[DAGCombine] Fix getStoreMergeAndAliasCandidates's AA-enabled chain walking

When AA is being used, non-aliasing stores are canonicalized to use the same
chain, and DAGCombiner::getStoreMergeAndAliasCandidates can take advantage of
this by looking only as users of a store's chain operand. However, user
iteration is not result-number specific, we need to check that the use is as a
chain operand, and not via some other operand. It is certainly possible to have
another potentially-aliasing store, which shares the first's base pointer, and
uses the first's chain's node via some other operand.

Failure to catch this situation caused, at least in the included test case, an
assert later because the relative sequence-number ordering caused later
replacement to create a cycle in the DAG.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@248698 91177308-0d34-0410-b5e6-96231b3b80d8
This commit is contained in:
Hal Finkel 2015-09-28 08:02:14 +00:00
parent 273cdae7e9
commit 76dfa41615
2 changed files with 43 additions and 0 deletions

View File

@ -10986,6 +10986,8 @@ void DAGCombiner::getStoreMergeAndAliasCandidates(
SDValue Chain = St->getChain();
for (auto I = Chain->use_begin(), E = Chain->use_end(); I != E; ++I) {
if (StoreSDNode *OtherST = dyn_cast<StoreSDNode>(*I)) {
if (I.getOperandNo() != 0)
continue;
if (OtherST->isVolatile() || OtherST->isIndexed())
continue;

View File

@ -0,0 +1,41 @@
; RUN: llc < %s | FileCheck %s
target datalayout = "E-m:e-i64:64-n32:64"
target triple = "powerpc64-unknown-linux-gnu"
@_ZNSs4_Rep20_S_empty_rep_storageE = external global [0 x i64], align 8
; Function Attrs: nounwind
define void @_ZN5clang7tooling15RefactoringTool10runAndSaveEPNS0_21FrontendActionFactoryE() #0 align 2 {
entry:
br i1 undef, label %_ZN4llvm18IntrusiveRefCntPtrIN5clang13DiagnosticIDsEEC2EPS2_.exit, label %return
; CHECK: @_ZN5clang7tooling15RefactoringTool10runAndSaveEPNS0_21FrontendActionFactoryE
_ZN4llvm18IntrusiveRefCntPtrIN5clang13DiagnosticIDsEEC2EPS2_.exit: ; preds = %entry
%call2 = call noalias i8* @_Znwm() #3
%ref_cnt.i.i = bitcast i8* %call2 to i32*
store <2 x i8*> <i8* bitcast (i64* getelementptr inbounds ([0 x i64], [0 x i64]* @_ZNSs4_Rep20_S_empty_rep_storageE, i64 0, i64 3) to i8*), i8* bitcast (i64* getelementptr inbounds ([0 x i64], [0 x i64]* @_ZNSs4_Rep20_S_empty_rep_storageE, i64 0, i64 3) to i8*)>, <2 x i8*>* undef, align 8
%IgnoreWarnings.i = getelementptr inbounds i8, i8* %call2, i64 4
%0 = bitcast i8* %IgnoreWarnings.i to i32*
call void @llvm.memset.p0i8.i64(i8* null, i8 0, i64 48, i32 8, i1 false) #4
store i32 251658240, i32* %0, align 4
store i256 37662610426935100959726589394453639584271499769928088551424, i256* null, align 8
store i32 1, i32* %ref_cnt.i.i, align 4
unreachable
return: ; preds = %entry
ret void
}
; Function Attrs: nobuiltin
declare noalias i8* @_Znwm() #1
; Function Attrs: nounwind argmemonly
declare void @llvm.memset.p0i8.i64(i8* nocapture, i8, i64, i32, i1) #2
attributes #0 = { nounwind "target-cpu"="pwr7" }
attributes #1 = { nobuiltin "target-cpu"="pwr7" }
attributes #2 = { nounwind argmemonly }
attributes #3 = { builtin nounwind }
attributes #4 = { nounwind }