[mips][FastISel] Implement intrinsics memset, memcopy & memmove.

[vkalintiris]

Summary

Implement the intrinsics memset, memcopy and memmove in MIPS FastISel.
Make some needed infrastructure fixes so that this can work.

Based on a patch by Reed Kotler.

[dsanders]

lib/Target/Mips/MipsFastISel.cpp

494–495

Given that InMBB only makes sense when V is an Instruction, why don't we test it on line 494?

1233–1241

Not required for this patch but we shouldn't be using JALR for the majority of calls

test/CodeGen/Mips/Fast-ISel/memtest1.ll

18

Should also check the jalr, the arguments and the return.

Likewise for mov(), and clear() below

53–61

Nit: Unnecessary.

[vkalintiris]

Addressed review comments.

[dsanders]

LGTM

lib/Target/Mips/MipsFastISel.cpp

480

Nit: Could you add a comment about why we require the same BB?

lib/Target/Mips/MipsFastISel.cpp

class MipsFastISel final : public FastISel {
   const TargetInstrInfo &TII;
   const TargetLowering &TLI;
   MipsFunctionInfo *MFI;
 
   // Convenience variables to avoid some queries.
   LLVMContext *Context;
 
   bool fastLowerCall(CallLoweringInfo &CLI) override;
+  bool fastLowerIntrinsicCall(const IntrinsicInst *II) override;
 
   bool TargetSupported;
   bool UnsupportedFPMode; // To allow fast-isel to proceed and just not handle
   // floating point but not reject doing fast-isel in other
   // situations
 
 private:
   // Selection routines.
private:
 
   unsigned emitLogicalOp(unsigned ISDOpc, MVT RetVT, const Value *LHS,
                          const Value *RHS);
 
   unsigned materializeFP(const ConstantFP *CFP, MVT VT);
   unsigned materializeGV(const GlobalValue *GV, MVT VT);
   unsigned materializeInt(const Constant *C, MVT VT);
   unsigned materialize32BitInt(int64_t Imm, const TargetRegisterClass *RC);
+  unsigned materializeExternalCallSym(const char *SynName);
 
   MachineInstrBuilder emitInst(unsigned Opc) {
     return BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(Opc));
   }
   MachineInstrBuilder emitInst(unsigned Opc, unsigned DstReg) {
     return BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(Opc),
                    DstReg);
   }
if ((GV->hasInternalLinkage() ||
     emitInst(Mips::ADDiu, TempReg)
         .addReg(DestReg)
         .addGlobalAddress(GV, 0, MipsII::MO_ABS_LO);
     DestReg = TempReg;
   }
   return DestReg;
 }
 
+unsigned MipsFastISel::materializeExternalCallSym(const char *SymName) {
+  const TargetRegisterClass *RC = &Mips::GPR32RegClass;
+  unsigned DestReg = createResultReg(RC);
+  emitInst(Mips::LW, DestReg)
+      .addReg(MFI->getGlobalBaseReg())
+      .addExternalSymbol(SymName, MipsII::MO_GOT);
+  return DestReg;
+}
+
 // Materialize a constant into a register, and return the register
 // number (or zero if we failed to handle it).
 unsigned MipsFastISel::fastMaterializeConstant(const Constant *C) {
   EVT CEVT = TLI.getValueType(C->getType(), true);
 
   // Only handle simple types.
   if (!CEVT.isSimple())
     return 0;
case Instruction::Alloca: {
     break;
   }
   }
   Addr.setReg(getRegForValue(Obj));
   return Addr.getReg() != 0;
 }
 
 bool MipsFastISel::computeCallAddress(const Value *V, Address &Addr) {
-  const GlobalValue *GV = dyn_cast<GlobalValue>(V);
-  if (GV && isa<Function>(GV) && cast<Function>(GV)->isIntrinsic())
-    return false;
-  if (!GV)
-    return false;
+  const User *U = nullptr;
+  unsigned Opcode = Instruction::UserOp1;
+
+  if (const auto *I = dyn_cast<Instruction>(V)) {
+    if (I->getParent() == FuncInfo.MBB->getBasicBlock()) {
+      Opcode = I->getOpcode();
+      U = I;
+    }
+  } else if (const auto *C = dyn_cast<ConstantExpr>(V)) {
+    Opcode = C->getOpcode();
+    U = C;
+  }
+
+  switch (Opcode) {
+  default:
+    break;
+  case Instruction::BitCast:
+    // Look past bitcasts if its operand is in the same BB.
+      return computeCallAddress(U->getOperand(0), Addr);
+    break;
+  case Instruction::IntToPtr:
+    // Look past no-op inttoptrs if its operand is in the same BB.
+    if (TLI.getValueType(U->getOperand(0)->getType()) == TLI.getPointerTy())
+      return computeCallAddress(U->getOperand(0), Addr);
+    break;
+  case Instruction::PtrToInt:
+    // Look past no-op ptrtoints if its operand is in the same BB.
+    if (TLI.getValueType(U->getType()) == TLI.getPointerTy())
+      return computeCallAddress(U->getOperand(0), Addr);
+    break;
+  }
+
   if (const GlobalValue *GV = dyn_cast<GlobalValue>(V)) {
     Addr.setGlobalValue(GV);
     return true;
   }
+
+  // If all else fails, try to materialize the value in a register.
+  if (!Addr.getGlobalValue()) {
+    Addr.setReg(getRegForValue(V));
+    return Addr.getReg() != 0;
+  }
+
   return false;
 }
 
 bool MipsFastISel::isTypeLegal(Type *Ty, MVT &VT) {
   EVT evt = TLI.getValueType(Ty, true);
   // Only handle simple types.
   if (evt == MVT::Other || !evt.isSimple())
     return false;
bool MipsFastISel::finishCall(CallLoweringInfo &CLI, MVT RetVT,
   return true;
 }
 
 bool MipsFastISel::fastLowerCall(CallLoweringInfo &CLI) {
   CallingConv::ID CC = CLI.CallConv;
   bool IsTailCall = CLI.IsTailCall;
   bool IsVarArg = CLI.IsVarArg;
   const Value *Callee = CLI.Callee;
-  // const char *SymName = CLI.SymName;
+  const char *SymName = CLI.SymName;
 
   // Allow SelectionDAG isel to handle tail calls.
   if (IsTailCall)
     return false;
 
   // Let SDISel handle vararg functions.
   if (IsVarArg)
     return false;
bool MipsFastISel::fastLowerCall(CallLoweringInfo &CLI) {
   if (!computeCallAddress(Callee, Addr))
     return false;
 
   // Handle the arguments now that we've gotten them.
   unsigned NumBytes;
   if (!processCallArgs(CLI, OutVTs, NumBytes))
     return false;
 
+  if (!Addr.getGlobalValue())
+    return false;
+
   // Issue the call.
-  unsigned DestAddress = materializeGV(Addr.getGlobalValue(), MVT::i32);
+  unsigned DestAddress;
+  if (SymName)
+    DestAddress = materializeExternalCallSym(SymName);
+  else
+    DestAddress = materializeGV(Addr.getGlobalValue(), MVT::i32);
   emitInst(TargetOpcode::COPY, Mips::T9).addReg(DestAddress);
   MachineInstrBuilder MIB =
       BuildMI(*FuncInfo.MBB, FuncInfo.InsertPt, DbgLoc, TII.get(Mips::JALR),
               Mips::RA).addReg(Mips::T9);
 
   // Add implicit physical register uses to the call.
   for (auto Reg : CLI.OutRegs)
     MIB.addReg(Reg, RegState::Implicit);
 
   // Add a register mask with the call-preserved registers.
   // Proper defs for return values will be added by setPhysRegsDeadExcept().
   MIB.addRegMask(TRI.getCallPreservedMask(*FuncInfo.MF, CC));
 
   CLI.Call = MIB;
 
   // Finish off the call including any return values.
   return finishCall(CLI, RetVT, NumBytes);
 }
 
+bool MipsFastISel::fastLowerIntrinsicCall(const IntrinsicInst *II) {
+  switch (II->getIntrinsicID()) {
+  default:
+    return false;
+  case Intrinsic::memcpy:
+  case Intrinsic::memmove: {
+    const auto *MTI = cast<MemTransferInst>(II);
+    // Don't handle volatile.
+    if (MTI->isVolatile())
+      return false;
+    if (!MTI->getLength()->getType()->isIntegerTy(32))
+      return false;
+    const char *IntrMemName = isa<MemCpyInst>(II) ? "memcpy" : "memmove";
+    return lowerCallTo(II, IntrMemName, II->getNumArgOperands() - 2);
+  }
+  case Intrinsic::memset: {
+    const MemSetInst *MSI = cast<MemSetInst>(II);
+    // Don't handle volatile.
+    if (MSI->isVolatile())
+      return false;
+    if (!MSI->getLength()->getType()->isIntegerTy(32))
+      return false;
+    return lowerCallTo(II, "memset", II->getNumArgOperands() - 2);
+  }
+  }
+  return false;
+}
+
 bool MipsFastISel::selectRet(const Instruction *I) {
   const Function &F = *I->getParent()->getParent();
   const ReturnInst *Ret = cast<ReturnInst>(I);
 
   if (!FuncInfo.CanLowerReturn)
     return false;
 
   // Build a list of return value registers.

test/CodeGen/Mips/Fast-ISel/memtest1.ll

+; RUN: llc < %s -march=mipsel -mcpu=mips32 -O0 -relocation-model=pic \
+; RUN:     -fast-isel=true -mips-fast-isel -fast-isel-abort=1 | FileCheck %s \
+; RUN:     -check-prefix=ALL -check-prefix=32R1
+; RUN: llc < %s -march=mipsel -mcpu=mips32r2 -O0 -relocation-model=pic \
+; RUN:     -fast-isel=true -mips-fast-isel -fast-isel-abort=1 | FileCheck %s \
+; RUN:     -check-prefix=ALL -check-prefix=32R2
+
+@str = private unnamed_addr constant [12 x i8] c"hello there\00", align 1
+@src = global i8* getelementptr inbounds ([12 x i8], [12 x i8]* @str, i32 0, i32 0), align 4
+@i = global i32 12, align 4
+@dest = common global [50 x i8] zeroinitializer, align 1
+
+declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture readonly, i32, i32, i1)
+declare void @llvm.memmove.p0i8.p0i8.i32(i8* nocapture, i8* nocapture readonly, i32, i32, i1)
+declare void @llvm.memset.p0i8.i32(i8* nocapture, i8, i32, i32, i1)
+
+define void @cpy(i8* %src, i32 %i) {
+  ; ALL-LABEL:  cpy:
+
+  ; ALL-DAG:        lw    $[[T0:[0-9]+]], %got(dest)(${{[0-9]+}})
+  ; ALL-DAG:        sw    $4, 24($sp)
+  ; ALL-DAG:        move  $4, $[[T0]]
+  ; ALL-DAG:        sw    $5, 20($sp)
+  ; ALL-DAG:        lw    $[[T1:[0-9]+]], 24($sp)
+  ; ALL-DAG:        move  $5, $[[T1]]
+  ; ALL-DAG:        lw    $6, 20($sp)
+  ; ALL-DAG:        lw    $[[T2:[0-9]+]], %got(memcpy)(${{[0-9]+}})
+  ; ALL:            jalr  $[[T2]]
+  ; ALL-NEXT:       nop
+  ; ALL-NOT:        {{.*}}$2{{.*}}
+  call void @llvm.memcpy.p0i8.p0i8.i32(i8* getelementptr inbounds ([50 x i8], [50 x i8]* @dest, i32 0, i32 0),
+                                       i8* %src, i32 %i, i32 1, i1 false)
+  ret void
+}
+
+define void @mov(i8* %src, i32 %i) {
+  ; ALL-LABEL:  mov:
+
+
+  ; ALL-DAG:        lw    $[[T0:[0-9]+]], %got(dest)(${{[0-9]+}})
+  ; ALL-DAG:        sw    $4, 24($sp)
+  ; ALL-DAG:        move  $4, $[[T0]]
+  ; ALL-DAG:        sw    $5, 20($sp)
+  ; ALL-DAG:        lw    $[[T1:[0-9]+]], 24($sp)
+  ; ALL-DAG:        move  $5, $[[T1]]
+  ; ALL-DAG:        lw    $6, 20($sp)
+  ; ALL-DAG:        lw    $[[T2:[0-9]+]], %got(memmove)(${{[0-9]+}})
+  ; ALL:            jalr  $[[T2]]
+  ; ALL-NEXT:       nop
+  ; ALL-NOT:        {{.*}}$2{{.*}}
+  call void @llvm.memmove.p0i8.p0i8.i32(i8* getelementptr inbounds ([50 x i8], [50 x i8]* @dest, i32 0, i32 0),
+                                        i8* %src, i32 %i, i32 1, i1 false)
+  ret void
+}
+
+define void @clear(i32 %i) {
+  ; ALL-LABEL:  clear:
+
+  ; ALL-DAG:        lw    $[[T0:[0-9]+]], %got(dest)(${{[0-9]+}})
+  ; ALL-DAG:        sw    $4, 16($sp)
+  ; ALL-DAG:        move  $4, $[[T0]]
+  ; ALL-DAG:        addiu $[[T1:[0-9]+]], $zero, 42
+  ; 32R1-DAG:       sll   $[[T2:[0-9]+]], $[[T1]], 24
+  ; 32R1-DAG:       sra   $5, $[[T2]], 24
+  ; 32R2-DAG:       seb   $5, $[[T1]]
+  ; ALL-DAG:        lw    $6, 16($sp)
+  ; ALL-DAG:        lw    $[[T2:[0-9]+]], %got(memset)(${{[0-9]+}})
+  ; ALL:            jalr  $[[T2]]
+  ; ALL-NEXT:       nop
+  ; ALL-NOT:        {{.*}}$2{{.*}}
+  call void @llvm.memset.p0i8.i32(i8* getelementptr inbounds ([50 x i8], [50 x i8]* @dest, i32 0, i32 0),
+                                  i8 42, i32 %i, i32 1, i1 false)
+  ret void
+}