[X86] Put no-op ADJCALLSTACK markers around all dynamic lowerings

[rnk]

Summary

These ADJCALLSTACK markers don't generate code, but they keep dynamic
alloca code that calls chkstk out of the prologue.

[qcolombet]

Hi Reid,

LGTM.

Thanks,
Q.

llvm/trunk/lib/Target/X86/X86ISelLowering.cpp

 X86TargetLowering::LowerDYNAMIC_STACKALLOC(SDValue Op,
                                            SelectionDAG &DAG) const {
   MachineFunction &MF = DAG.getMachineFunction();
   bool SplitStack = MF.shouldSplitStack();
   bool Lower = (Subtarget->isOSWindows() && !Subtarget->isTargetMachO()) ||
                SplitStack;
   SDLoc dl(Op);
 
-  if (!Lower) {
-    const TargetLowering &TLI = DAG.getTargetLoweringInfo();
-    SDNode* Node = Op.getNode();
+  // Get the inputs.
+  SDNode *Node = Op.getNode();
+  SDValue Chain = Op.getOperand(0);
+  SDValue Size  = Op.getOperand(1);
+  unsigned Align = cast<ConstantSDNode>(Op.getOperand(2))->getZExtValue();
+  EVT VT = Node->getValueType(0);
+
+  // Chain the dynamic stack allocation so that it doesn't modify the stack
+  // pointer when other instructions are using the stack.
+  Chain = DAG.getCALLSEQ_START(Chain, DAG.getIntPtrConstant(0, dl, true), dl);
 
+  bool Is64Bit = Subtarget->is64Bit();
+  MVT SPTy = getPointerTy(DAG.getDataLayout());
+
+  SDValue Result;
+  if (!Lower) {
+    const TargetLowering &TLI = DAG.getTargetLoweringInfo();
     unsigned SPReg = TLI.getStackPointerRegisterToSaveRestore();
     assert(SPReg && "Target cannot require DYNAMIC_STACKALLOC expansion and"
                     " not tell us which reg is the stack pointer!");
     EVT VT = Node->getValueType(0);
-    SDValue Tmp1 = SDValue(Node, 0);
-    SDValue Tmp2 = SDValue(Node, 1);
     SDValue Tmp3 = Node->getOperand(2);
-    SDValue Chain = Tmp1.getOperand(0);
 
-    // Chain the dynamic stack allocation so that it doesn't modify the stack
-    // pointer when other instructions are using the stack.
-    Chain = DAG.getCALLSEQ_START(Chain, DAG.getIntPtrConstant(0, dl, true),
-        SDLoc(Node));
-
-    SDValue Size = Tmp2.getOperand(1);
     SDValue SP = DAG.getCopyFromReg(Chain, dl, SPReg, VT);
     Chain = SP.getValue(1);
     unsigned Align = cast<ConstantSDNode>(Tmp3)->getZExtValue();
     const TargetFrameLowering &TFI = *Subtarget->getFrameLowering();
     unsigned StackAlign = TFI.getStackAlignment();
-    Tmp1 = DAG.getNode(ISD::SUB, dl, VT, SP, Size); // Value
+    Result = DAG.getNode(ISD::SUB, dl, VT, SP, Size); // Value
     if (Align > StackAlign)
-      Tmp1 = DAG.getNode(ISD::AND, dl, VT, Tmp1,
+      Result = DAG.getNode(ISD::AND, dl, VT, Result,
                          DAG.getConstant(-(uint64_t)Align, dl, VT));
-    Chain = DAG.getCopyToReg(Chain, dl, SPReg, Tmp1); // Output chain
-
-    Tmp2 = DAG.getCALLSEQ_END(Chain, DAG.getIntPtrConstant(0, dl, true),
-        DAG.getIntPtrConstant(0, dl, true), SDValue(),
-        SDLoc(Node));
-
-    SDValue Ops[2] = { Tmp1, Tmp2 };
-    return DAG.getMergeValues(Ops, dl);
-  }
-
-  // Get the inputs.
-  SDValue Chain = Op.getOperand(0);
-  SDValue Size  = Op.getOperand(1);
-  unsigned Align = cast<ConstantSDNode>(Op.getOperand(2))->getZExtValue();
-  EVT VT = Op.getNode()->getValueType(0);
-
-  bool Is64Bit = Subtarget->is64Bit();
-  MVT SPTy = getPointerTy(DAG.getDataLayout());
-
-  if (SplitStack) {
+    Chain = DAG.getCopyToReg(Chain, dl, SPReg, Result); // Output chain
+  } else if (SplitStack) {
     MachineRegisterInfo &MRI = MF.getRegInfo();
 
     if (Is64Bit) {
       // The 64 bit implementation of segmented stacks needs to clobber both r10
       // r11. This makes it impossible to use it along with nested parameters.
       const Function *F = MF.getFunction();
 
       for (Function::const_arg_iterator I = F->arg_begin(), E = F->arg_end();
            I != E; ++I)
         if (I->hasNestAttr())
           report_fatal_error("Cannot use segmented stacks with functions that "
                              "have nested arguments.");
     }
 
     const TargetRegisterClass *AddrRegClass = getRegClassFor(SPTy);
     unsigned Vreg = MRI.createVirtualRegister(AddrRegClass);
     Chain = DAG.getCopyToReg(Chain, dl, Vreg, Size);
-    SDValue Value = DAG.getNode(X86ISD::SEG_ALLOCA, dl, SPTy, Chain,
+    Result = DAG.getNode(X86ISD::SEG_ALLOCA, dl, SPTy, Chain,
                                 DAG.getRegister(Vreg, SPTy));
-    SDValue Ops1[2] = { Value, Chain };
-    return DAG.getMergeValues(Ops1, dl);
   } else {
     SDValue Flag;
     const unsigned Reg = (Subtarget->isTarget64BitLP64() ? X86::RAX : X86::EAX);
 
     Chain = DAG.getCopyToReg(Chain, dl, Reg, Size, Flag);
     Flag = Chain.getValue(1);
     SDVTList NodeTys = DAG.getVTList(MVT::Other, MVT::Glue);
 
     Chain = DAG.getNode(X86ISD::WIN_ALLOCA, dl, NodeTys, Chain, Flag);
 
     const X86RegisterInfo *RegInfo = Subtarget->getRegisterInfo();
     unsigned SPReg = RegInfo->getStackRegister();
     SDValue SP = DAG.getCopyFromReg(Chain, dl, SPReg, SPTy);
     Chain = SP.getValue(1);
 
     if (Align) {
       SP = DAG.getNode(ISD::AND, dl, VT, SP.getValue(0),
                        DAG.getConstant(-(uint64_t)Align, dl, VT));
       Chain = DAG.getCopyToReg(Chain, dl, SPReg, SP);
     }
 
-    SDValue Ops1[2] = { SP, Chain };
-    return DAG.getMergeValues(Ops1, dl);
+    Result = SP;
   }
+
+  Chain = DAG.getCALLSEQ_END(Chain, DAG.getIntPtrConstant(0, dl, true),
+                             DAG.getIntPtrConstant(0, dl, true), SDValue(), dl);
+
+  SDValue Ops[2] = {Result, Chain};
+  return DAG.getMergeValues(Ops, dl);
 }
 
 SDValue X86TargetLowering::LowerVASTART(SDValue Op, SelectionDAG &DAG) const {
   MachineFunction &MF = DAG.getMachineFunction();
   auto PtrVT = getPointerTy(MF.getDataLayout());
   X86MachineFunctionInfo *FuncInfo = MF.getInfo<X86MachineFunctionInfo>();
 
   const Value *SV = cast<SrcValueSDNode>(Op.getOperand(2))->getValue();

llvm/trunk/test/CodeGen/X86/inalloca-stdcall.ll

 ; RUN: llc < %s -mtriple=i686-pc-win32 | FileCheck %s
 
 %Foo = type { i32, i32 }
 
 declare x86_stdcallcc void @f(%Foo* inalloca %a)
 declare x86_stdcallcc void @i(i32 %a)
 
 define void @g() {
 ; CHECK-LABEL: _g:
   %b = alloca inalloca %Foo
 ; CHECK: movl    $8, %eax
 ; CHECK: calll   __chkstk
   %f1 = getelementptr %Foo, %Foo* %b, i32 0, i32 0
   %f2 = getelementptr %Foo, %Foo* %b, i32 0, i32 1
   store i32 13, i32* %f1
   store i32 42, i32* %f2
-; CHECK: movl    $13, (%esp)
-; CHECK: movl    $42, 4(%esp)
+; CHECK: movl %esp, %eax
+; CHECK: movl    $13, (%eax)
+; CHECK: movl    $42, 4(%eax)
   call x86_stdcallcc void @f(%Foo* inalloca %b)
 ; CHECK: calll   _f@8
 ; CHECK-NOT: %esp
 ; CHECK: pushl
 ; CHECK: calll   _i@4
   call x86_stdcallcc void @i(i32 0)
   ret void
 }

llvm/trunk/test/CodeGen/X86/inalloca.ll

 ; RUN: llc < %s -mtriple=i686-pc-win32 | FileCheck %s
 
 %Foo = type { i32, i32 }
 
 declare void @f(%Foo* inalloca %b)
 
 define void @a() {
 ; CHECK-LABEL: _a:
 entry:
   %b = alloca inalloca %Foo
 ; CHECK: movl    $8, %eax
 ; CHECK: calll   __chkstk
   %f1 = getelementptr %Foo, %Foo* %b, i32 0, i32 0
   %f2 = getelementptr %Foo, %Foo* %b, i32 0, i32 1
   store i32 13, i32* %f1
   store i32 42, i32* %f2
-; CHECK: movl    $13, (%esp)
-; CHECK: movl    $42, 4(%esp)
+; CHECK: movl %esp, %eax
+; CHECK: movl    $13, (%eax)
+; CHECK: movl    $42, 4(%eax)
   call void @f(%Foo* inalloca %b)
 ; CHECK: calll   _f
   ret void
 }
 
 declare void @inreg_with_inalloca(i32 inreg %a, %Foo* inalloca %b)
 
 define void @b() {
 ; CHECK-LABEL: _b:
 entry:
   %b = alloca inalloca %Foo
 ; CHECK: movl    $8, %eax
 ; CHECK: calll   __chkstk
   %f1 = getelementptr %Foo, %Foo* %b, i32 0, i32 0
   %f2 = getelementptr %Foo, %Foo* %b, i32 0, i32 1
   store i32 13, i32* %f1
   store i32 42, i32* %f2
-; CHECK: movl    $13, (%esp)
-; CHECK: movl    $42, 4(%esp)
+; CHECK: movl %esp, %eax
+; CHECK: movl    $13, (%eax)
+; CHECK: movl    $42, 4(%eax)
   call void @inreg_with_inalloca(i32 inreg 1, %Foo* inalloca %b)
 ; CHECK: movl    $1, %eax
 ; CHECK: calll   _inreg_with_inalloca
   ret void
 }
 
 declare x86_thiscallcc void @thiscall_with_inalloca(i8* %a, %Foo* inalloca %b)
 
 define void @c() {
 ; CHECK-LABEL: _c:
 entry:
   %b = alloca inalloca %Foo
 ; CHECK: movl    $8, %eax
 ; CHECK: calll   __chkstk
   %f1 = getelementptr %Foo, %Foo* %b, i32 0, i32 0
   %f2 = getelementptr %Foo, %Foo* %b, i32 0, i32 1
   store i32 13, i32* %f1
   store i32 42, i32* %f2
-; CHECK-DAG: movl    $13, (%esp)
-; CHECK-DAG: movl    $42, 4(%esp)
+; CHECK: movl %esp, %eax
+; CHECK-DAG: movl    $13, (%eax)
+; CHECK-DAG: movl    $42, 4(%eax)
   call x86_thiscallcc void @thiscall_with_inalloca(i8* null, %Foo* inalloca %b)
 ; CHECK-DAG: xorl    %ecx, %ecx
 ; CHECK: calll   _thiscall_with_inalloca
   ret void
 }

llvm/trunk/test/CodeGen/X86/shrink-wrap-chkstk.ll

+; RUN: llc < %s -enable-shrink-wrap=true | FileCheck %s
+
+; chkstk cannot come before the usual prologue, since it adjusts ESP.
+
+target datalayout = "e-m:x-p:32:32-i64:64-f80:32-n8:16:32-a:0:32-S32"
+target triple = "i686-pc-windows-msvc18.0.0"
+
+%struct.S = type { [12 x i8] }
+
+define x86_thiscallcc void @call_inalloca(i1 %x) {
+entry:
+  %argmem = alloca inalloca <{ %struct.S }>, align 4
+  %argidx1 = getelementptr inbounds <{ %struct.S }>, <{ %struct.S }>* %argmem, i32 0, i32 0, i32 0, i32 0
+  %argidx2 = getelementptr inbounds <{ %struct.S }>, <{ %struct.S }>* %argmem, i32 0, i32 0, i32 0, i32 1
+  store i8 42, i8* %argidx2, align 4
+  br i1 %x, label %bb1, label %bb2
+
+bb1:
+  store i8 42, i8* %argidx1, align 4
+  br label %bb2
+
+bb2:
+  call void @inalloca_params(<{ %struct.S }>* inalloca nonnull %argmem)
+  ret void
+}
+
+; CHECK-LABEL: _call_inalloca: # @call_inalloca
+; CHECK: pushl %ebp
+; CHECK: movl %esp, %ebp
+; CHECK: movl $12, %eax
+; CHECK: calll __chkstk
+; CHECK: calll _inalloca_params
+; CHECK: movl %ebp, %esp
+; CHECK: popl %ebp
+; CHECK: retl
+
+declare void @inalloca_params(<{ %struct.S }>* inalloca)