OpenMP-codegen: Correctly pass function arguments to subfunctions

Before we only checked if certain instructions can be expanded by us. Now we
check any value, including function arguments.

llvm-svn: 246425
This commit is contained in:
Tobias Grosser 2015-08-31 09:05:43 +00:00
parent f10a32014d
commit f93451802a
2 changed files with 66 additions and 11 deletions

View File

@ -184,17 +184,13 @@ struct FindValuesUser {
/// @brief Extract the values and SCEVs needed to generate code for a block.
static int findValuesInBlock(struct FindValuesUser &User, const ScopStmt *Stmt,
const BasicBlock *BB) {
// Check all the operands of instructions in the basic block.
for (const Instruction &Inst : *BB) {
for (Value *SrcVal : Inst.operands()) {
if (Instruction *OpInst = dyn_cast<Instruction>(SrcVal))
if (canSynthesize(OpInst, &User.LI, &User.SE, &User.R)) {
User.SCEVs.insert(
User.SE.getSCEVAtScope(OpInst, User.LI.getLoopFor(BB)));
continue;
}
}
}
for (const Instruction &Inst : *BB)
for (Value *SrcVal : Inst.operands())
if (canSynthesize(SrcVal, &User.LI, &User.SE, &User.R)) {
User.SCEVs.insert(
User.SE.getSCEVAtScope(SrcVal, User.LI.getLoopFor(BB)));
continue;
}
return 0;
}

View File

@ -0,0 +1,59 @@
; RUN: opt %loadPolly -polly-detect-unprofitable -polly-parallel \
; RUN: -polly-parallel-force -polly-codegen -S -verify-dom-info < %s \
; RUN: | FileCheck %s -check-prefix=IR
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"
; IR: @GOMP_parallel_loop_runtime_start
@longLimit = external global [9 x [23 x i32]], align 16
@shortLimit = external global [9 x [14 x i32]], align 16
define void @init_layer3(i32 %down_sample_sblimit) #0 {
entry:
br label %for.cond.463.preheader
for.cond.463.preheader: ; preds = %entry
br label %for.cond.499.preheader
for.cond.533.preheader: ; preds = %for.inc.530
unreachable
for.cond.499.preheader: ; preds = %for.inc.530, %for.cond.463.preheader
%indvars.iv140 = phi i64 [ 0, %for.cond.463.preheader ], [ %indvars.iv.next141, %for.inc.530 ]
%arrayidx483 = getelementptr inbounds [9 x [23 x i32]], [9 x [23 x i32]]* @longLimit, i64 0, i64 %indvars.iv140, i64 0
store i32 undef, i32* %arrayidx483, align 4, !tbaa !1
%arrayidx487 = getelementptr inbounds [9 x [23 x i32]], [9 x [23 x i32]]* @longLimit, i64 0, i64 %indvars.iv140, i64 0
%tmp = load i32, i32* %arrayidx487, align 4, !tbaa !1
%indvars.iv.next135 = add nuw nsw i64 0, 1
br label %for.body.502
for.body.502: ; preds = %for.inc.527, %for.cond.499.preheader
%indvars.iv137 = phi i64 [ 0, %for.cond.499.preheader ], [ %indvars.iv.next138, %for.inc.527 ]
%arrayidx518 = getelementptr inbounds [9 x [14 x i32]], [9 x [14 x i32]]* @shortLimit, i64 0, i64 %indvars.iv140, i64 %indvars.iv137
%tmp1 = load i32, i32* %arrayidx518, align 4, !tbaa !1
%cmp519 = icmp sgt i32 %tmp1, %down_sample_sblimit
br i1 %cmp519, label %if.then.521, label %for.inc.527
if.then.521: ; preds = %for.body.502
br label %for.inc.527
for.inc.527: ; preds = %if.then.521, %for.body.502
%indvars.iv.next138 = add nuw nsw i64 %indvars.iv137, 1
%exitcond139 = icmp ne i64 %indvars.iv.next138, 14
br i1 %exitcond139, label %for.body.502, label %for.inc.530
for.inc.530: ; preds = %for.inc.527
%indvars.iv.next141 = add nuw nsw i64 %indvars.iv140, 1
%exitcond142 = icmp ne i64 %indvars.iv.next141, 9
br i1 %exitcond142, label %for.cond.499.preheader, label %for.cond.533.preheader
}
attributes #0 = { nounwind uwtable "disable-tail-calls"="false" "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "target-cpu"="x86-64" "target-features"="+sse,+sse2" "unsafe-fp-math"="false" "use-soft-float"="false" }
!llvm.ident = !{!0}
!0 = !{!"clang version 3.8.0 (trunk 246359)"}
!1 = !{!2, !2, i64 0}
!2 = !{!"int", !3, i64 0}
!3 = !{!"omnipotent char", !4, i64 0}
!4 = !{!"Simple C/C++ TBAA"}