %3) #3
call void @llvm.lifetime.end.p0i8(i64 4, i8* nonnull %2) #3
call void @llvm.lifetime.end.p0i8(i64 4, i8* nonnull %1) #3
br label %omp.precond.end
omp.precond.end: ; preds = %omp.loop.exit, %entry
ret void
}
; Function Attrs: argmemonly nounwind willreturn
declare void @llvm.lifetime.start.p0i8(i64 immarg, i8* nocapture) #2
; Function Attrs: argmemonly nounwind willreturn
declare void @llvm.lifetime.end.p0i8(i64 immarg, i8* nocapture) #2
declare dso_local void @__kmpc_for_static_init_4(%struct.ident_t*, i32, i32, i32*, i32*, i32*, i32*, i32, i32) local_unnamed_addr
; Function Attrs: nounwind
declare void @__kmpc_for_static_fini(%struct.ident_t*, i32) local_unnamed_addr #3
; Function Attrs: nounwind
declare !callback !1 void @__kmpc_fork_call(%struct.ident_t*, i32, void (i32*, i32*, ...)*, ...) local_unnamed_addr #3
attributes #0 = { nounwind uwtable }
attributes #1 = { alwaysinline nofree norecurse nounwind uwtable }
attributes #2 = { nounwind }
!llvm.module.flags = !{!0}
!0 = !{i32 7, !"openmp", i32 50}
!1 = !{!2}
!2 = !{i64 2, i64 -1, i64 -1, i1 true}
;.
; CHECK: attributes #[[ATTR0:[0-9]+]] = { nounwind uwtable }
; CHECK: attributes #[[ATTR1:[0-9]+]] = { alwaysinline nofree norecurse nounwind uwtable }
; CHECK: attributes #[[ATTR2:[0-9]+]] = { nocallback nofree nosync nounwind willreturn memory(argmem: readwrite) }
; CHECK: attributes #[[ATTR3:[0-9]+]] = { memory(readwrite) }
;.
; CHECK: [[META0:![0-9]+]] = !{i32 7, !"openmp", i32 50}
; CHECK: [[META1:![0-9]+]] = !{!2}
; CHECK: [[META2:![0-9]+]] = !{i64 2, i64 -1, i64 -1, i1 true}
;.