From 5b6af7163d0a76864bed610c4e884aeffa16b302 Mon Sep 17 00:00:00 2001 From: Josh Magee Date: Thu, 19 Dec 2013 03:17:11 +0000 Subject: [PATCH] [stackprotector] Use analysis from the StackProtector pass for stack layout in PEI a nd LocalStackSlot passes. This changes the MachineFrameInfo API to use the new SSPLayoutKind information produced by the StackProtector pass (instead of a boolean flag) and updates a few pass dependencies (to preserve the SSP analysis). The stack layout follows the same approach used prior to this change - i.e., only LargeArray stack objects will be placed near the canary and everything else will be laid out normally. After this change, structures containing large arrays will also be placed near the canary - a case previously missed by the old implementation. Out of tree targets will need to update their usage of MachineFrameInfo::CreateStackObject to remove the MayNeedSP argument. The next patch will implement the rules for sspstrong and sspreq. The end goal is to support ssp-strong stack layout rules. WIP. Differential Revision: http://llvm-reviews.chandlerc.com/D2158 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@197653 91177308-0d34-0410-b5e6-96231b3b80d8 --- include/llvm/CodeGen/MachineFrameInfo.h | 21 +- include/llvm/CodeGen/StackProtector.h | 2 +- lib/CodeGen/LocalStackSlotAllocation.cpp | 60 ++++- lib/CodeGen/MachineFunction.cpp | 13 +- lib/CodeGen/MachineFunctionPass.cpp | 1 + lib/CodeGen/Passes.cpp | 4 +- lib/CodeGen/PrologEpilogInserter.cpp | 46 +++- .../SelectionDAG/FunctionLoweringInfo.cpp | 10 +- .../SelectionDAG/SelectionDAGBuilder.cpp | 2 +- lib/CodeGen/StackProtector.cpp | 1 - lib/Target/Hexagon/HexagonRemoveSZExtArgs.cpp | 1 + lib/Target/NVPTX/NVPTXAllocaHoisting.h | 1 + lib/Target/NVPTX/NVPTXLowerAggrCopies.h | 1 + lib/Target/NVPTX/NVPTXSplitBBatBar.h | 1 + lib/Target/X86/X86FastISel.cpp | 2 + lib/Target/XCore/XCoreISelLowering.cpp | 2 +- test/CodeGen/ARM/ssp-data-layout.ll | 207 +++++++++++++++ test/CodeGen/X86/ssp-data-layout.ll | 238 ++++++++++++++++++ 18 files changed, 559 insertions(+), 54 deletions(-) create mode 100644 test/CodeGen/ARM/ssp-data-layout.ll create mode 100644 test/CodeGen/X86/ssp-data-layout.ll diff --git a/include/llvm/CodeGen/MachineFrameInfo.h b/include/llvm/CodeGen/MachineFrameInfo.h index d5a0c166412..0699863311f 100644 --- a/include/llvm/CodeGen/MachineFrameInfo.h +++ b/include/llvm/CodeGen/MachineFrameInfo.h @@ -101,11 +101,6 @@ class MachineFrameInfo { // cannot alias any other memory objects. bool isSpillSlot; - // MayNeedSP - If true the stack object triggered the creation of the stack - // protector. We should allocate this object right after the stack - // protector. - bool MayNeedSP; - /// Alloca - If this stack object is originated from an Alloca instruction /// this value saves the original IR allocation. Can be NULL. const AllocaInst *Alloca; @@ -115,9 +110,9 @@ class MachineFrameInfo { bool PreAllocated; StackObject(uint64_t Sz, unsigned Al, int64_t SP, bool IM, - bool isSS, bool NSP, const AllocaInst *Val) + bool isSS, const AllocaInst *Val) : SPOffset(SP), Size(Sz), Alignment(Al), isImmutable(IM), - isSpillSlot(isSS), MayNeedSP(NSP), Alloca(Val), PreAllocated(false) {} + isSpillSlot(isSS), Alloca(Val), PreAllocated(false) {} }; const TargetMachine &TM; @@ -406,14 +401,6 @@ public: return Objects[ObjectIdx+NumFixedObjects].Alloca; } - /// NeedsStackProtector - Returns true if the object may need stack - /// protectors. - bool MayNeedStackProtector(int ObjectIdx) const { - assert(unsigned(ObjectIdx+NumFixedObjects) < Objects.size() && - "Invalid Object Idx!"); - return Objects[ObjectIdx+NumFixedObjects].MayNeedSP; - } - /// getObjectOffset - Return the assigned stack offset of the specified object /// from the incoming stack pointer. /// @@ -531,7 +518,7 @@ public: /// a nonnegative identifier to represent it. /// int CreateStackObject(uint64_t Size, unsigned Alignment, bool isSS, - bool MayNeedSP = false, const AllocaInst *Alloca = 0); + const AllocaInst *Alloca = 0); /// CreateSpillStackObject - Create a new statically sized stack object that /// represents a spill slot, returning a nonnegative identifier to represent @@ -551,7 +538,7 @@ public: /// variable sized object is created, whether or not the index returned is /// actually used. /// - int CreateVariableSizedObject(unsigned Alignment); + int CreateVariableSizedObject(unsigned Alignment, const AllocaInst *Alloca); /// getCalleeSavedInfo - Returns a reference to call saved info vector for the /// current function. diff --git a/include/llvm/CodeGen/StackProtector.h b/include/llvm/CodeGen/StackProtector.h index d09a933a663..7815a49bc24 100644 --- a/include/llvm/CodeGen/StackProtector.h +++ b/include/llvm/CodeGen/StackProtector.h @@ -20,11 +20,11 @@ #include "llvm/ADT/SmallPtrSet.h" #include "llvm/ADT/Triple.h" #include "llvm/ADT/ValueMap.h" +#include "llvm/Analysis/Dominators.h" #include "llvm/Pass.h" #include "llvm/Target/TargetLowering.h" namespace llvm { -class DominatorTree; class Function; class Module; class PHINode; diff --git a/lib/CodeGen/LocalStackSlotAllocation.cpp b/lib/CodeGen/LocalStackSlotAllocation.cpp index 80a6da0f93b..f521548b213 100644 --- a/lib/CodeGen/LocalStackSlotAllocation.cpp +++ b/lib/CodeGen/LocalStackSlotAllocation.cpp @@ -17,12 +17,14 @@ #define DEBUG_TYPE "localstackalloc" #include "llvm/CodeGen/Passes.h" #include "llvm/ADT/STLExtras.h" +#include "llvm/ADT/SetVector.h" #include "llvm/ADT/SmallSet.h" #include "llvm/ADT/Statistic.h" #include "llvm/CodeGen/MachineFrameInfo.h" #include "llvm/CodeGen/MachineFunction.h" #include "llvm/CodeGen/MachineFunctionPass.h" #include "llvm/CodeGen/MachineRegisterInfo.h" +#include "llvm/CodeGen/StackProtector.h" #include "llvm/IR/Constants.h" #include "llvm/IR/DerivedTypes.h" #include "llvm/IR/Instructions.h" @@ -60,18 +62,27 @@ namespace { class LocalStackSlotPass: public MachineFunctionPass { SmallVector LocalOffsets; + /// StackObjSet - A set of stack object indexes + typedef SmallSetVector StackObjSet; void AdjustStackOffset(MachineFrameInfo *MFI, int FrameIdx, int64_t &Offset, bool StackGrowsDown, unsigned &MaxAlign); + void AssignProtectedObjSet(const StackObjSet &UnassignedObjs, + SmallSet &ProtectedObjs, + MachineFrameInfo *MFI, bool StackGrowsDown, + int64_t &Offset, unsigned &MaxAlign); void calculateFrameObjectOffsets(MachineFunction &Fn); bool insertFrameReferenceRegisters(MachineFunction &Fn); public: static char ID; // Pass identification, replacement for typeid - explicit LocalStackSlotPass() : MachineFunctionPass(ID) { } + explicit LocalStackSlotPass() : MachineFunctionPass(ID) { + initializeLocalStackSlotPassPass(*PassRegistry::getPassRegistry()); + } bool runOnMachineFunction(MachineFunction &MF); virtual void getAnalysisUsage(AnalysisUsage &AU) const { AU.setPreservesCFG(); + AU.addRequired(); MachineFunctionPass::getAnalysisUsage(AU); } @@ -81,8 +92,12 @@ namespace { char LocalStackSlotPass::ID = 0; char &llvm::LocalStackSlotAllocationID = LocalStackSlotPass::ID; -INITIALIZE_PASS(LocalStackSlotPass, "localstackalloc", - "Local Stack Slot Allocation", false, false) +INITIALIZE_PASS_BEGIN(LocalStackSlotPass, "localstackalloc", + "Local Stack Slot Allocation", false, false) +INITIALIZE_PASS_DEPENDENCY(StackProtector) +INITIALIZE_PASS_END(LocalStackSlotPass, "localstackalloc", + "Local Stack Slot Allocation", false, false) + bool LocalStackSlotPass::runOnMachineFunction(MachineFunction &MF) { MachineFrameInfo *MFI = MF.getFrameInfo(); @@ -145,6 +160,22 @@ void LocalStackSlotPass::AdjustStackOffset(MachineFrameInfo *MFI, ++NumAllocations; } +/// AssignProtectedObjSet - Helper function to assign large stack objects (i.e., +/// those required to be close to the Stack Protector) to stack offsets. +void LocalStackSlotPass::AssignProtectedObjSet(const StackObjSet &UnassignedObjs, + SmallSet &ProtectedObjs, + MachineFrameInfo *MFI, + bool StackGrowsDown, int64_t &Offset, + unsigned &MaxAlign) { + + for (StackObjSet::const_iterator I = UnassignedObjs.begin(), + E = UnassignedObjs.end(); I != E; ++I) { + int i = *I; + AdjustStackOffset(MFI, i, Offset, StackGrowsDown, MaxAlign); + ProtectedObjs.insert(i); + } +} + /// calculateFrameObjectOffsets - Calculate actual frame offsets for all of the /// abstract stack objects. /// @@ -156,11 +187,13 @@ void LocalStackSlotPass::calculateFrameObjectOffsets(MachineFunction &Fn) { TFI.getStackGrowthDirection() == TargetFrameLowering::StackGrowsDown; int64_t Offset = 0; unsigned MaxAlign = 0; + StackProtector *SP = &getAnalysis(); // Make sure that the stack protector comes before the local variables on the // stack. - SmallSet LargeStackObjs; + SmallSet ProtectedObjs; if (MFI->getStackProtectorIndex() >= 0) { + StackObjSet LargeArrayObjs; AdjustStackOffset(MFI, MFI->getStackProtectorIndex(), Offset, StackGrowsDown, MaxAlign); @@ -170,12 +203,21 @@ void LocalStackSlotPass::calculateFrameObjectOffsets(MachineFunction &Fn) { continue; if (MFI->getStackProtectorIndex() == (int)i) continue; - if (!MFI->MayNeedStackProtector(i)) - continue; - AdjustStackOffset(MFI, i, Offset, StackGrowsDown, MaxAlign); - LargeStackObjs.insert(i); + switch (SP->getSSPLayout(MFI->getObjectAllocation(i))) { + case StackProtector::SSPLK_None: + case StackProtector::SSPLK_SmallArray: + case StackProtector::SSPLK_AddrOf: + continue; + case StackProtector::SSPLK_LargeArray: + LargeArrayObjs.insert(i); + continue; + } + llvm_unreachable("Unexpected SSPLayoutKind."); } + + AssignProtectedObjSet(LargeArrayObjs, ProtectedObjs, MFI, StackGrowsDown, + Offset, MaxAlign); } // Then assign frame offsets to stack objects that are not used to spill @@ -185,7 +227,7 @@ void LocalStackSlotPass::calculateFrameObjectOffsets(MachineFunction &Fn) { continue; if (MFI->getStackProtectorIndex() == (int)i) continue; - if (LargeStackObjs.count(i)) + if (ProtectedObjs.count(i)) continue; AdjustStackOffset(MFI, i, Offset, StackGrowsDown, MaxAlign); diff --git a/lib/CodeGen/MachineFunction.cpp b/lib/CodeGen/MachineFunction.cpp index 96a5ccb9691..80956bc4f7e 100644 --- a/lib/CodeGen/MachineFunction.cpp +++ b/lib/CodeGen/MachineFunction.cpp @@ -499,14 +499,13 @@ static inline unsigned clampStackAlignment(bool ShouldClamp, unsigned Align, /// a nonnegative identifier to represent it. /// int MachineFrameInfo::CreateStackObject(uint64_t Size, unsigned Alignment, - bool isSS, bool MayNeedSP, const AllocaInst *Alloca) { + bool isSS, const AllocaInst *Alloca) { assert(Size != 0 && "Cannot allocate zero size stack objects!"); Alignment = clampStackAlignment(!getFrameLowering()->isStackRealignable() || !RealignOption, Alignment, getFrameLowering()->getStackAlignment()); - Objects.push_back(StackObject(Size, Alignment, 0, false, isSS, MayNeedSP, - Alloca)); + Objects.push_back(StackObject(Size, Alignment, 0, false, isSS, Alloca)); int Index = (int)Objects.size() - NumFixedObjects - 1; assert(Index >= 0 && "Bad frame index!"); ensureMaxAlignment(Alignment); @@ -523,7 +522,7 @@ int MachineFrameInfo::CreateSpillStackObject(uint64_t Size, clampStackAlignment(!getFrameLowering()->isStackRealignable() || !RealignOption, Alignment, getFrameLowering()->getStackAlignment()); - CreateStackObject(Size, Alignment, true, false); + CreateStackObject(Size, Alignment, true); int Index = (int)Objects.size() - NumFixedObjects - 1; ensureMaxAlignment(Alignment); return Index; @@ -534,13 +533,14 @@ int MachineFrameInfo::CreateSpillStackObject(uint64_t Size, /// variable sized object is created, whether or not the index returned is /// actually used. /// -int MachineFrameInfo::CreateVariableSizedObject(unsigned Alignment) { +int MachineFrameInfo::CreateVariableSizedObject(unsigned Alignment, + const AllocaInst *Alloca) { HasVarSizedObjects = true; Alignment = clampStackAlignment(!getFrameLowering()->isStackRealignable() || !RealignOption, Alignment, getFrameLowering()->getStackAlignment()); - Objects.push_back(StackObject(0, Alignment, 0, false, false, true, 0)); + Objects.push_back(StackObject(0, Alignment, 0, false, false, Alloca)); ensureMaxAlignment(Alignment); return (int)Objects.size()-NumFixedObjects-1; } @@ -565,7 +565,6 @@ int MachineFrameInfo::CreateFixedObject(uint64_t Size, int64_t SPOffset, Align, getFrameLowering()->getStackAlignment()); Objects.insert(Objects.begin(), StackObject(Size, Align, SPOffset, Immutable, /*isSS*/ false, - /*NeedSP*/ false, /*Alloca*/ 0)); return -++NumFixedObjects; } diff --git a/lib/CodeGen/MachineFunctionPass.cpp b/lib/CodeGen/MachineFunctionPass.cpp index 674cc80a006..789f2042a07 100644 --- a/lib/CodeGen/MachineFunctionPass.cpp +++ b/lib/CodeGen/MachineFunctionPass.cpp @@ -51,6 +51,7 @@ void MachineFunctionPass::getAnalysisUsage(AnalysisUsage &AU) const { AU.addPreserved("domfrontier"); AU.addPreserved("loops"); AU.addPreserved("lda"); + AU.addPreserved("stack-protector"); FunctionPass::getAnalysisUsage(AU); } diff --git a/lib/CodeGen/Passes.cpp b/lib/CodeGen/Passes.cpp index c9a55f07886..f31a843a94c 100644 --- a/lib/CodeGen/Passes.cpp +++ b/lib/CodeGen/Passes.cpp @@ -427,10 +427,10 @@ void TargetPassConfig::addCodeGenPrepare() { /// Add common passes that perform LLVM IR to IR transforms in preparation for /// instruction selection. void TargetPassConfig::addISelPrepare() { - addPass(createStackProtectorPass(TM)); - addPreISel(); + addPass(createStackProtectorPass(TM)); + if (PrintISelInput) addPass(createPrintFunctionPass("\n\n" "*** Final LLVM Code input to ISel ***\n", diff --git a/lib/CodeGen/PrologEpilogInserter.cpp b/lib/CodeGen/PrologEpilogInserter.cpp index 9e699799912..6a43efbec7e 100644 --- a/lib/CodeGen/PrologEpilogInserter.cpp +++ b/lib/CodeGen/PrologEpilogInserter.cpp @@ -20,6 +20,7 @@ #include "PrologEpilogInserter.h" #include "llvm/ADT/IndexedMap.h" #include "llvm/ADT/STLExtras.h" +#include "llvm/ADT/SetVector.h" #include "llvm/ADT/SmallSet.h" #include "llvm/ADT/Statistic.h" #include "llvm/CodeGen/MachineDominators.h" @@ -30,6 +31,7 @@ #include "llvm/CodeGen/MachineRegisterInfo.h" #include "llvm/CodeGen/RegisterScavenging.h" #include "llvm/IR/DiagnosticInfo.h" +#include "llvm/CodeGen/StackProtector.h" #include "llvm/IR/InlineAsm.h" #include "llvm/IR/LLVMContext.h" #include "llvm/Support/CommandLine.h" @@ -56,6 +58,7 @@ INITIALIZE_PASS_BEGIN(PEI, "prologepilog", "Prologue/Epilogue Insertion", false, false) INITIALIZE_PASS_DEPENDENCY(MachineLoopInfo) INITIALIZE_PASS_DEPENDENCY(MachineDominatorTree) +INITIALIZE_PASS_DEPENDENCY(StackProtector) INITIALIZE_PASS_DEPENDENCY(TargetPassConfig) INITIALIZE_PASS_END(PEI, "prologepilog", "Prologue/Epilogue Insertion & Frame Finalization", @@ -69,6 +72,7 @@ void PEI::getAnalysisUsage(AnalysisUsage &AU) const { AU.setPreservesCFG(); AU.addPreserved(); AU.addPreserved(); + AU.addRequired(); AU.addRequired(); MachineFunctionPass::getAnalysisUsage(AU); } @@ -97,6 +101,9 @@ void PEI::calculateSets(MachineFunction &Fn) { return; } +/// StackObjSet - A set of stack object indexes +typedef SmallSetVector StackObjSet; + /// runOnMachineFunction - Insert prolog/epilog code and replace abstract /// frame indexes with appropriate references. /// @@ -412,11 +419,28 @@ AdjustStackOffset(MachineFrameInfo *MFI, int FrameIdx, } } +/// AssignProtectedObjSet - Helper function to assign large stack objects (i.e., +/// those required to be close to the Stack Protector) to stack offsets. +static void +AssignProtectedObjSet(const StackObjSet &UnassignedObjs, + SmallSet &ProtectedObjs, + MachineFrameInfo *MFI, bool StackGrowsDown, + int64_t &Offset, unsigned &MaxAlign) { + + for (StackObjSet::const_iterator I = UnassignedObjs.begin(), + E = UnassignedObjs.end(); I != E; ++I) { + int i = *I; + AdjustStackOffset(MFI, i, StackGrowsDown, Offset, MaxAlign); + ProtectedObjs.insert(i); + } +} + /// calculateFrameObjectOffsets - Calculate actual frame offsets for all of the /// abstract stack objects. /// void PEI::calculateFrameObjectOffsets(MachineFunction &Fn) { const TargetFrameLowering &TFI = *Fn.getTarget().getFrameLowering(); + StackProtector *SP = &getAnalysis(); bool StackGrowsDown = TFI.getStackGrowthDirection() == TargetFrameLowering::StackGrowsDown; @@ -526,8 +550,9 @@ void PEI::calculateFrameObjectOffsets(MachineFunction &Fn) { // Make sure that the stack protector comes before the local variables on the // stack. - SmallSet LargeStackObjs; + SmallSet ProtectedObjs; if (MFI->getStackProtectorIndex() >= 0) { + StackObjSet LargeArrayObjs; AdjustStackOffset(MFI, MFI->getStackProtectorIndex(), StackGrowsDown, Offset, MaxAlign); @@ -544,12 +569,21 @@ void PEI::calculateFrameObjectOffsets(MachineFunction &Fn) { continue; if (MFI->getStackProtectorIndex() == (int)i) continue; - if (!MFI->MayNeedStackProtector(i)) - continue; - AdjustStackOffset(MFI, i, StackGrowsDown, Offset, MaxAlign); - LargeStackObjs.insert(i); + switch (SP->getSSPLayout(MFI->getObjectAllocation(i))) { + case StackProtector::SSPLK_None: + case StackProtector::SSPLK_SmallArray: + case StackProtector::SSPLK_AddrOf: + continue; + case StackProtector::SSPLK_LargeArray: + LargeArrayObjs.insert(i); + continue; + } + llvm_unreachable("Unexpected SSPLayoutKind."); } + + AssignProtectedObjSet(LargeArrayObjs, ProtectedObjs, MFI, StackGrowsDown, + Offset, MaxAlign); } // Then assign frame offsets to stack objects that are not used to spill @@ -566,7 +600,7 @@ void PEI::calculateFrameObjectOffsets(MachineFunction &Fn) { continue; if (MFI->getStackProtectorIndex() == (int)i) continue; - if (LargeStackObjs.count(i)) + if (ProtectedObjs.count(i)) continue; AdjustStackOffset(MFI, i, StackGrowsDown, Offset, MaxAlign); diff --git a/lib/CodeGen/SelectionDAG/FunctionLoweringInfo.cpp b/lib/CodeGen/SelectionDAG/FunctionLoweringInfo.cpp index 4309dc1d48c..82e97f40fc0 100644 --- a/lib/CodeGen/SelectionDAG/FunctionLoweringInfo.cpp +++ b/lib/CodeGen/SelectionDAG/FunctionLoweringInfo.cpp @@ -85,16 +85,8 @@ void FunctionLoweringInfo::set(const Function &fn, MachineFunction &mf) { TySize *= CUI->getZExtValue(); // Get total allocated size. if (TySize == 0) TySize = 1; // Don't create zero-sized stack objects. - // The object may need to be placed onto the stack near the stack - // protector if one exists. Determine here if this object is a suitable - // candidate. I.e., it would trigger the creation of a stack protector. - bool MayNeedSP = - (AI->isArrayAllocation() || - (TySize >= 8 && isa(Ty) && - cast(Ty)->getElementType()->isIntegerTy(8))); StaticAllocaMap[AI] = - MF->getFrameInfo()->CreateStackObject(TySize, Align, false, - MayNeedSP, AI); + MF->getFrameInfo()->CreateStackObject(TySize, Align, false, AI); } for (; BB != EB; ++BB) diff --git a/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp b/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp index e58856f024f..e40a7c44ade 100644 --- a/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp +++ b/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp @@ -3380,7 +3380,7 @@ void SelectionDAGBuilder::visitAlloca(const AllocaInst &I) { // Inform the Frame Information that we have just allocated a variable-sized // object. - FuncInfo.MF->getFrameInfo()->CreateVariableSizedObject(Align ? Align : 1); + FuncInfo.MF->getFrameInfo()->CreateVariableSizedObject(Align ? Align : 1, &I); } void SelectionDAGBuilder::visitLoad(const LoadInst &I) { diff --git a/lib/CodeGen/StackProtector.cpp b/lib/CodeGen/StackProtector.cpp index 90204499715..19fb0edcc3c 100644 --- a/lib/CodeGen/StackProtector.cpp +++ b/lib/CodeGen/StackProtector.cpp @@ -20,7 +20,6 @@ #include "llvm/CodeGen/Passes.h" #include "llvm/ADT/SmallPtrSet.h" #include "llvm/ADT/Statistic.h" -#include "llvm/Analysis/Dominators.h" #include "llvm/Analysis/ValueTracking.h" #include "llvm/IR/Attributes.h" #include "llvm/IR/Constants.h" diff --git a/lib/Target/Hexagon/HexagonRemoveSZExtArgs.cpp b/lib/Target/Hexagon/HexagonRemoveSZExtArgs.cpp index 44234e85dc6..552cac22eef 100644 --- a/lib/Target/Hexagon/HexagonRemoveSZExtArgs.cpp +++ b/lib/Target/Hexagon/HexagonRemoveSZExtArgs.cpp @@ -42,6 +42,7 @@ namespace { virtual void getAnalysisUsage(AnalysisUsage &AU) const { AU.addRequired(); AU.addPreserved(); + AU.addPreserved("stack-protector"); FunctionPass::getAnalysisUsage(AU); } }; diff --git a/lib/Target/NVPTX/NVPTXAllocaHoisting.h b/lib/Target/NVPTX/NVPTXAllocaHoisting.h index 19d73c5783c..d3305701a5b 100644 --- a/lib/Target/NVPTX/NVPTXAllocaHoisting.h +++ b/lib/Target/NVPTX/NVPTXAllocaHoisting.h @@ -32,6 +32,7 @@ public: void getAnalysisUsage(AnalysisUsage &AU) const { AU.addRequired(); + AU.addPreserved("stack-protector"); AU.addPreserved(); } diff --git a/lib/Target/NVPTX/NVPTXLowerAggrCopies.h b/lib/Target/NVPTX/NVPTXLowerAggrCopies.h index 286e753fa92..780ed4974b2 100644 --- a/lib/Target/NVPTX/NVPTXLowerAggrCopies.h +++ b/lib/Target/NVPTX/NVPTXLowerAggrCopies.h @@ -29,6 +29,7 @@ struct NVPTXLowerAggrCopies : public FunctionPass { void getAnalysisUsage(AnalysisUsage &AU) const { AU.addRequired(); + AU.addPreserved("stack-protector"); AU.addPreserved(); } diff --git a/lib/Target/NVPTX/NVPTXSplitBBatBar.h b/lib/Target/NVPTX/NVPTXSplitBBatBar.h index bdafba9075a..03391feb3b4 100644 --- a/lib/Target/NVPTX/NVPTXSplitBBatBar.h +++ b/lib/Target/NVPTX/NVPTXSplitBBatBar.h @@ -26,6 +26,7 @@ struct NVPTXSplitBBatBar : public FunctionPass { NVPTXSplitBBatBar() : FunctionPass(ID) {} void getAnalysisUsage(AnalysisUsage &AU) const { + AU.addPreserved("stack-protector"); AU.addPreserved(); } virtual bool runOnFunction(Function &F); diff --git a/lib/Target/X86/X86FastISel.cpp b/lib/Target/X86/X86FastISel.cpp index 28a6ac0ab25..972b82fa946 100644 --- a/lib/Target/X86/X86FastISel.cpp +++ b/lib/Target/X86/X86FastISel.cpp @@ -1696,6 +1696,8 @@ bool X86FastISel::X86VisitIntrinsicCall(const IntrinsicInst &I) { const Value *Op1 = I.getArgOperand(0); // The guard's value. const AllocaInst *Slot = cast(I.getArgOperand(1)); + MFI.setStackProtectorIndex(FuncInfo.StaticAllocaMap[Slot]); + // Grab the frame index. X86AddressMode AM; if (!X86SelectAddress(Slot, AM)) return false; diff --git a/lib/Target/XCore/XCoreISelLowering.cpp b/lib/Target/XCore/XCoreISelLowering.cpp index 6b0a2d0835e..668a2486396 100644 --- a/lib/Target/XCore/XCoreISelLowering.cpp +++ b/lib/Target/XCore/XCoreISelLowering.cpp @@ -1246,7 +1246,7 @@ XCoreTargetLowering::LowerCCCArguments(SDValue Chain, unsigned Size = ArgDI->Flags.getByValSize(); unsigned Align = std::max(StackSlotSize, ArgDI->Flags.getByValAlign()); // Create a new object on the stack and copy the pointee into it. - int FI = MFI->CreateStackObject(Size, Align, false, false); + int FI = MFI->CreateStackObject(Size, Align, false); SDValue FIN = DAG.getFrameIndex(FI, MVT::i32); InVals.push_back(FIN); MemOps.push_back(DAG.getMemcpy(Chain, dl, FIN, ArgDI->SDV, diff --git a/test/CodeGen/ARM/ssp-data-layout.ll b/test/CodeGen/ARM/ssp-data-layout.ll new file mode 100644 index 00000000000..afcdfdee1aa --- /dev/null +++ b/test/CodeGen/ARM/ssp-data-layout.ll @@ -0,0 +1,207 @@ +; RUN: llc < %s -disable-fp-elim -march=arm -o - | FileCheck %s +; This test is fairly fragile. The goal is to ensure that "large" stack +; objects are allocated closest to the stack protector (i.e., farthest away +; from the Stack Pointer.) In standard SSP mode this means that large (>= +; ssp-buffer-size) arrays and structures containing such arrays are +; closet to the protector. With sspstrong and sspreq this means large +; arrays/structures-with-arrays are closest, followed by small (< ssp-buffer-size) +; arrays/structures-with-arrays, and then addr-taken variables. +; +; Ideally, we only want verify that the objects appear in the correct groups +; and that the groups have the correct relative stack offset. The ordering +; within a group is not relevant to this test. Unfortunately, there is not +; an elegant way to do this, so just match the offset for each object. + +%struct.struct_large_char = type { [8 x i8] } +%struct.struct_large_char2 = type { [2 x i8], [8 x i8] } +%struct.struct_small_char = type { [2 x i8] } +%struct.struct_large_nonchar = type { [8 x i32] } +%struct.struct_small_nonchar = type { [2 x i16] } + +define void @layout_ssp() ssp { +entry: +; Expected stack layout for ssp is +; 180 large_char . Group 1, nested arrays, arrays >= ssp-buffer-size +; 172 struct_large_char . +; 168 scalar1 | Everything else +; 164 scalar2 +; 160 scalar3 +; 156 addr-of +; 152 small_nonchar (84+68) +; 112 large_nonchar +; 110 small_char +; 108 struct_small_char +; 72 struct_large_nonchar +; 68 struct_small_nonchar + +; CHECK: layout_ssp: +; r[[SP]] is used as an offset into the stack later +; CHECK: add r[[SP:[0-9]+]], sp, #68 + +; CHECK: bl get_scalar1 +; CHECK: str r0, [sp, #168] +; CHECK: bl end_scalar1 + +; CHECK: bl get_scalar2 +; CHECK: str r0, [sp, #164] +; CHECK: bl end_scalar2 + +; CHECK: bl get_scalar3 +; CHECK: str r0, [sp, #160] +; CHECK: bl end_scalar3 + +; CHECK: bl get_addrof +; CHECK: str r0, [sp, #156] +; CHECK: bl end_addrof + +; CHECK: get_small_nonchar +; CHECK: strh r0, [r[[SP]], #84] +; CHECK: bl end_small_nonchar + +; CHECK: bl get_large_nonchar +; CHECK: str r0, [sp, #112] +; CHECK: bl end_large_nonchar + +; CHECK: bl get_small_char +; CHECK: strb r0, [sp, #110] +; CHECK: bl end_small_char + +; CHECK: bl get_large_char +; CHECK: strb r0, [sp, #180] +; CHECK: bl end_large_char + +; CHECK: bl get_struct_large_char +; CHECK: strb r0, [sp, #172] +; CHECK: bl end_struct_large_char + +; CHECK: bl get_struct_small_char +; CHECK: strb r0, [sp, #108] +; CHECK: bl end_struct_small_char + +; CHECK: bl get_struct_large_nonchar +; CHECK:str r0, [sp, #72] +; CHECK: bl end_struct_large_nonchar + +; CHECK: bl get_struct_small_nonchar +; CHECK: strh r0, [r[[SP]]] +; CHECK: bl end_struct_small_nonchar + %x = alloca i32, align 4 + %y = alloca i32, align 4 + %z = alloca i32, align 4 + %ptr = alloca i32, align 4 + %small2 = alloca [2 x i16], align 2 + %large2 = alloca [8 x i32], align 16 + %small = alloca [2 x i8], align 1 + %large = alloca [8 x i8], align 1 + %a = alloca %struct.struct_large_char, align 1 + %b = alloca %struct.struct_small_char, align 1 + %c = alloca %struct.struct_large_nonchar, align 8 + %d = alloca %struct.struct_small_nonchar, align 2 + %call = call i32 @get_scalar1() + store i32 %call, i32* %x, align 4 + call void @end_scalar1() + %call1 = call i32 @get_scalar2() + store i32 %call1, i32* %y, align 4 + call void @end_scalar2() + %call2 = call i32 @get_scalar3() + store i32 %call2, i32* %z, align 4 + call void @end_scalar3() + %call3 = call i32 @get_addrof() + store i32 %call3, i32* %ptr, align 4 + call void @end_addrof() + %call4 = call signext i16 @get_small_nonchar() + %arrayidx = getelementptr inbounds [2 x i16]* %small2, i32 0, i64 0 + store i16 %call4, i16* %arrayidx, align 2 + call void @end_small_nonchar() + %call5 = call i32 @get_large_nonchar() + %arrayidx6 = getelementptr inbounds [8 x i32]* %large2, i32 0, i64 0 + store i32 %call5, i32* %arrayidx6, align 4 + call void @end_large_nonchar() + %call7 = call signext i8 @get_small_char() + %arrayidx8 = getelementptr inbounds [2 x i8]* %small, i32 0, i64 0 + store i8 %call7, i8* %arrayidx8, align 1 + call void @end_small_char() + %call9 = call signext i8 @get_large_char() + %arrayidx10 = getelementptr inbounds [8 x i8]* %large, i32 0, i64 0 + store i8 %call9, i8* %arrayidx10, align 1 + call void @end_large_char() + %call11 = call signext i8 @get_struct_large_char() + %foo = getelementptr inbounds %struct.struct_large_char* %a, i32 0, i32 0 + %arrayidx12 = getelementptr inbounds [8 x i8]* %foo, i32 0, i64 0 + store i8 %call11, i8* %arrayidx12, align 1 + call void @end_struct_large_char() + %call13 = call signext i8 @get_struct_small_char() + %foo14 = getelementptr inbounds %struct.struct_small_char* %b, i32 0, i32 0 + %arrayidx15 = getelementptr inbounds [2 x i8]* %foo14, i32 0, i64 0 + store i8 %call13, i8* %arrayidx15, align 1 + call void @end_struct_small_char() + %call16 = call i32 @get_struct_large_nonchar() + %foo17 = getelementptr inbounds %struct.struct_large_nonchar* %c, i32 0, i32 0 + %arrayidx18 = getelementptr inbounds [8 x i32]* %foo17, i32 0, i64 0 + store i32 %call16, i32* %arrayidx18, align 4 + call void @end_struct_large_nonchar() + %call19 = call signext i16 @get_struct_small_nonchar() + %foo20 = getelementptr inbounds %struct.struct_small_nonchar* %d, i32 0, i32 0 + %arrayidx21 = getelementptr inbounds [2 x i16]* %foo20, i32 0, i64 0 + store i16 %call19, i16* %arrayidx21, align 2 + call void @end_struct_small_nonchar() + %arraydecay = getelementptr inbounds [8 x i8]* %large, i32 0, i32 0 + %arraydecay22 = getelementptr inbounds [2 x i8]* %small, i32 0, i32 0 + %arraydecay23 = getelementptr inbounds [8 x i32]* %large2, i32 0, i32 0 + %arraydecay24 = getelementptr inbounds [2 x i16]* %small2, i32 0, i32 0 + %0 = load i32* %x, align 4 + %1 = load i32* %y, align 4 + %2 = load i32* %z, align 4 + %coerce.dive = getelementptr %struct.struct_large_char* %a, i32 0, i32 0 + %3 = bitcast [8 x i8]* %coerce.dive to i64* + %4 = load i64* %3, align 1 + %coerce.dive25 = getelementptr %struct.struct_small_char* %b, i32 0, i32 0 + %5 = bitcast [2 x i8]* %coerce.dive25 to i16* + %6 = load i16* %5, align 1 + %coerce.dive26 = getelementptr %struct.struct_small_nonchar* %d, i32 0, i32 0 + %7 = bitcast [2 x i16]* %coerce.dive26 to i32* + %8 = load i32* %7, align 1 + call void @takes_all(i64 %4, i16 %6, %struct.struct_large_nonchar* byval align 8 %c, i32 %8, i8* %arraydecay, i8* %arraydecay22, i32* %arraydecay23, i16* %arraydecay24, i32* %ptr, i32 %0, i32 %1, i32 %2) + ret void +} + +declare i32 @get_scalar1() +declare void @end_scalar1() + +declare i32 @get_scalar2() +declare void @end_scalar2() + +declare i32 @get_scalar3() +declare void @end_scalar3() + +declare i32 @get_addrof() +declare void @end_addrof() + +declare signext i16 @get_small_nonchar() +declare void @end_small_nonchar() + +declare i32 @get_large_nonchar() +declare void @end_large_nonchar() + +declare signext i8 @get_small_char() +declare void @end_small_char() + +declare signext i8 @get_large_char() +declare void @end_large_char() + +declare signext i8 @get_struct_large_char() +declare void @end_struct_large_char() + +declare signext i8 @get_struct_large_char2() +declare void @end_struct_large_char2() + +declare signext i8 @get_struct_small_char() +declare void @end_struct_small_char() + +declare i32 @get_struct_large_nonchar() +declare void @end_struct_large_nonchar() + +declare signext i16 @get_struct_small_nonchar() +declare void @end_struct_small_nonchar() + +declare void @takes_all(i64, i16, %struct.struct_large_nonchar* byval align 8, i32, i8*, i8*, i32*, i16*, i32*, i32, i32, i32) diff --git a/test/CodeGen/X86/ssp-data-layout.ll b/test/CodeGen/X86/ssp-data-layout.ll new file mode 100644 index 00000000000..72194af96f1 --- /dev/null +++ b/test/CodeGen/X86/ssp-data-layout.ll @@ -0,0 +1,238 @@ +; RUN: llc < %s -disable-fp-elim -mtriple=x86_64-pc-linux-gnu -mcpu=corei7 -o - | FileCheck %s +; This test is fairly fragile. The goal is to ensure that "large" stack +; objects are allocated closest to the stack protector (i.e., farthest away +; from the Stack Pointer.) In standard SSP mode this means that large (>= +; ssp-buffer-size) arrays and structures containing such arrays are +; closet to the protector. With sspstrong and sspreq this means large +; arrays/structures-with-arrays are closest, followed by small (< ssp-buffer-size) +; arrays/structures-with-arrays, and then addr-taken variables. +; +; Ideally, we only want verify that the objects appear in the correct groups +; and that the groups have the correct relative stack offset. The ordering +; within a group is not relevant to this test. Unfortunately, there is not +; an elegant way to do this, so just match the offset for each object. +; RUN: llc < %s -disable-fp-elim -mtriple=x86_64-unknown-unknown -O0 -mcpu=corei7 -o - \ +; RUN: | FileCheck --check-prefix=FAST-NON-LIN %s +; FastISel was not setting the StackProtectorIndex when lowering +; Intrinsic::stackprotector and as a result the stack re-arrangement code was +; never applied. This problem only shows up on non-Linux platforms because on +; Linux the stack protector cookie is loaded from a special address space which +; always triggers standard ISel. Run a basic test to ensure that at -O0 +; on a non-linux target the data layout rules are triggered. + +%struct.struct_large_char = type { [8 x i8] } +%struct.struct_large_char2 = type { [2 x i8], [8 x i8] } +%struct.struct_small_char = type { [2 x i8] } +%struct.struct_large_nonchar = type { [8 x i32] } +%struct.struct_small_nonchar = type { [2 x i16] } + +define void @layout_ssp() ssp { +entry: +; Expected stack layout for ssp is +; -16 large_char . Group 1, nested arrays, arrays >= ssp-buffer-size +; -24 struct_large_char . +; -28 scalar1 | Everything else +; -32 scalar2 +; -36 scalar3 +; -40 addr-of +; -44 small_nonchar +; -80 large_nonchar +; -82 small_char +; -88 struct_small_char +; -120 struct_large_nonchar +; -128 struct_small_nonchar + +; CHECK: layout_ssp: +; CHECK: call{{l|q}} get_scalar1 +; CHECK: movl %eax, -28( +; CHECK: call{{l|q}} end_scalar1 + +; CHECK: call{{l|q}} get_scalar2 +; CHECK: movl %eax, -32( +; CHECK: call{{l|q}} end_scalar2 + +; CHECK: call{{l|q}} get_scalar3 +; CHECK: movl %eax, -36( +; CHECK: call{{l|q}} end_scalar3 + +; CHECK: call{{l|q}} get_addrof +; CHECK: movl %eax, -40( +; CHECK: call{{l|q}} end_addrof + +; CHECK: get_small_nonchar +; CHECK: movw %ax, -44( +; CHECK: call{{l|q}} end_small_nonchar + +; CHECK: call{{l|q}} get_large_nonchar +; CHECK: movl %eax, -80( +; CHECK: call{{l|q}} end_large_nonchar + +; CHECK: call{{l|q}} get_small_char +; CHECK: movb %al, -82( +; CHECK: call{{l|q}} end_small_char + +; CHECK: call{{l|q}} get_large_char +; CHECK: movb %al, -16( +; CHECK: call{{l|q}} end_large_char + +; CHECK: call{{l|q}} get_struct_large_char +; CHECK: movb %al, -24( +; CHECK: call{{l|q}} end_struct_large_char + +; CHECK: call{{l|q}} get_struct_small_char +; CHECK: movb %al, -88( +; CHECK: call{{l|q}} end_struct_small_char + +; CHECK: call{{l|q}} get_struct_large_nonchar +; CHECK: movl %eax, -120( +; CHECK: call{{l|q}} end_struct_large_nonchar + +; CHECK: call{{l|q}} get_struct_small_nonchar +; CHECK: movw %ax, -128( +; CHECK: call{{l|q}} end_struct_small_nonchar + %x = alloca i32, align 4 + %y = alloca i32, align 4 + %z = alloca i32, align 4 + %ptr = alloca i32, align 4 + %small2 = alloca [2 x i16], align 2 + %large2 = alloca [8 x i32], align 16 + %small = alloca [2 x i8], align 1 + %large = alloca [8 x i8], align 1 + %a = alloca %struct.struct_large_char, align 1 + %b = alloca %struct.struct_small_char, align 1 + %c = alloca %struct.struct_large_nonchar, align 8 + %d = alloca %struct.struct_small_nonchar, align 2 + %call = call i32 @get_scalar1() + store i32 %call, i32* %x, align 4 + call void @end_scalar1() + %call1 = call i32 @get_scalar2() + store i32 %call1, i32* %y, align 4 + call void @end_scalar2() + %call2 = call i32 @get_scalar3() + store i32 %call2, i32* %z, align 4 + call void @end_scalar3() + %call3 = call i32 @get_addrof() + store i32 %call3, i32* %ptr, align 4 + call void @end_addrof() + %call4 = call signext i16 @get_small_nonchar() + %arrayidx = getelementptr inbounds [2 x i16]* %small2, i32 0, i64 0 + store i16 %call4, i16* %arrayidx, align 2 + call void @end_small_nonchar() + %call5 = call i32 @get_large_nonchar() + %arrayidx6 = getelementptr inbounds [8 x i32]* %large2, i32 0, i64 0 + store i32 %call5, i32* %arrayidx6, align 4 + call void @end_large_nonchar() + %call7 = call signext i8 @get_small_char() + %arrayidx8 = getelementptr inbounds [2 x i8]* %small, i32 0, i64 0 + store i8 %call7, i8* %arrayidx8, align 1 + call void @end_small_char() + %call9 = call signext i8 @get_large_char() + %arrayidx10 = getelementptr inbounds [8 x i8]* %large, i32 0, i64 0 + store i8 %call9, i8* %arrayidx10, align 1 + call void @end_large_char() + %call11 = call signext i8 @get_struct_large_char() + %foo = getelementptr inbounds %struct.struct_large_char* %a, i32 0, i32 0 + %arrayidx12 = getelementptr inbounds [8 x i8]* %foo, i32 0, i64 0 + store i8 %call11, i8* %arrayidx12, align 1 + call void @end_struct_large_char() + %call13 = call signext i8 @get_struct_small_char() + %foo14 = getelementptr inbounds %struct.struct_small_char* %b, i32 0, i32 0 + %arrayidx15 = getelementptr inbounds [2 x i8]* %foo14, i32 0, i64 0 + store i8 %call13, i8* %arrayidx15, align 1 + call void @end_struct_small_char() + %call16 = call i32 @get_struct_large_nonchar() + %foo17 = getelementptr inbounds %struct.struct_large_nonchar* %c, i32 0, i32 0 + %arrayidx18 = getelementptr inbounds [8 x i32]* %foo17, i32 0, i64 0 + store i32 %call16, i32* %arrayidx18, align 4 + call void @end_struct_large_nonchar() + %call19 = call signext i16 @get_struct_small_nonchar() + %foo20 = getelementptr inbounds %struct.struct_small_nonchar* %d, i32 0, i32 0 + %arrayidx21 = getelementptr inbounds [2 x i16]* %foo20, i32 0, i64 0 + store i16 %call19, i16* %arrayidx21, align 2 + call void @end_struct_small_nonchar() + %arraydecay = getelementptr inbounds [8 x i8]* %large, i32 0, i32 0 + %arraydecay22 = getelementptr inbounds [2 x i8]* %small, i32 0, i32 0 + %arraydecay23 = getelementptr inbounds [8 x i32]* %large2, i32 0, i32 0 + %arraydecay24 = getelementptr inbounds [2 x i16]* %small2, i32 0, i32 0 + %0 = load i32* %x, align 4 + %1 = load i32* %y, align 4 + %2 = load i32* %z, align 4 + %coerce.dive = getelementptr %struct.struct_large_char* %a, i32 0, i32 0 + %3 = bitcast [8 x i8]* %coerce.dive to i64* + %4 = load i64* %3, align 1 + %coerce.dive25 = getelementptr %struct.struct_small_char* %b, i32 0, i32 0 + %5 = bitcast [2 x i8]* %coerce.dive25 to i16* + %6 = load i16* %5, align 1 + %coerce.dive26 = getelementptr %struct.struct_small_nonchar* %d, i32 0, i32 0 + %7 = bitcast [2 x i16]* %coerce.dive26 to i32* + %8 = load i32* %7, align 1 + call void @takes_all(i64 %4, i16 %6, %struct.struct_large_nonchar* byval align 8 %c, i32 %8, i8* %arraydecay, i8* %arraydecay22, i32* %arraydecay23, i16* %arraydecay24, i32* %ptr, i32 %0, i32 %1, i32 %2) + ret void +} + +define void @fast_non_linux() ssp { +entry: +; FAST-NON-LIN: fast_non_linux: +; FAST-NON-LIN: call{{l|q}} get_scalar1 +; FAST-NON-LIN: movl %eax, -20( +; FAST-NON-LIN: call{{l|q}} end_scalar1 + +; FAST-NON-LIN: call{{l|q}} get_large_char +; FAST-NON-LIN: movb %al, -16( +; FAST-NON-LIN: call{{l|q}} end_large_char + %x = alloca i32, align 4 + %large = alloca [8 x i8], align 1 + %call = call i32 @get_scalar1() + store i32 %call, i32* %x, align 4 + call void @end_scalar1() + %call1 = call signext i8 @get_large_char() + %arrayidx = getelementptr inbounds [8 x i8]* %large, i32 0, i64 0 + store i8 %call1, i8* %arrayidx, align 1 + call void @end_large_char() + %0 = load i32* %x, align 4 + %arraydecay = getelementptr inbounds [8 x i8]* %large, i32 0, i32 0 + call void @takes_two(i32 %0, i8* %arraydecay) + ret void +} + +declare i32 @get_scalar1() +declare void @end_scalar1() + +declare i32 @get_scalar2() +declare void @end_scalar2() + +declare i32 @get_scalar3() +declare void @end_scalar3() + +declare i32 @get_addrof() +declare void @end_addrof() + +declare signext i16 @get_small_nonchar() +declare void @end_small_nonchar() + +declare i32 @get_large_nonchar() +declare void @end_large_nonchar() + +declare signext i8 @get_small_char() +declare void @end_small_char() + +declare signext i8 @get_large_char() +declare void @end_large_char() + +declare signext i8 @get_struct_large_char() +declare void @end_struct_large_char() + +declare signext i8 @get_struct_large_char2() +declare void @end_struct_large_char2() + +declare signext i8 @get_struct_small_char() +declare void @end_struct_small_char() + +declare i32 @get_struct_large_nonchar() +declare void @end_struct_large_nonchar() + +declare signext i16 @get_struct_small_nonchar() +declare void @end_struct_small_nonchar() + +declare void @takes_all(i64, i16, %struct.struct_large_nonchar* byval align 8, i32, i8*, i8*, i32*, i16*, i32*, i32, i32, i32) +declare void @takes_two(i32, i8*)