[llvm] r304272 - NewGVN: Fix PR 33185 by checking whether we need to recursively
Daniel Berlin via llvm-commits
llvm-commits at lists.llvm.org
Tue May 30 18:47:32 PDT 2017
Author: dannyb
Date: Tue May 30 20:47:32 2017
New Revision: 304272
URL: http://llvm.org/viewvc/llvm-project?rev=304272&view=rev
Log:
NewGVN: Fix PR 33185 by checking whether we need to recursively
generate a phi of ops, which we don't currently support.
Added:
llvm/trunk/test/Transforms/NewGVN/pr33185.ll
Modified:
llvm/trunk/lib/Transforms/Scalar/NewGVN.cpp
Modified: llvm/trunk/lib/Transforms/Scalar/NewGVN.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Scalar/NewGVN.cpp?rev=304272&r1=304271&r2=304272&view=diff
==============================================================================
--- llvm/trunk/lib/Transforms/Scalar/NewGVN.cpp (original)
+++ llvm/trunk/lib/Transforms/Scalar/NewGVN.cpp Tue May 30 20:47:32 2017
@@ -613,7 +613,7 @@ private:
return CClass;
}
void initializeCongruenceClasses(Function &F);
- const Expression *makePossiblePhiOfOps(Instruction *, bool,
+ const Expression *makePossiblePhiOfOps(Instruction *,
SmallPtrSetImpl<Value *> &);
void addPhiOfOps(PHINode *Op, BasicBlock *BB, Instruction *ExistingValue);
@@ -1937,7 +1937,8 @@ void NewGVN::touchAndErase(Map &M, const
}
void NewGVN::addAdditionalUsers(Value *To, Value *User) const {
- AdditionalUsers[To].insert(User);
+ if (isa<Instruction>(To))
+ AdditionalUsers[To].insert(User);
}
void NewGVN::markUsersTouched(Value *V) {
@@ -2423,7 +2424,7 @@ static bool okayForPHIOfOps(const Instru
// When we see an instruction that is an op of phis, generate the equivalent phi
// of ops form.
const Expression *
-NewGVN::makePossiblePhiOfOps(Instruction *I, bool HasBackedge,
+NewGVN::makePossiblePhiOfOps(Instruction *I,
SmallPtrSetImpl<Value *> &Visited) {
if (!okayForPHIOfOps(I))
return nullptr;
@@ -2438,24 +2439,6 @@ NewGVN::makePossiblePhiOfOps(Instruction
return nullptr;
unsigned IDFSNum = InstrToDFSNum(I);
- // Pretty much all of the instructions we can convert to phi of ops over a
- // backedge that are adds, are really induction variables, and those are
- // pretty much pointless to convert. This is very coarse-grained for a
- // test, so if we do find some value, we can change it later.
- // But otherwise, what can happen is we convert the induction variable from
- //
- // i = phi (0, tmp)
- // tmp = i + 1
- //
- // to
- // i = phi (0, tmpphi)
- // tmpphi = phi(1, tmpphi+1)
- //
- // Which we don't want to happen. We could just avoid this for all non-cycle
- // free phis, and we made go that route.
- if (HasBackedge && I->getOpcode() == Instruction::Add)
- return nullptr;
-
SmallPtrSet<const Value *, 8> ProcessedPHIs;
// TODO: We don't do phi translation on memory accesses because it's
// complicated. For a load, we'd need to be able to simulate a new memoryuse,
@@ -2470,6 +2453,16 @@ NewGVN::makePossiblePhiOfOps(Instruction
// Convert op of phis to phi of ops
for (auto &Op : I->operands()) {
+ // TODO: We can't handle expressions that must be recursively translated
+ // IE
+ // a = phi (b, c)
+ // f = use a
+ // g = f + phi of something
+ // To properly make a phi of ops for g, we'd have to properly translate and
+ // use the instruction for f. We should add this by splitting out the
+ // instruction creation we do below.
+ if (isa<Instruction>(Op) && PHINodeUses.count(cast<Instruction>(Op)))
+ return nullptr;
if (!isa<PHINode>(Op))
continue;
auto *OpPHI = cast<PHINode>(Op);
@@ -2782,8 +2775,7 @@ void NewGVN::valueNumberInstruction(Inst
// Make a phi of ops if necessary
if (Symbolized && !isa<ConstantExpression>(Symbolized) &&
!isa<VariableExpression>(Symbolized) && PHINodeUses.count(I)) {
- // FIXME: Backedge argument
- auto *PHIE = makePossiblePhiOfOps(I, false, Visited);
+ auto *PHIE = makePossiblePhiOfOps(I, Visited);
if (PHIE)
Symbolized = PHIE;
}
Added: llvm/trunk/test/Transforms/NewGVN/pr33185.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/NewGVN/pr33185.ll?rev=304272&view=auto
==============================================================================
--- llvm/trunk/test/Transforms/NewGVN/pr33185.ll (added)
+++ llvm/trunk/test/Transforms/NewGVN/pr33185.ll Tue May 30 20:47:32 2017
@@ -0,0 +1,59 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
+; RUN: opt -newgvn -S %s | FileCheck %s
+
+ at a = local_unnamed_addr global i32 9, align 4
+ at .str4 = private unnamed_addr constant [6 x i8] c"D:%d\0A\00", align 1
+
+define i32 @main() local_unnamed_addr {
+; CHECK-LABEL: @main(
+; CHECK-NEXT: entry:
+; CHECK-NEXT: [[TMP:%.*]] = load i32, i32* @a, align 4
+; CHECK-NEXT: [[CMP1_I:%.*]] = icmp ne i32 [[TMP]], 0
+; CHECK-NEXT: br label [[FOR_BODY_I:%.*]]
+; CHECK: for.body.i:
+; CHECK-NEXT: [[TMP1:%.*]] = phi i1 [ true, [[ENTRY:%.*]] ], [ false, [[COND_END_I:%.*]] ]
+; CHECK-NEXT: [[F_08_I:%.*]] = phi i32 [ 0, [[ENTRY]] ], [ [[INC_I:%.*]], [[COND_END_I]] ]
+; CHECK-NEXT: [[MUL_I:%.*]] = select i1 [[CMP1_I]], i32 [[F_08_I]], i32 0
+; CHECK-NEXT: br i1 [[TMP1]], label [[COND_END_I]], label [[COND_TRUE_I:%.*]]
+; CHECK: cond.true.i:
+; CHECK-NEXT: [[DIV_I:%.*]] = udiv i32 [[MUL_I]], [[F_08_I]]
+; CHECK-NEXT: br label [[COND_END_I]]
+; CHECK: cond.end.i:
+; CHECK-NEXT: [[COND_I:%.*]] = phi i32 [ [[DIV_I]], [[COND_TRUE_I]] ], [ 0, [[FOR_BODY_I]] ]
+; CHECK-NEXT: [[INC_I]] = add nuw nsw i32 [[F_08_I]], 1
+; CHECK-NEXT: [[EXITCOND_I:%.*]] = icmp eq i32 [[INC_I]], 4
+; CHECK-NEXT: br i1 [[EXITCOND_I]], label [[FN1_EXIT:%.*]], label [[FOR_BODY_I]]
+; CHECK: fn1.exit:
+; CHECK-NEXT: [[CALL4:%.*]] = tail call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([6 x i8], [6 x i8]* @.str4, i64 0, i64 0), i32 [[COND_I]])
+; CHECK-NEXT: ret i32 0
+;
+entry:
+ %tmp = load i32, i32* @a, align 4
+ %cmp1.i = icmp ne i32 %tmp, 0
+ br label %for.body.i
+
+for.body.i:
+ %tmp1 = phi i1 [ true, %entry ], [ false, %cond.end.i ]
+ %f.08.i = phi i32 [ 0, %entry ], [ %inc.i, %cond.end.i ]
+ %mul.i = select i1 %cmp1.i, i32 %f.08.i, i32 0
+ br i1 %tmp1, label %cond.end.i, label %cond.true.i
+
+cond.true.i:
+ ;; Ensure we don't replace this divide with a phi of ops that merges the wrong loop iteration value
+ %div.i = udiv i32 %mul.i, %f.08.i
+ br label %cond.end.i
+
+cond.end.i:
+ %cond.i = phi i32 [ %div.i, %cond.true.i ], [ 0, %for.body.i ]
+ %inc.i = add nuw nsw i32 %f.08.i, 1
+ %exitcond.i = icmp eq i32 %inc.i, 4
+ br i1 %exitcond.i, label %fn1.exit, label %for.body.i
+
+fn1.exit:
+ %cond.i.lcssa = phi i32 [ %cond.i, %cond.end.i ]
+ %call4= tail call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([6 x i8], [6 x i8]* @.str4, i64 0, i64 0), i32 %cond.i.lcssa)
+ ret i32 0
+}
+
+declare i32 @printf(i8* nocapture readonly, ...)
+
More information about the llvm-commits
mailing list