[CodeGen][X86] Fix lowering of tailcalls when -ms-hotpatch is used

llvm/include/llvm/Support/TargetOpcodes.def

@@ -170,15 +170,10 @@ HANDLE_TARGET_OPCODE(LOCAL_ESCAPE)
 /// comparisons into existing memory operations.
 HANDLE_TARGET_OPCODE(FAULTING_OP)
 
-/// Wraps a machine instruction to add patchability constraints.  An
-/// instruction wrapped in PATCHABLE_OP has to either have a minimum
+/// Precedes a machine instruction to add patchability constraints.  An
+/// instruction after PATCHABLE_OP has to either have a minimum
 /// size or be preceded with a nop of that size.  The first operand is
-/// an immediate denoting the minimum size of the instruction, the
-/// second operand is an immediate denoting the opcode of the original
-/// instruction.  The rest of the operands are the operands of the
-/// original instruction.
-/// PATCHABLE_OP can be used as second operand to only insert a nop of
-/// required size.
+/// an immediate denoting the minimum size of the following instruction.
 HANDLE_TARGET_OPCODE(PATCHABLE_OP)
 
 /// This is a marker instruction which gets translated into a nop sled, useful

llvm/lib/CodeGen/PatchableFunction.cpp

@@ -38,58 +38,28 @@ struct PatchableFunction : public MachineFunctionPass {
 }
 
 bool PatchableFunction::runOnMachineFunction(MachineFunction &MF) {
+  MachineBasicBlock &FirstMBB = *MF.begin();
+
   if (MF.getFunction().hasFnAttribute("patchable-function-entry")) {
-    MachineBasicBlock &FirstMBB = *MF.begin();
     const TargetInstrInfo *TII = MF.getSubtarget().getInstrInfo();
     // The initial .loc covers PATCHABLE_FUNCTION_ENTER.
     BuildMI(FirstMBB, FirstMBB.begin(), DebugLoc(),
             TII->get(TargetOpcode::PATCHABLE_FUNCTION_ENTER));
     return true;
-  }
-
-  if (!MF.getFunction().hasFnAttribute("patchable-function"))
-    return false;
-
+  } else if (MF.getFunction().hasFnAttribute("patchable-function")) {
 #ifndef NDEBUG
-  Attribute PatchAttr = MF.getFunction().getFnAttribute("patchable-function");
-  StringRef PatchType = PatchAttr.getValueAsString();
-  assert(PatchType == "prologue-short-redirect" && "Only possibility today!");
+    Attribute PatchAttr = MF.getFunction().getFnAttribute("patchable-function");
+    StringRef PatchType = PatchAttr.getValueAsString();
+    assert(PatchType == "prologue-short-redirect" && "Only possibility today!");
 #endif
-
-  auto &FirstMBB = *MF.begin();
-  auto *TII = MF.getSubtarget().getInstrInfo();
-
-  MachineBasicBlock::iterator FirstActualI = llvm::find_if(
-      FirstMBB, [](const MachineInstr &MI) { return !MI.isMetaInstruction(); });
-
-  if (FirstActualI == FirstMBB.end()) {
-    // As of Microsoft documentation on /hotpatch feature, we must ensure that
-    // "the first instruction of each function is at least two bytes, and no
-    // jump within the function goes to the first instruction"
-
-    // When the first MBB is empty, insert a patchable no-op. This ensures the
-    // first instruction is patchable in two special cases:
-    // - the function is empty (e.g. unreachable)
-    // - the function jumps back to the first instruction, which is in a
-    // successor MBB.
-    BuildMI(&FirstMBB, DebugLoc(), TII->get(TargetOpcode::PATCHABLE_OP))
-        .addImm(2)
-        .addImm(TargetOpcode::PATCHABLE_OP);
+    auto *TII = MF.getSubtarget().getInstrInfo();
+    BuildMI(FirstMBB, FirstMBB.begin(), DebugLoc(),
+            TII->get(TargetOpcode::PATCHABLE_OP))
+        .addImm(2);
     MF.ensureAlignment(Align(16));
     return true;
   }
-
-  auto MIB = BuildMI(FirstMBB, FirstActualI, FirstActualI->getDebugLoc(),
-                     TII->get(TargetOpcode::PATCHABLE_OP))
-                 .addImm(2)
-                 .addImm(FirstActualI->getOpcode());
-
-  for (auto &MO : FirstActualI->operands())
-    MIB.add(MO);
-
-  FirstActualI->eraseFromParent();
-  MF.ensureAlignment(Align(16));
-  return true;
+  return false;
 }
 
 char PatchableFunction::ID = 0;

llvm/lib/Target/X86/X86MCInstLower.cpp

@@ -948,24 +948,26 @@ void X86AsmPrinter::LowerASAN_CHECK_MEMACCESS(const MachineInstr &MI) {
 
 void X86AsmPrinter::LowerPATCHABLE_OP(const MachineInstr &MI,
                                       X86MCInstLower &MCIL) {
-  // PATCHABLE_OP minsize, opcode, operands
+  // PATCHABLE_OP minsize
 
   NoAutoPaddingScope NoPadScope(*OutStreamer);
 
-  unsigned MinSize = MI.getOperand(0).getImm();
-  unsigned Opcode = MI.getOperand(1).getImm();
-  // Opcode PATCHABLE_OP is a special case: there is no instruction to wrap,
-  // simply emit a nop of size MinSize.
-  bool EmptyInst = (Opcode == TargetOpcode::PATCHABLE_OP);
-
-  MCInst MCI;
-  MCI.setOpcode(Opcode);
-  for (auto &MO : drop_begin(MI.operands(), 2))
-    if (auto MaybeOperand = MCIL.LowerMachineOperand(&MI, MO))
-      MCI.addOperand(*MaybeOperand);
+  // Find the next MachineInstr in this MBB.
+  const MachineInstr *NextMI = MI.getNextNode();
+  while (NextMI) {
+    if (!NextMI->isMetaInstruction())
+      break;
+    NextMI = NextMI->getNextNode();
+  }
 
   SmallString<256> Code;
-  if (!EmptyInst) {
+  unsigned MinSize = MI.getOperand(0).getImm();
+
+  if (NextMI) {
+    // Lower the next MachineInstr to find its byte size.
+    MCInst MCI;
+    MCIL.Lower(NextMI, MCI);
+
     SmallVector<MCFixup, 4> Fixups;
     CodeEmitter->encodeInstruction(MCI, Code, Fixups, getSubtargetInfo());
   }
@@ -981,21 +983,12 @@ void X86AsmPrinter::LowerPATCHABLE_OP(const MachineInstr &MI,
       OutStreamer->emitInstruction(
           MCInstBuilder(X86::MOV32rr_REV).addReg(X86::EDI).addReg(X86::EDI),
           *Subtarget);
-    } else if (MinSize == 2 && Opcode == X86::PUSH64r) {
-      // This is an optimization that lets us get away without emitting a nop in
-      // many cases.
-      //
-      // NB! In some cases the encoding for PUSH64r (e.g. PUSH64r %r9) takes two
-      // bytes too, so the check on MinSize is important.
-      MCI.setOpcode(X86::PUSH64rmr);
     } else {
       unsigned NopSize = emitNop(*OutStreamer, MinSize, Subtarget);
       assert(NopSize == MinSize && "Could not implement MinSize!");
       (void)NopSize;
     }
   }
-  if (!EmptyInst)
-    OutStreamer->emitInstruction(MCI, getSubtargetInfo());
 }
 
 // Lower a stackmap of the form:

llvm/test/CodeGen/X86/patchable-prologue-tailcall.ll

@@ -0,0 +1,55 @@
+; RUN: llc -verify-machineinstrs < %s | FileCheck %s --check-prefix=CHECK
+
+; CHECK: "?mi_new_test@@YAPEAX_K@Z":
+; CHECK-NEXT: # %bb.0:
+; CHECK-NEXT: jmp     mi_new                          # TAILCALL
+
+; CHECK: "?builtin_malloc_test@@YAPEAX_K@Z":
+; CHECK-NEXT: # %bb.0:
+; CHECK-NEXT: jmp     malloc                          # TAILCALL
+
+; // Built with: clang-cl.exe /c patchable-prologue-tailcall.cpp /O2 /hotpatch -Xclang -emit-llvm
+;
+; typedef unsigned long long size_t;
+; 
+; extern "C" {
+;     void* mi_new(size_t size);
+;  }
+;
+; void *mi_new_test(size_t count)
+; {
+;     return mi_new(count);
+; }
+; 
+; void *builtin_malloc_test(size_t count)
+; {
+;     return __builtin_malloc(count);
+; }
+
+target datalayout = "e-m:w-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-pc-windows-msvc19.38.33133"
+
+define dso_local noundef ptr @"?mi_new_test@@YAPEAX_K@Z"(i64 noundef %count) local_unnamed_addr "patchable-function"="prologue-short-redirect" {
+entry:
+  %call = tail call ptr @mi_new(i64 noundef %count)
+  ret ptr %call
+}
+
+declare dso_local ptr @mi_new(i64 noundef) local_unnamed_addr
+
+define dso_local noalias noundef ptr @"?builtin_malloc_test@@YAPEAX_K@Z"(i64 noundef %count) local_unnamed_addr "patchable-function"="prologue-short-redirect" {
+entry:
+  %call = tail call ptr @malloc(i64 noundef %count)
+  ret ptr %call
+}
+
+declare dso_local noalias noundef ptr @malloc(i64 noundef) local_unnamed_addr #0
+
+attributes #0 = { allockind("alloc,uninitialized") allocsize(0) memory(inaccessiblemem: readwrite) "alloc-family"="malloc" }
+
+!llvm.module.flags = !{!0, !1, !2, !3}
+
+!0 = !{i32 1, !"wchar_size", i32 2}
+!1 = !{i32 8, !"PIC Level", i32 2}
+!2 = !{i32 7, !"uwtable", i32 2}
+!3 = !{i32 1, !"MaxTLSAlign", i32 65536}

llvm/test/CodeGen/X86/patchable-prologue.ll

@@ -32,7 +32,8 @@ define void @f0() "patchable-function"="prologue-short-redirect" {
 
 define void @f1() "patchable-function"="prologue-short-redirect" "frame-pointer"="all" {
 ; CHECK-LABEL: _f1
-; CHECK-NEXT: ff f5 	pushq	%rbp
+; CHECK-NEXT: 66 90     nop
+; CHECK-NEXT: 55		pushq	%rbp
 
 ; CHECK-ALIGN: 	.p2align	4, 0x90
 ; CHECK-ALIGN: _f1:
@@ -47,6 +48,7 @@ define void @f1() "patchable-function"="prologue-short-redirect" "frame-pointer"
 ; 64: f1:
 ; 64-NEXT: .seh_proc f1
 ; 64-NEXT: # %bb.0:
+; 64-NEXT: xchgw %ax, %ax
 ; 64-NEXT: pushq   %rbp
 
   ret void