From owner-svn-src-stable-10@FreeBSD.ORG Wed Mar 26 07:42:48 2014 Return-Path: Delivered-To: svn-src-stable-10@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [8.8.178.115]) (using TLSv1 with cipher ADH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id 95757D4A; Wed, 26 Mar 2014 07:42:48 +0000 (UTC) Received: from svn.freebsd.org (svn.freebsd.org [IPv6:2001:1900:2254:2068::e6a:0]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mx1.freebsd.org (Postfix) with ESMTPS id 6745837B; Wed, 26 Mar 2014 07:42:48 +0000 (UTC) Received: from svn.freebsd.org ([127.0.1.70]) by svn.freebsd.org (8.14.8/8.14.8) with ESMTP id s2Q7gm8c058506; Wed, 26 Mar 2014 07:42:48 GMT (envelope-from dim@svn.freebsd.org) Received: (from dim@localhost) by svn.freebsd.org (8.14.8/8.14.8/Submit) id s2Q7gkPX058491; Wed, 26 Mar 2014 07:42:46 GMT (envelope-from dim@svn.freebsd.org) Message-Id: <201403260742.s2Q7gkPX058491@svn.freebsd.org> From: Dimitry Andric Date: Wed, 26 Mar 2014 07:42:46 +0000 (UTC) To: src-committers@freebsd.org, svn-src-all@freebsd.org, svn-src-stable@freebsd.org, svn-src-stable-10@freebsd.org Subject: svn commit: r263765 - in stable: 10/contrib/llvm/include/llvm/CodeGen 10/contrib/llvm/lib/CodeGen 10/contrib/llvm/lib/CodeGen/SelectionDAG 10/contrib/llvm/lib/MC/MCParser 10/contrib/llvm/lib/Target... X-SVN-Group: stable-10 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-BeenThere: svn-src-stable-10@freebsd.org X-Mailman-Version: 2.1.17 Precedence: list List-Id: SVN commit messages for only the 10-stable src tree List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 26 Mar 2014 07:42:48 -0000 Author: dim Date: Wed Mar 26 07:42:43 2014 New Revision: 263765 URL: http://svnweb.freebsd.org/changeset/base/263765 Log: MFC r263312: Pull in r196939 from upstream llvm trunk (by Reid Kleckner): Reland "Fix miscompile of MS inline assembly with stack realignment" This re-lands commit r196876, which was reverted in r196879. The tests have been fixed to pass on platforms with a stack alignment larger than 4. Update to clang side tests will land shortly. Pull in r196986 from upstream llvm trunk (by Reid Kleckner): Revert the backend fatal error from r196939 The combination of inline asm, stack realignment, and dynamic allocas turns out to be too common to reject out of hand. ASan inserts empy inline asm fragments and uses aligned allocas. Compiling any trivial function containing a dynamic alloca with ASan is enough to trigger the check. XFAIL the test cases that would be miscompiled and add one that uses the relevant functionality. Pull in r202930 from upstream llvm trunk (by Hans Wennborg): Check for dynamic allocas and inline asm that clobbers sp before building selection dag (PR19012) In X86SelectionDagInfo::EmitTargetCodeForMemcpy we check with MachineFrameInfo to make sure that ESI isn't used as a base pointer register before we choose to emit rep movs (which clobbers esi). The problem is that MachineFrameInfo wouldn't know about dynamic allocas or inline asm that clobbers the stack pointer until SelectionDAGBuilder has encountered them. This patch fixes the problem by checking for such things when building the FunctionLoweringInfo. Differential Revision: http://llvm-reviews.chandlerc.com/D2954 Together, these commits fix the problem encountered in the devel/emacs port on the i386 architecture, where a combination of stack realignment, alloca() and memcpy() could incidentally clobber the %esi register, leading to segfaults in the temacs build-time utility. See also: http://llvm.org/PR18171 and http://llvm.org/PR19012 Reported by: ashish PR: ports/183064 MFC r263313: Pull in r203311 from upstream llvm trunk (by Arnold Schwaighofer): ISel: Make VSELECT selection terminate in cases where the condition type has to be split and the result type widened. When the condition of a vselect has to be split it makes no sense widening the vselect and thereby widening the condition. We end up in an endless loop of widening (vselect result type) and splitting (condition mask type) doing this. Instead, split both the condition and the vselect and widen the result. I ran this over the test suite with i686 and mattr=+sse and saw no regressions. Fixes PR18036. With this fix the original problem case from the graphics/rawtherapee port (posted in http://llvm.org/PR18036 ) now compiles within ~97MB RSS. Reported by: mandree MFC r263320: Add separate patch files for all the customizations we have currently applied to our copy of llvm/clang. These can be applied in alphabetical order to a pristine llvm/clang 3.4 release source tree, to result in the same version used in FreeBSD. This is intended to clearly document all the changes until now, which mostly consist of cherry pickings from the respective upstream trunks, plus a number of hand-written FreeBSD-specific ones. Hopefully those can eventually be cleaned up and sent upstream too. Added: stable/10/contrib/llvm/patches/ - copied from r263320, head/contrib/llvm/patches/ Modified: stable/10/contrib/llvm/include/llvm/CodeGen/FunctionLoweringInfo.h stable/10/contrib/llvm/include/llvm/CodeGen/MachineFrameInfo.h stable/10/contrib/llvm/include/llvm/CodeGen/MachineFunction.h stable/10/contrib/llvm/lib/CodeGen/MachineFunction.cpp stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/FunctionLoweringInfo.cpp stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp stable/10/contrib/llvm/lib/MC/MCParser/AsmParser.cpp stable/10/contrib/llvm/lib/Target/X86/X86FrameLowering.cpp stable/10/contrib/llvm/lib/Target/X86/X86RegisterInfo.cpp Directory Properties: stable/10/ (props changed) Changes in other areas also in this revision: Added: stable/9/contrib/llvm/patches/ - copied from r263320, head/contrib/llvm/patches/ Modified: stable/9/contrib/llvm/include/llvm/CodeGen/FunctionLoweringInfo.h stable/9/contrib/llvm/include/llvm/CodeGen/MachineFrameInfo.h stable/9/contrib/llvm/include/llvm/CodeGen/MachineFunction.h stable/9/contrib/llvm/lib/CodeGen/MachineFunction.cpp stable/9/contrib/llvm/lib/CodeGen/SelectionDAG/FunctionLoweringInfo.cpp stable/9/contrib/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp stable/9/contrib/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp stable/9/contrib/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp stable/9/contrib/llvm/lib/MC/MCParser/AsmParser.cpp stable/9/contrib/llvm/lib/Target/X86/X86FrameLowering.cpp stable/9/contrib/llvm/lib/Target/X86/X86RegisterInfo.cpp Directory Properties: stable/9/ (props changed) stable/9/contrib/ (props changed) stable/9/contrib/llvm/ (props changed) Modified: stable/10/contrib/llvm/include/llvm/CodeGen/FunctionLoweringInfo.h ============================================================================== --- stable/10/contrib/llvm/include/llvm/CodeGen/FunctionLoweringInfo.h Wed Mar 26 07:35:24 2014 (r263764) +++ stable/10/contrib/llvm/include/llvm/CodeGen/FunctionLoweringInfo.h Wed Mar 26 07:42:43 2014 (r263765) @@ -41,6 +41,7 @@ class MachineBasicBlock; class MachineFunction; class MachineModuleInfo; class MachineRegisterInfo; +class SelectionDAG; class TargetLowering; class Value; @@ -125,7 +126,7 @@ public: /// set - Initialize this FunctionLoweringInfo with the given Function /// and its associated MachineFunction. /// - void set(const Function &Fn, MachineFunction &MF); + void set(const Function &Fn, MachineFunction &MF, SelectionDAG *DAG); /// clear - Clear out all the function-specific state. This returns this /// FunctionLoweringInfo to an empty state, ready to be used for a Modified: stable/10/contrib/llvm/include/llvm/CodeGen/MachineFrameInfo.h ============================================================================== --- stable/10/contrib/llvm/include/llvm/CodeGen/MachineFrameInfo.h Wed Mar 26 07:35:24 2014 (r263764) +++ stable/10/contrib/llvm/include/llvm/CodeGen/MachineFrameInfo.h Wed Mar 26 07:42:43 2014 (r263765) @@ -223,6 +223,10 @@ class MachineFrameInfo { /// Whether the "realign-stack" option is on. bool RealignOption; + /// True if the function includes inline assembly that adjusts the stack + /// pointer. + bool HasInlineAsmWithSPAdjust; + const TargetFrameLowering *getFrameLowering() const; public: explicit MachineFrameInfo(const TargetMachine &TM, bool RealignOpt) @@ -240,6 +244,7 @@ public: LocalFrameSize = 0; LocalFrameMaxAlign = 0; UseLocalStackAllocationBlock = false; + HasInlineAsmWithSPAdjust = false; } /// hasStackObjects - Return true if there are any stack objects in this @@ -451,6 +456,10 @@ public: bool hasCalls() const { return HasCalls; } void setHasCalls(bool V) { HasCalls = V; } + /// Returns true if the function contains any stack-adjusting inline assembly. + bool hasInlineAsmWithSPAdjust() const { return HasInlineAsmWithSPAdjust; } + void setHasInlineAsmWithSPAdjust(bool B) { HasInlineAsmWithSPAdjust = B; } + /// getMaxCallFrameSize - Return the maximum size of a call frame that must be /// allocated for an outgoing function call. This is only available if /// CallFrameSetup/Destroy pseudo instructions are used by the target, and @@ -521,7 +530,7 @@ 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. Modified: stable/10/contrib/llvm/include/llvm/CodeGen/MachineFunction.h ============================================================================== --- stable/10/contrib/llvm/include/llvm/CodeGen/MachineFunction.h Wed Mar 26 07:35:24 2014 (r263764) +++ stable/10/contrib/llvm/include/llvm/CodeGen/MachineFunction.h Wed Mar 26 07:42:43 2014 (r263765) @@ -131,8 +131,8 @@ class MachineFunction { /// about the control flow of such functions. bool ExposesReturnsTwice; - /// True if the function includes MS-style inline assembly. - bool HasMSInlineAsm; + /// True if the function includes any inline assembly. + bool HasInlineAsm; MachineFunction(const MachineFunction &) LLVM_DELETED_FUNCTION; void operator=(const MachineFunction&) LLVM_DELETED_FUNCTION; @@ -218,15 +218,14 @@ public: ExposesReturnsTwice = B; } - /// Returns true if the function contains any MS-style inline assembly. - bool hasMSInlineAsm() const { - return HasMSInlineAsm; + /// Returns true if the function contains any inline assembly. + bool hasInlineAsm() const { + return HasInlineAsm; } - /// Set a flag that indicates that the function contains MS-style inline - /// assembly. - void setHasMSInlineAsm(bool B) { - HasMSInlineAsm = B; + /// Set a flag that indicates that the function contains inline assembly. + void setHasInlineAsm(bool B) { + HasInlineAsm = B; } /// getInfo - Keep track of various per-function pieces of information for Modified: stable/10/contrib/llvm/lib/CodeGen/MachineFunction.cpp ============================================================================== --- stable/10/contrib/llvm/lib/CodeGen/MachineFunction.cpp Wed Mar 26 07:35:24 2014 (r263764) +++ stable/10/contrib/llvm/lib/CodeGen/MachineFunction.cpp Wed Mar 26 07:42:43 2014 (r263765) @@ -525,13 +525,14 @@ int MachineFrameInfo::CreateSpillStackOb /// 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, true, Alloca)); ensureMaxAlignment(Alignment); return (int)Objects.size()-NumFixedObjects-1; } Modified: stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/FunctionLoweringInfo.cpp ============================================================================== --- stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/FunctionLoweringInfo.cpp Wed Mar 26 07:35:24 2014 (r263764) +++ stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/FunctionLoweringInfo.cpp Wed Mar 26 07:42:43 2014 (r263765) @@ -33,6 +33,7 @@ #include "llvm/Support/ErrorHandling.h" #include "llvm/Support/MathExtras.h" #include "llvm/Target/TargetInstrInfo.h" +#include "llvm/Target/TargetFrameLowering.h" #include "llvm/Target/TargetLowering.h" #include "llvm/Target/TargetOptions.h" #include "llvm/Target/TargetRegisterInfo.h" @@ -55,7 +56,8 @@ static bool isUsedOutsideOfDefiningBlock return false; } -void FunctionLoweringInfo::set(const Function &fn, MachineFunction &mf) { +void FunctionLoweringInfo::set(const Function &fn, MachineFunction &mf, + SelectionDAG *DAG) { const TargetLowering *TLI = TM.getTargetLowering(); Fn = &fn; @@ -100,6 +102,43 @@ void FunctionLoweringInfo::set(const Fun for (; BB != EB; ++BB) for (BasicBlock::const_iterator I = BB->begin(), E = BB->end(); I != E; ++I) { + // Look for dynamic allocas. + if (const AllocaInst *AI = dyn_cast(I)) { + if (!AI->isStaticAlloca()) { + unsigned Align = std::max( + (unsigned)TLI->getDataLayout()->getPrefTypeAlignment( + AI->getAllocatedType()), + AI->getAlignment()); + unsigned StackAlign = TM.getFrameLowering()->getStackAlignment(); + if (Align <= StackAlign) + Align = 0; + // Inform the Frame Information that we have variable-sized objects. + MF->getFrameInfo()->CreateVariableSizedObject(Align ? Align : 1, AI); + } + } + + // Look for inline asm that clobbers the SP register. + if (isa(I) || isa(I)) { + ImmutableCallSite CS(I); + if (const InlineAsm *IA = dyn_cast(CS.getCalledValue())) { + unsigned SP = TLI->getStackPointerRegisterToSaveRestore(); + std::vector Ops = + TLI->ParseConstraints(CS); + for (size_t I = 0, E = Ops.size(); I != E; ++I) { + TargetLowering::AsmOperandInfo &Op = Ops[I]; + if (Op.Type == InlineAsm::isClobber) { + // Clobbers don't have SDValue operands, hence SDValue(). + TLI->ComputeConstraintToUse(Op, SDValue(), DAG); + std::pair PhysReg = + TLI->getRegForInlineAsmConstraint(Op.ConstraintCode, + Op.ConstraintVT); + if (PhysReg.first == SP) + MF->getFrameInfo()->setHasInlineAsmWithSPAdjust(true); + } + } + } + } + // Mark values used outside their block as exported, by allocating // a virtual register for them. if (isUsedOutsideOfDefiningBlock(I)) Modified: stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp ============================================================================== --- stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp Wed Mar 26 07:35:24 2014 (r263764) +++ stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/LegalizeVectorTypes.cpp Wed Mar 26 07:42:43 2014 (r263765) @@ -2180,6 +2180,17 @@ SDValue DAGTypeLegalizer::WidenVecRes_SE if (getTypeAction(CondVT) == TargetLowering::TypeWidenVector) Cond1 = GetWidenedVector(Cond1); + // If we have to split the condition there is no point in widening the + // select. This would result in an cycle of widening the select -> + // widening the condition operand -> splitting the condition operand -> + // splitting the select -> widening the select. Instead split this select + // further and widen the resulting type. + if (getTypeAction(CondVT) == TargetLowering::TypeSplitVector) { + SDValue SplitSelect = SplitVecOp_VSELECT(N, 0); + SDValue Res = ModifyToType(SplitSelect, WidenVT); + return Res; + } + if (Cond1.getValueType() != CondWidenVT) Cond1 = ModifyToType(Cond1, CondWidenVT); } Modified: stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp ============================================================================== --- stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp Wed Mar 26 07:35:24 2014 (r263764) +++ stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp Wed Mar 26 07:42:43 2014 (r263765) @@ -851,12 +851,20 @@ void RegsForValue::AddInlineAsmOperands( SDValue Res = DAG.getTargetConstant(Flag, MVT::i32); Ops.push_back(Res); + unsigned SP = TLI.getStackPointerRegisterToSaveRestore(); for (unsigned Value = 0, Reg = 0, e = ValueVTs.size(); Value != e; ++Value) { unsigned NumRegs = TLI.getNumRegisters(*DAG.getContext(), ValueVTs[Value]); MVT RegisterVT = RegVTs[Value]; for (unsigned i = 0; i != NumRegs; ++i) { assert(Reg < Regs.size() && "Mismatch in # registers expected"); - Ops.push_back(DAG.getRegister(Regs[Reg++], RegisterVT)); + unsigned TheReg = Regs[Reg++]; + Ops.push_back(DAG.getRegister(TheReg, RegisterVT)); + + if (TheReg == SP && Code == InlineAsm::Kind_Clobber) { + // If we clobbered the stack pointer, MFI should know about it. + assert(DAG.getMachineFunction().getFrameInfo()-> + hasInlineAsmWithSPAdjust()); + } } } } @@ -3370,9 +3378,7 @@ void SelectionDAGBuilder::visitAlloca(co setValue(&I, DSA); DAG.setRoot(DSA.getValue(1)); - // Inform the Frame Information that we have just allocated a variable-sized - // object. - FuncInfo.MF->getFrameInfo()->CreateVariableSizedObject(Align ? Align : 1); + assert(FuncInfo.MF->getFrameInfo()->hasVarSizedObjects()); } void SelectionDAGBuilder::visitLoad(const LoadInst &I) { Modified: stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp ============================================================================== --- stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp Wed Mar 26 07:35:24 2014 (r263764) +++ stable/10/contrib/llvm/lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp Wed Mar 26 07:42:43 2014 (r263765) @@ -419,7 +419,7 @@ bool SelectionDAGISel::runOnMachineFunct SplitCriticalSideEffectEdges(const_cast(Fn), this); CurDAG->init(*MF, TTI, TLI); - FuncInfo->set(Fn, *MF); + FuncInfo->set(Fn, *MF, CurDAG); if (UseMBPI && OptLevel != CodeGenOpt::None) FuncInfo->BPI = &getAnalysis(); @@ -428,7 +428,8 @@ bool SelectionDAGISel::runOnMachineFunct SDB->init(GFI, *AA, LibInfo); - MF->setHasMSInlineAsm(false); + MF->setHasInlineAsm(false); + SelectAllBasicBlocks(Fn); // If the first basic block in the function has live ins that need to be @@ -511,7 +512,7 @@ bool SelectionDAGISel::runOnMachineFunct for (MachineFunction::const_iterator I = MF->begin(), E = MF->end(); I != E; ++I) { - if (MFI->hasCalls() && MF->hasMSInlineAsm()) + if (MFI->hasCalls() && MF->hasInlineAsm()) break; const MachineBasicBlock *MBB = I; @@ -522,8 +523,8 @@ bool SelectionDAGISel::runOnMachineFunct II->isStackAligningInlineAsm()) { MFI->setHasCalls(true); } - if (II->isMSInlineAsm()) { - MF->setHasMSInlineAsm(true); + if (II->isInlineAsm()) { + MF->setHasInlineAsm(true); } } } Modified: stable/10/contrib/llvm/lib/MC/MCParser/AsmParser.cpp ============================================================================== --- stable/10/contrib/llvm/lib/MC/MCParser/AsmParser.cpp Wed Mar 26 07:35:24 2014 (r263764) +++ stable/10/contrib/llvm/lib/MC/MCParser/AsmParser.cpp Wed Mar 26 07:42:43 2014 (r263765) @@ -4192,6 +4192,11 @@ bool AsmParser::parseMSInlineAsm( AsmStrRewrites.push_back(AsmRewrite(AOK_Input, Start, SymName.size())); } } + + // Consider implicit defs to be clobbers. Think of cpuid and push. + const uint16_t *ImpDefs = Desc.getImplicitDefs(); + for (unsigned I = 0, E = Desc.getNumImplicitDefs(); I != E; ++I) + ClobberRegs.push_back(ImpDefs[I]); } // Set the number of Outputs and Inputs. Modified: stable/10/contrib/llvm/lib/Target/X86/X86FrameLowering.cpp ============================================================================== --- stable/10/contrib/llvm/lib/Target/X86/X86FrameLowering.cpp Wed Mar 26 07:35:24 2014 (r263764) +++ stable/10/contrib/llvm/lib/Target/X86/X86FrameLowering.cpp Wed Mar 26 07:42:43 2014 (r263765) @@ -50,7 +50,7 @@ bool X86FrameLowering::hasFP(const Machi return (MF.getTarget().Options.DisableFramePointerElim(MF) || RegInfo->needsStackRealignment(MF) || MFI->hasVarSizedObjects() || - MFI->isFrameAddressTaken() || MF.hasMSInlineAsm() || + MFI->isFrameAddressTaken() || MFI->hasInlineAsmWithSPAdjust() || MF.getInfo()->getForceFramePointer() || MMI.callsUnwindInit() || MMI.callsEHReturn()); } Modified: stable/10/contrib/llvm/lib/Target/X86/X86RegisterInfo.cpp ============================================================================== --- stable/10/contrib/llvm/lib/Target/X86/X86RegisterInfo.cpp Wed Mar 26 07:35:24 2014 (r263764) +++ stable/10/contrib/llvm/lib/Target/X86/X86RegisterInfo.cpp Wed Mar 26 07:42:43 2014 (r263765) @@ -403,18 +403,15 @@ bool X86RegisterInfo::hasBasePointer(con if (!EnableBasePointer) return false; - // When we need stack realignment and there are dynamic allocas, we can't - // reference off of the stack pointer, so we reserve a base pointer. - // - // This is also true if the function contain MS-style inline assembly. We - // do this because if any stack changes occur in the inline assembly, e.g., - // "pusha", then any C local variable or C argument references in the - // inline assembly will be wrong because the SP is not properly tracked. - if ((needsStackRealignment(MF) && MFI->hasVarSizedObjects()) || - MF.hasMSInlineAsm()) - return true; - - return false; + // When we need stack realignment, we can't address the stack from the frame + // pointer. When we have dynamic allocas or stack-adjusting inline asm, we + // can't address variables from the stack pointer. MS inline asm can + // reference locals while also adjusting the stack pointer. When we can't + // use both the SP and the FP, we need a separate base pointer register. + bool CantUseFP = needsStackRealignment(MF); + bool CantUseSP = + MFI->hasVarSizedObjects() || MFI->hasInlineAsmWithSPAdjust(); + return CantUseFP && CantUseSP; } bool X86RegisterInfo::canRealignStack(const MachineFunction &MF) const {