llvm-project/llvm/lib/Target/Sparc/DelaySlotFiller.cpp

512 lines
15 KiB
C++

//===-- DelaySlotFiller.cpp - SPARC delay slot filler ---------------------===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//
//
// This is a simple local pass that attempts to fill delay slots with useful
// instructions. If no instructions can be moved into the delay slot, then a
// NOP is placed.
//===----------------------------------------------------------------------===//
#include "Sparc.h"
#include "SparcSubtarget.h"
#include "llvm/ADT/SmallSet.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/CodeGen/MachineFunctionPass.h"
#include "llvm/CodeGen/MachineInstrBuilder.h"
#include "llvm/CodeGen/MachineRegisterInfo.h"
#include "llvm/CodeGen/TargetInstrInfo.h"
#include "llvm/CodeGen/TargetRegisterInfo.h"
#include "llvm/Support/CommandLine.h"
#include "llvm/Target/TargetMachine.h"
using namespace llvm;
#define DEBUG_TYPE "delay-slot-filler"
STATISTIC(FilledSlots, "Number of delay slots filled");
static cl::opt<bool> DisableDelaySlotFiller(
"disable-sparc-delay-filler",
cl::init(false),
cl::desc("Disable the Sparc delay slot filler."),
cl::Hidden);
namespace {
struct Filler : public MachineFunctionPass {
const SparcSubtarget *Subtarget = nullptr;
static char ID;
Filler() : MachineFunctionPass(ID) {}
StringRef getPassName() const override { return "SPARC Delay Slot Filler"; }
bool runOnMachineBasicBlock(MachineBasicBlock &MBB);
bool runOnMachineFunction(MachineFunction &F) override {
bool Changed = false;
Subtarget = &F.getSubtarget<SparcSubtarget>();
// This pass invalidates liveness information when it reorders
// instructions to fill delay slot.
F.getRegInfo().invalidateLiveness();
for (MachineFunction::iterator FI = F.begin(), FE = F.end();
FI != FE; ++FI)
Changed |= runOnMachineBasicBlock(*FI);
return Changed;
}
MachineFunctionProperties getRequiredProperties() const override {
return MachineFunctionProperties().set(
MachineFunctionProperties::Property::NoVRegs);
}
void insertCallDefsUses(MachineBasicBlock::iterator MI,
SmallSet<unsigned, 32>& RegDefs,
SmallSet<unsigned, 32>& RegUses);
void insertDefsUses(MachineBasicBlock::iterator MI,
SmallSet<unsigned, 32>& RegDefs,
SmallSet<unsigned, 32>& RegUses);
bool IsRegInSet(SmallSet<unsigned, 32>& RegSet,
unsigned Reg);
bool delayHasHazard(MachineBasicBlock::iterator candidate,
bool &sawLoad, bool &sawStore,
SmallSet<unsigned, 32> &RegDefs,
SmallSet<unsigned, 32> &RegUses);
MachineBasicBlock::iterator
findDelayInstr(MachineBasicBlock &MBB, MachineBasicBlock::iterator slot);
bool needsUnimp(MachineBasicBlock::iterator I, unsigned &StructSize);
bool tryCombineRestoreWithPrevInst(MachineBasicBlock &MBB,
MachineBasicBlock::iterator MBBI);
};
char Filler::ID = 0;
} // end of anonymous namespace
/// createSparcDelaySlotFillerPass - Returns a pass that fills in delay
/// slots in Sparc MachineFunctions
///
FunctionPass *llvm::createSparcDelaySlotFillerPass() {
return new Filler;
}
/// runOnMachineBasicBlock - Fill in delay slots for the given basic block.
/// We assume there is only one delay slot per delayed instruction.
///
bool Filler::runOnMachineBasicBlock(MachineBasicBlock &MBB) {
bool Changed = false;
Subtarget = &MBB.getParent()->getSubtarget<SparcSubtarget>();
const TargetInstrInfo *TII = Subtarget->getInstrInfo();
for (MachineBasicBlock::iterator I = MBB.begin(); I != MBB.end(); ) {
MachineBasicBlock::iterator MI = I;
++I;
// If MI is restore, try combining it with previous inst.
if (!DisableDelaySlotFiller &&
(MI->getOpcode() == SP::RESTORErr
|| MI->getOpcode() == SP::RESTOREri)) {
Changed |= tryCombineRestoreWithPrevInst(MBB, MI);
continue;
}
// TODO: If we ever want to support v7, this needs to be extended
// to cover all floating point operations.
if (!Subtarget->isV9() &&
(MI->getOpcode() == SP::FCMPS || MI->getOpcode() == SP::FCMPD
|| MI->getOpcode() == SP::FCMPQ)) {
BuildMI(MBB, I, MI->getDebugLoc(), TII->get(SP::NOP));
Changed = true;
continue;
}
// If MI has no delay slot, skip.
if (!MI->hasDelaySlot())
continue;
MachineBasicBlock::iterator D = MBB.end();
if (!DisableDelaySlotFiller)
D = findDelayInstr(MBB, MI);
++FilledSlots;
Changed = true;
if (D == MBB.end())
BuildMI(MBB, I, MI->getDebugLoc(), TII->get(SP::NOP));
else
MBB.splice(I, &MBB, D);
unsigned structSize = 0;
if (needsUnimp(MI, structSize)) {
MachineBasicBlock::iterator J = MI;
++J; // skip the delay filler.
assert (J != MBB.end() && "MI needs a delay instruction.");
BuildMI(MBB, ++J, MI->getDebugLoc(),
TII->get(SP::UNIMP)).addImm(structSize);
// Bundle the delay filler and unimp with the instruction.
MIBundleBuilder(MBB, MachineBasicBlock::iterator(MI), J);
} else {
MIBundleBuilder(MBB, MachineBasicBlock::iterator(MI), I);
}
}
return Changed;
}
MachineBasicBlock::iterator
Filler::findDelayInstr(MachineBasicBlock &MBB,
MachineBasicBlock::iterator slot)
{
SmallSet<unsigned, 32> RegDefs;
SmallSet<unsigned, 32> RegUses;
bool sawLoad = false;
bool sawStore = false;
if (slot == MBB.begin())
return MBB.end();
if (slot->getOpcode() == SP::RET || slot->getOpcode() == SP::TLS_CALL)
return MBB.end();
if (slot->getOpcode() == SP::RETL) {
MachineBasicBlock::iterator J = slot;
--J;
if (J->getOpcode() == SP::RESTORErr
|| J->getOpcode() == SP::RESTOREri) {
// change retl to ret.
slot->setDesc(Subtarget->getInstrInfo()->get(SP::RET));
return J;
}
}
// Call's delay filler can def some of call's uses.
if (slot->isCall())
insertCallDefsUses(slot, RegDefs, RegUses);
else
insertDefsUses(slot, RegDefs, RegUses);
bool done = false;
MachineBasicBlock::iterator I = slot;
while (!done) {
done = (I == MBB.begin());
if (!done)
--I;
// skip debug instruction
if (I->isDebugInstr())
continue;
if (I->hasUnmodeledSideEffects() || I->isInlineAsm() || I->isPosition() ||
I->hasDelaySlot() || I->isBundledWithSucc())
break;
if (delayHasHazard(I, sawLoad, sawStore, RegDefs, RegUses)) {
insertDefsUses(I, RegDefs, RegUses);
continue;
}
return I;
}
return MBB.end();
}
bool Filler::delayHasHazard(MachineBasicBlock::iterator candidate,
bool &sawLoad,
bool &sawStore,
SmallSet<unsigned, 32> &RegDefs,
SmallSet<unsigned, 32> &RegUses)
{
if (candidate->isImplicitDef() || candidate->isKill())
return true;
if (candidate->mayLoad()) {
sawLoad = true;
if (sawStore)
return true;
}
if (candidate->mayStore()) {
if (sawStore)
return true;
sawStore = true;
if (sawLoad)
return true;
}
for (unsigned i = 0, e = candidate->getNumOperands(); i!= e; ++i) {
const MachineOperand &MO = candidate->getOperand(i);
if (!MO.isReg())
continue; // skip
Register Reg = MO.getReg();
if (MO.isDef()) {
// check whether Reg is defined or used before delay slot.
if (IsRegInSet(RegDefs, Reg) || IsRegInSet(RegUses, Reg))
return true;
}
if (MO.isUse()) {
// check whether Reg is defined before delay slot.
if (IsRegInSet(RegDefs, Reg))
return true;
}
}
unsigned Opcode = candidate->getOpcode();
// LD and LDD may have NOPs inserted afterwards in the case of some LEON
// processors, so we can't use the delay slot if this feature is switched-on.
if (Subtarget->insertNOPLoad()
&&
Opcode >= SP::LDDArr && Opcode <= SP::LDrr)
return true;
// Same as above for FDIV and FSQRT on some LEON processors.
if (Subtarget->fixAllFDIVSQRT()
&&
Opcode >= SP::FDIVD && Opcode <= SP::FSQRTD)
return true;
return false;
}
void Filler::insertCallDefsUses(MachineBasicBlock::iterator MI,
SmallSet<unsigned, 32>& RegDefs,
SmallSet<unsigned, 32>& RegUses)
{
// Call defines o7, which is visible to the instruction in delay slot.
RegDefs.insert(SP::O7);
switch(MI->getOpcode()) {
default: llvm_unreachable("Unknown opcode.");
case SP::CALL: break;
case SP::CALLrr:
case SP::CALLri:
assert(MI->getNumOperands() >= 2);
const MachineOperand &Reg = MI->getOperand(0);
assert(Reg.isReg() && "CALL first operand is not a register.");
assert(Reg.isUse() && "CALL first operand is not a use.");
RegUses.insert(Reg.getReg());
const MachineOperand &Operand1 = MI->getOperand(1);
if (Operand1.isImm() || Operand1.isGlobal())
break;
assert(Operand1.isReg() && "CALLrr second operand is not a register.");
assert(Operand1.isUse() && "CALLrr second operand is not a use.");
RegUses.insert(Operand1.getReg());
break;
}
}
// Insert Defs and Uses of MI into the sets RegDefs and RegUses.
void Filler::insertDefsUses(MachineBasicBlock::iterator MI,
SmallSet<unsigned, 32>& RegDefs,
SmallSet<unsigned, 32>& RegUses)
{
for (unsigned i = 0, e = MI->getNumOperands(); i != e; ++i) {
const MachineOperand &MO = MI->getOperand(i);
if (!MO.isReg())
continue;
Register Reg = MO.getReg();
if (Reg == 0)
continue;
if (MO.isDef())
RegDefs.insert(Reg);
if (MO.isUse()) {
// Implicit register uses of retl are return values and
// retl does not use them.
if (MO.isImplicit() && MI->getOpcode() == SP::RETL)
continue;
RegUses.insert(Reg);
}
}
}
// returns true if the Reg or its alias is in the RegSet.
bool Filler::IsRegInSet(SmallSet<unsigned, 32>& RegSet, unsigned Reg)
{
// Check Reg and all aliased Registers.
for (MCRegAliasIterator AI(Reg, Subtarget->getRegisterInfo(), true);
AI.isValid(); ++AI)
if (RegSet.count(*AI))
return true;
return false;
}
bool Filler::needsUnimp(MachineBasicBlock::iterator I, unsigned &StructSize)
{
if (!I->isCall())
return false;
unsigned structSizeOpNum = 0;
switch (I->getOpcode()) {
default: llvm_unreachable("Unknown call opcode.");
case SP::CALL: structSizeOpNum = 1; break;
case SP::CALLrr:
case SP::CALLri: structSizeOpNum = 2; break;
case SP::TLS_CALL: return false;
}
const MachineOperand &MO = I->getOperand(structSizeOpNum);
if (!MO.isImm())
return false;
StructSize = MO.getImm();
return true;
}
static bool combineRestoreADD(MachineBasicBlock::iterator RestoreMI,
MachineBasicBlock::iterator AddMI,
const TargetInstrInfo *TII)
{
// Before: add <op0>, <op1>, %i[0-7]
// restore %g0, %g0, %i[0-7]
//
// After : restore <op0>, <op1>, %o[0-7]
Register reg = AddMI->getOperand(0).getReg();
if (reg < SP::I0 || reg > SP::I7)
return false;
// Erase RESTORE.
RestoreMI->eraseFromParent();
// Change ADD to RESTORE.
AddMI->setDesc(TII->get((AddMI->getOpcode() == SP::ADDrr)
? SP::RESTORErr
: SP::RESTOREri));
// Map the destination register.
AddMI->getOperand(0).setReg(reg - SP::I0 + SP::O0);
return true;
}
static bool combineRestoreOR(MachineBasicBlock::iterator RestoreMI,
MachineBasicBlock::iterator OrMI,
const TargetInstrInfo *TII)
{
// Before: or <op0>, <op1>, %i[0-7]
// restore %g0, %g0, %i[0-7]
// and <op0> or <op1> is zero,
//
// After : restore <op0>, <op1>, %o[0-7]
Register reg = OrMI->getOperand(0).getReg();
if (reg < SP::I0 || reg > SP::I7)
return false;
// check whether it is a copy.
if (OrMI->getOpcode() == SP::ORrr
&& OrMI->getOperand(1).getReg() != SP::G0
&& OrMI->getOperand(2).getReg() != SP::G0)
return false;
if (OrMI->getOpcode() == SP::ORri
&& OrMI->getOperand(1).getReg() != SP::G0
&& (!OrMI->getOperand(2).isImm() || OrMI->getOperand(2).getImm() != 0))
return false;
// Erase RESTORE.
RestoreMI->eraseFromParent();
// Change OR to RESTORE.
OrMI->setDesc(TII->get((OrMI->getOpcode() == SP::ORrr)
? SP::RESTORErr
: SP::RESTOREri));
// Map the destination register.
OrMI->getOperand(0).setReg(reg - SP::I0 + SP::O0);
return true;
}
static bool combineRestoreSETHIi(MachineBasicBlock::iterator RestoreMI,
MachineBasicBlock::iterator SetHiMI,
const TargetInstrInfo *TII)
{
// Before: sethi imm3, %i[0-7]
// restore %g0, %g0, %g0
//
// After : restore %g0, (imm3<<10), %o[0-7]
Register reg = SetHiMI->getOperand(0).getReg();
if (reg < SP::I0 || reg > SP::I7)
return false;
if (!SetHiMI->getOperand(1).isImm())
return false;
int64_t imm = SetHiMI->getOperand(1).getImm();
// Is it a 3 bit immediate?
if (!isInt<3>(imm))
return false;
// Make it a 13 bit immediate.
imm = (imm << 10) & 0x1FFF;
assert(RestoreMI->getOpcode() == SP::RESTORErr);
RestoreMI->setDesc(TII->get(SP::RESTOREri));
RestoreMI->getOperand(0).setReg(reg - SP::I0 + SP::O0);
RestoreMI->getOperand(1).setReg(SP::G0);
RestoreMI->getOperand(2).ChangeToImmediate(imm);
// Erase the original SETHI.
SetHiMI->eraseFromParent();
return true;
}
bool Filler::tryCombineRestoreWithPrevInst(MachineBasicBlock &MBB,
MachineBasicBlock::iterator MBBI)
{
// No previous instruction.
if (MBBI == MBB.begin())
return false;
// assert that MBBI is a "restore %g0, %g0, %g0".
assert(MBBI->getOpcode() == SP::RESTORErr
&& MBBI->getOperand(0).getReg() == SP::G0
&& MBBI->getOperand(1).getReg() == SP::G0
&& MBBI->getOperand(2).getReg() == SP::G0);
MachineBasicBlock::iterator PrevInst = std::prev(MBBI);
// It cannot be combined with a bundled instruction.
if (PrevInst->isBundledWithSucc())
return false;
const TargetInstrInfo *TII = Subtarget->getInstrInfo();
switch (PrevInst->getOpcode()) {
default: break;
case SP::ADDrr:
case SP::ADDri: return combineRestoreADD(MBBI, PrevInst, TII); break;
case SP::ORrr:
case SP::ORri: return combineRestoreOR(MBBI, PrevInst, TII); break;
case SP::SETHIi: return combineRestoreSETHIi(MBBI, PrevInst, TII); break;
}
// It cannot combine with the previous instruction.
return false;
}