llvm-6502/lib/Target/ARM/ARMConstantIslandPass.cpp

1894 lines
71 KiB
C++
Raw Normal View History

//===-- ARMConstantIslandPass.cpp - ARM constant islands --------*- C++ -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
// This file contains a pass that splits the constant pool up into 'islands'
// which are scattered through-out the function. This is required due to the
// limited pc-relative displacements that ARM has.
//
//===----------------------------------------------------------------------===//
#define DEBUG_TYPE "arm-cp-islands"
#include "ARM.h"
#include "ARMAddressingModes.h"
#include "ARMMachineFunctionInfo.h"
#include "ARMInstrInfo.h"
#include "Thumb2InstrInfo.h"
#include "llvm/CodeGen/MachineConstantPool.h"
#include "llvm/CodeGen/MachineFunctionPass.h"
#include "llvm/CodeGen/MachineJumpTableInfo.h"
#include "llvm/Target/TargetData.h"
#include "llvm/Target/TargetMachine.h"
#include "llvm/Support/Debug.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/Support/raw_ostream.h"
#include "llvm/ADT/SmallSet.h"
#include "llvm/ADT/SmallVector.h"
#include "llvm/ADT/STLExtras.h"
#include "llvm/ADT/Statistic.h"
#include "llvm/Support/CommandLine.h"
#include <algorithm>
using namespace llvm;
STATISTIC(NumCPEs, "Number of constpool entries");
STATISTIC(NumSplit, "Number of uncond branches inserted");
STATISTIC(NumCBrFixed, "Number of cond branches fixed");
STATISTIC(NumUBrFixed, "Number of uncond branches fixed");
STATISTIC(NumTBs, "Number of table branches generated");
STATISTIC(NumT2CPShrunk, "Number of Thumb2 constantpool instructions shrunk");
STATISTIC(NumT2BrShrunk, "Number of Thumb2 immediate branches shrunk");
STATISTIC(NumCBZ, "Number of CBZ / CBNZ formed");
STATISTIC(NumJTMoved, "Number of jump table destination blocks moved");
STATISTIC(NumJTInserted, "Number of jump table intermediate blocks inserted");
static cl::opt<bool>
AdjustJumpTableBlocks("arm-adjust-jump-tables", cl::Hidden, cl::init(true),
cl::desc("Adjust basic block layout to better use TB[BH]"));
namespace {
/// ARMConstantIslands - Due to limited PC-relative displacements, ARM
/// requires constant pool entries to be scattered among the instructions
/// inside a function. To do this, it completely ignores the normal LLVM
/// constant pool; instead, it places constants wherever it feels like with
/// special instructions.
///
/// The terminology used in this pass includes:
/// Islands - Clumps of constants placed in the function.
/// Water - Potential places where an island could be formed.
/// CPE - A constant pool entry that has been placed somewhere, which
/// tracks a list of users.
class ARMConstantIslands : public MachineFunctionPass {
/// BBSizes - The size of each MachineBasicBlock in bytes of code, indexed
/// by MBB Number. The two-byte pads required for Thumb alignment are
/// counted as part of the following block (i.e., the offset and size for
/// a padded block will both be ==2 mod 4).
std::vector<unsigned> BBSizes;
/// BBOffsets - the offset of each MBB in bytes, starting from 0.
/// The two-byte pads required for Thumb alignment are counted as part of
/// the following block.
std::vector<unsigned> BBOffsets;
/// WaterList - A sorted list of basic blocks where islands could be placed
/// (i.e. blocks that don't fall through to the following block, due
/// to a return, unreachable, or unconditional branch).
std::vector<MachineBasicBlock*> WaterList;
/// NewWaterList - The subset of WaterList that was created since the
/// previous iteration by inserting unconditional branches.
SmallSet<MachineBasicBlock*, 4> NewWaterList;
typedef std::vector<MachineBasicBlock*>::iterator water_iterator;
/// CPUser - One user of a constant pool, keeping the machine instruction
/// pointer, the constant pool being referenced, and the max displacement
/// allowed from the instruction to the CP. The HighWaterMark records the
/// highest basic block where a new CPEntry can be placed. To ensure this
/// pass terminates, the CP entries are initially placed at the end of the
/// function and then move monotonically to lower addresses. The
/// exception to this rule is when the current CP entry for a particular
/// CPUser is out of range, but there is another CP entry for the same
/// constant value in range. We want to use the existing in-range CP
/// entry, but if it later moves out of range, the search for new water
/// should resume where it left off. The HighWaterMark is used to record
/// that point.
struct CPUser {
MachineInstr *MI;
MachineInstr *CPEMI;
MachineBasicBlock *HighWaterMark;
unsigned MaxDisp;
bool NegOk;
bool IsSoImm;
CPUser(MachineInstr *mi, MachineInstr *cpemi, unsigned maxdisp,
bool neg, bool soimm)
: MI(mi), CPEMI(cpemi), MaxDisp(maxdisp), NegOk(neg), IsSoImm(soimm) {
HighWaterMark = CPEMI->getParent();
}
};
/// CPUsers - Keep track of all of the machine instructions that use various
/// constant pools and their max displacement.
std::vector<CPUser> CPUsers;
/// CPEntry - One per constant pool entry, keeping the machine instruction
/// pointer, the constpool index, and the number of CPUser's which
/// reference this entry.
struct CPEntry {
MachineInstr *CPEMI;
unsigned CPI;
unsigned RefCount;
CPEntry(MachineInstr *cpemi, unsigned cpi, unsigned rc = 0)
: CPEMI(cpemi), CPI(cpi), RefCount(rc) {}
};
/// CPEntries - Keep track of all of the constant pool entry machine
/// instructions. For each original constpool index (i.e. those that
/// existed upon entry to this pass), it keeps a vector of entries.
/// Original elements are cloned as we go along; the clones are
/// put in the vector of the original element, but have distinct CPIs.
std::vector<std::vector<CPEntry> > CPEntries;
/// ImmBranch - One per immediate branch, keeping the machine instruction
/// pointer, conditional or unconditional, the max displacement,
/// and (if isCond is true) the corresponding unconditional branch
/// opcode.
struct ImmBranch {
MachineInstr *MI;
unsigned MaxDisp : 31;
bool isCond : 1;
int UncondBr;
ImmBranch(MachineInstr *mi, unsigned maxdisp, bool cond, int ubr)
: MI(mi), MaxDisp(maxdisp), isCond(cond), UncondBr(ubr) {}
};
/// ImmBranches - Keep track of all the immediate branch instructions.
///
std::vector<ImmBranch> ImmBranches;
/// PushPopMIs - Keep track of all the Thumb push / pop instructions.
///
SmallVector<MachineInstr*, 4> PushPopMIs;
/// T2JumpTables - Keep track of all the Thumb2 jumptable instructions.
SmallVector<MachineInstr*, 4> T2JumpTables;
/// HasFarJump - True if any far jump instruction has been emitted during
/// the branch fix up pass.
bool HasFarJump;
/// HasInlineAsm - True if the function contains inline assembly.
bool HasInlineAsm;
const ARMInstrInfo *TII;
const ARMSubtarget *STI;
ARMFunctionInfo *AFI;
bool isThumb;
bool isThumb1;
bool isThumb2;
public:
static char ID;
ARMConstantIslands() : MachineFunctionPass(ID) {}
virtual bool runOnMachineFunction(MachineFunction &MF);
virtual const char *getPassName() const {
return "ARM constant island placement and branch shortening pass";
}
private:
void DoInitialPlacement(MachineFunction &MF,
std::vector<MachineInstr*> &CPEMIs);
CPEntry *findConstPoolEntry(unsigned CPI, const MachineInstr *CPEMI);
void JumpTableFunctionScan(MachineFunction &MF);
void InitialFunctionScan(MachineFunction &MF,
const std::vector<MachineInstr*> &CPEMIs);
MachineBasicBlock *SplitBlockBeforeInstr(MachineInstr *MI);
void UpdateForInsertedWaterBlock(MachineBasicBlock *NewBB);
void AdjustBBOffsetsAfter(MachineBasicBlock *BB, int delta);
bool DecrementOldEntry(unsigned CPI, MachineInstr* CPEMI);
int LookForExistingCPEntry(CPUser& U, unsigned UserOffset);
bool LookForWater(CPUser&U, unsigned UserOffset, water_iterator &WaterIter);
void CreateNewWater(unsigned CPUserIndex, unsigned UserOffset,
MachineBasicBlock *&NewMBB);
bool HandleConstantPoolUser(MachineFunction &MF, unsigned CPUserIndex);
void RemoveDeadCPEMI(MachineInstr *CPEMI);
bool RemoveUnusedCPEntries();
bool CPEIsInRange(MachineInstr *MI, unsigned UserOffset,
MachineInstr *CPEMI, unsigned Disp, bool NegOk,
bool DoDump = false);
bool WaterIsInRange(unsigned UserOffset, MachineBasicBlock *Water,
CPUser &U);
bool OffsetIsInRange(unsigned UserOffset, unsigned TrialOffset,
unsigned Disp, bool NegativeOK, bool IsSoImm = false);
bool BBIsInRange(MachineInstr *MI, MachineBasicBlock *BB, unsigned Disp);
bool FixUpImmediateBr(MachineFunction &MF, ImmBranch &Br);
bool FixUpConditionalBr(MachineFunction &MF, ImmBranch &Br);
bool FixUpUnconditionalBr(MachineFunction &MF, ImmBranch &Br);
bool UndoLRSpillRestore();
bool OptimizeThumb2Instructions(MachineFunction &MF);
bool OptimizeThumb2Branches(MachineFunction &MF);
bool ReorderThumb2JumpTables(MachineFunction &MF);
bool OptimizeThumb2JumpTables(MachineFunction &MF);
MachineBasicBlock *AdjustJTTargetBlockForward(MachineBasicBlock *BB,
MachineBasicBlock *JTBB);
unsigned GetOffsetOf(MachineInstr *MI) const;
void dumpBBs();
void verify(MachineFunction &MF);
};
char ARMConstantIslands::ID = 0;
}
/// verify - check BBOffsets, BBSizes, alignment of islands
void ARMConstantIslands::verify(MachineFunction &MF) {
assert(BBOffsets.size() == BBSizes.size());
for (unsigned i = 1, e = BBOffsets.size(); i != e; ++i)
assert(BBOffsets[i-1]+BBSizes[i-1] == BBOffsets[i]);
if (!isThumb)
return;
#ifndef NDEBUG
for (MachineFunction::iterator MBBI = MF.begin(), E = MF.end();
MBBI != E; ++MBBI) {
MachineBasicBlock *MBB = MBBI;
if (!MBB->empty() &&
MBB->begin()->getOpcode() == ARM::CONSTPOOL_ENTRY) {
unsigned MBBId = MBB->getNumber();
assert(HasInlineAsm ||
(BBOffsets[MBBId]%4 == 0 && BBSizes[MBBId]%4 == 0) ||
(BBOffsets[MBBId]%4 != 0 && BBSizes[MBBId]%4 != 0));
}
}
for (unsigned i = 0, e = CPUsers.size(); i != e; ++i) {
CPUser &U = CPUsers[i];
unsigned UserOffset = GetOffsetOf(U.MI) + (isThumb ? 4 : 8);
unsigned CPEOffset = GetOffsetOf(U.CPEMI);
unsigned Disp = UserOffset < CPEOffset ? CPEOffset - UserOffset :
UserOffset - CPEOffset;
assert(Disp <= U.MaxDisp || "Constant pool entry out of range!");
}
#endif
}
/// print block size and offset information - debugging
void ARMConstantIslands::dumpBBs() {
for (unsigned J = 0, E = BBOffsets.size(); J !=E; ++J) {
DEBUG(errs() << "block " << J << " offset " << BBOffsets[J]
<< " size " << BBSizes[J] << "\n");
}
}
/// createARMConstantIslandPass - returns an instance of the constpool
/// island pass.
FunctionPass *llvm::createARMConstantIslandPass() {
return new ARMConstantIslands();
}
bool ARMConstantIslands::runOnMachineFunction(MachineFunction &MF) {
MachineConstantPool &MCP = *MF.getConstantPool();
TII = (const ARMInstrInfo*)MF.getTarget().getInstrInfo();
AFI = MF.getInfo<ARMFunctionInfo>();
STI = &MF.getTarget().getSubtarget<ARMSubtarget>();
isThumb = AFI->isThumbFunction();
isThumb1 = AFI->isThumb1OnlyFunction();
isThumb2 = AFI->isThumb2Function();
HasFarJump = false;
HasInlineAsm = false;
// Renumber all of the machine basic blocks in the function, guaranteeing that
// the numbers agree with the position of the block in the function.
MF.RenumberBlocks();
// Try to reorder and otherwise adjust the block layout to make good use
// of the TB[BH] instructions.
bool MadeChange = false;
if (isThumb2 && AdjustJumpTableBlocks) {
JumpTableFunctionScan(MF);
MadeChange |= ReorderThumb2JumpTables(MF);
// Data is out of date, so clear it. It'll be re-computed later.
T2JumpTables.clear();
// Blocks may have shifted around. Keep the numbering up to date.
MF.RenumberBlocks();
}
// Thumb1 functions containing constant pools get 4-byte alignment.
// This is so we can keep exact track of where the alignment padding goes.
// ARM and Thumb2 functions need to be 4-byte aligned.
if (!isThumb1)
MF.EnsureAlignment(2); // 2 = log2(4)
// Perform the initial placement of the constant pool entries. To start with,
// we put them all at the end of the function.
std::vector<MachineInstr*> CPEMIs;
if (!MCP.isEmpty()) {
DoInitialPlacement(MF, CPEMIs);
if (isThumb1)
MF.EnsureAlignment(2); // 2 = log2(4)
}
/// The next UID to take is the first unused one.
AFI->initConstPoolEntryUId(CPEMIs.size());
// Do the initial scan of the function, building up information about the
// sizes of each block, the location of all the water, and finding all of the
// constant pool users.
InitialFunctionScan(MF, CPEMIs);
CPEMIs.clear();
DEBUG(dumpBBs());
/// Remove dead constant pool entries.
RemoveUnusedCPEntries();
// Iteratively place constant pool entries and fix up branches until there
// is no change.
unsigned NoCPIters = 0, NoBRIters = 0;
while (true) {
bool CPChange = false;
for (unsigned i = 0, e = CPUsers.size(); i != e; ++i)
CPChange |= HandleConstantPoolUser(MF, i);
if (CPChange && ++NoCPIters > 30)
llvm_unreachable("Constant Island pass failed to converge!");
DEBUG(dumpBBs());
// Clear NewWaterList now. If we split a block for branches, it should
// appear as "new water" for the next iteration of constant pool placement.
NewWaterList.clear();
bool BRChange = false;
for (unsigned i = 0, e = ImmBranches.size(); i != e; ++i)
BRChange |= FixUpImmediateBr(MF, ImmBranches[i]);
if (BRChange && ++NoBRIters > 30)
llvm_unreachable("Branch Fix Up pass failed to converge!");
DEBUG(dumpBBs());
if (!CPChange && !BRChange)
break;
MadeChange = true;
}
// Shrink 32-bit Thumb2 branch, load, and store instructions.
if (isThumb2 && !STI->prefers32BitThumb())
MadeChange |= OptimizeThumb2Instructions(MF);
// After a while, this might be made debug-only, but it is not expensive.
verify(MF);
// If LR has been forced spilled and no far jump (i.e. BL) has been issued,
// undo the spill / restore of LR if possible.
if (isThumb && !HasFarJump && AFI->isLRSpilledForFarJump())
MadeChange |= UndoLRSpillRestore();
DEBUG(errs() << '\n'; dumpBBs());
BBSizes.clear();
BBOffsets.clear();
WaterList.clear();
CPUsers.clear();
CPEntries.clear();
ImmBranches.clear();
PushPopMIs.clear();
T2JumpTables.clear();
return MadeChange;
}
/// DoInitialPlacement - Perform the initial placement of the constant pool
/// entries. To start with, we put them all at the end of the function.
void ARMConstantIslands::DoInitialPlacement(MachineFunction &MF,
std::vector<MachineInstr*> &CPEMIs) {
// Create the basic block to hold the CPE's.
MachineBasicBlock *BB = MF.CreateMachineBasicBlock();
MF.push_back(BB);
// Add all of the constants from the constant pool to the end block, use an
// identity mapping of CPI's to CPE's.
const std::vector<MachineConstantPoolEntry> &CPs =
MF.getConstantPool()->getConstants();
const TargetData &TD = *MF.getTarget().getTargetData();
for (unsigned i = 0, e = CPs.size(); i != e; ++i) {
unsigned Size = TD.getTypeAllocSize(CPs[i].getType());
// Verify that all constant pool entries are a multiple of 4 bytes. If not,
// we would have to pad them out or something so that instructions stay
// aligned.
assert((Size & 3) == 0 && "CP Entry not multiple of 4 bytes!");
MachineInstr *CPEMI =
BuildMI(BB, DebugLoc(), TII->get(ARM::CONSTPOOL_ENTRY))
.addImm(i).addConstantPoolIndex(i).addImm(Size);
CPEMIs.push_back(CPEMI);
// Add a new CPEntry, but no corresponding CPUser yet.
std::vector<CPEntry> CPEs;
CPEs.push_back(CPEntry(CPEMI, i));
CPEntries.push_back(CPEs);
++NumCPEs;
DEBUG(errs() << "Moved CPI#" << i << " to end of function as #" << i
<< "\n");
}
}
/// BBHasFallthrough - Return true if the specified basic block can fallthrough
/// into the block immediately after it.
static bool BBHasFallthrough(MachineBasicBlock *MBB) {
// Get the next machine basic block in the function.
MachineFunction::iterator MBBI = MBB;
// Can't fall off end of function.
if (llvm::next(MBBI) == MBB->getParent()->end())
return false;
MachineBasicBlock *NextBB = llvm::next(MBBI);
for (MachineBasicBlock::succ_iterator I = MBB->succ_begin(),
E = MBB->succ_end(); I != E; ++I)
if (*I == NextBB)
return true;
return false;
}
/// findConstPoolEntry - Given the constpool index and CONSTPOOL_ENTRY MI,
/// look up the corresponding CPEntry.
ARMConstantIslands::CPEntry
*ARMConstantIslands::findConstPoolEntry(unsigned CPI,
const MachineInstr *CPEMI) {
std::vector<CPEntry> &CPEs = CPEntries[CPI];
// Number of entries per constpool index should be small, just do a
// linear search.
for (unsigned i = 0, e = CPEs.size(); i != e; ++i) {
if (CPEs[i].CPEMI == CPEMI)
return &CPEs[i];
}
return NULL;
}
/// JumpTableFunctionScan - Do a scan of the function, building up
/// information about the sizes of each block and the locations of all
/// the jump tables.
void ARMConstantIslands::JumpTableFunctionScan(MachineFunction &MF) {
for (MachineFunction::iterator MBBI = MF.begin(), E = MF.end();
MBBI != E; ++MBBI) {
MachineBasicBlock &MBB = *MBBI;
for (MachineBasicBlock::iterator I = MBB.begin(), E = MBB.end();
I != E; ++I)
if (I->getDesc().isBranch() && I->getOpcode() == ARM::t2BR_JT)
T2JumpTables.push_back(I);
}
}
/// InitialFunctionScan - Do the initial scan of the function, building up
/// information about the sizes of each block, the location of all the water,
/// and finding all of the constant pool users.
void ARMConstantIslands::InitialFunctionScan(MachineFunction &MF,
const std::vector<MachineInstr*> &CPEMIs) {
// First thing, see if the function has any inline assembly in it. If so,
// we have to be conservative about alignment assumptions, as we don't
// know for sure the size of any instructions in the inline assembly.
for (MachineFunction::iterator MBBI = MF.begin(), E = MF.end();
MBBI != E; ++MBBI) {
MachineBasicBlock &MBB = *MBBI;
for (MachineBasicBlock::iterator I = MBB.begin(), E = MBB.end();
I != E; ++I)
if (I->getOpcode() == ARM::INLINEASM)
HasInlineAsm = true;
}
// Now go back through the instructions and build up our data structures
unsigned Offset = 0;
for (MachineFunction::iterator MBBI = MF.begin(), E = MF.end();
MBBI != E; ++MBBI) {
MachineBasicBlock &MBB = *MBBI;
// If this block doesn't fall through into the next MBB, then this is
// 'water' that a constant pool island could be placed.
if (!BBHasFallthrough(&MBB))
WaterList.push_back(&MBB);
unsigned MBBSize = 0;
for (MachineBasicBlock::iterator I = MBB.begin(), E = MBB.end();
I != E; ++I) {
if (I->isDebugValue())
continue;
// Add instruction size to MBBSize.
MBBSize += TII->GetInstSizeInBytes(I);
int Opc = I->getOpcode();
if (I->getDesc().isBranch()) {
bool isCond = false;
unsigned Bits = 0;
unsigned Scale = 1;
int UOpc = Opc;
switch (Opc) {
default:
continue; // Ignore other JT branches
case ARM::tBR_JTr:
// A Thumb1 table jump may involve padding; for the offsets to
// be right, functions containing these must be 4-byte aligned.
// tBR_JTr expands to a mov pc followed by .align 2 and then the jump
// table entries. So this code checks whether offset of tBR_JTr + 2
// is aligned. That is held in Offset+MBBSize, which already has
// 2 added in for the size of the mov pc instruction.
MF.EnsureAlignment(2U);
if ((Offset+MBBSize)%4 != 0 || HasInlineAsm)
// FIXME: Add a pseudo ALIGN instruction instead.
MBBSize += 2; // padding
continue; // Does not get an entry in ImmBranches
case ARM::t2BR_JT:
T2JumpTables.push_back(I);
continue; // Does not get an entry in ImmBranches
case ARM::Bcc:
isCond = true;
UOpc = ARM::B;
// Fallthrough
case ARM::B:
Bits = 24;
Scale = 4;
break;
case ARM::tBcc:
isCond = true;
UOpc = ARM::tB;
Bits = 8;
Scale = 2;
break;
case ARM::tB:
Bits = 11;
Scale = 2;
break;
case ARM::t2Bcc:
isCond = true;
UOpc = ARM::t2B;
Bits = 20;
Scale = 2;
break;
case ARM::t2B:
Bits = 24;
Scale = 2;
break;
}
// Record this immediate branch.
unsigned MaxOffs = ((1 << (Bits-1))-1) * Scale;
ImmBranches.push_back(ImmBranch(I, MaxOffs, isCond, UOpc));
}
if (Opc == ARM::tPUSH || Opc == ARM::tPOP_RET)
PushPopMIs.push_back(I);
if (Opc == ARM::CONSTPOOL_ENTRY)
continue;
// Scan the instructions for constant pool operands.
for (unsigned op = 0, e = I->getNumOperands(); op != e; ++op)
if (I->getOperand(op).isCPI()) {
// We found one. The addressing mode tells us the max displacement
// from the PC that this instruction permits.
// Basic size info comes from the TSFlags field.
unsigned Bits = 0;
unsigned Scale = 1;
bool NegOk = false;
bool IsSoImm = false;
switch (Opc) {
default:
llvm_unreachable("Unknown addressing mode for CP reference!");
break;
// Taking the address of a CP entry.
case ARM::LEApcrel:
// This takes a SoImm, which is 8 bit immediate rotated. We'll
// pretend the maximum offset is 255 * 4. Since each instruction
// 4 byte wide, this is always correct. We'll check for other
// displacements that fits in a SoImm as well.
Bits = 8;
Scale = 4;
NegOk = true;
IsSoImm = true;
break;
case ARM::t2LEApcrel:
Bits = 12;
NegOk = true;
break;
case ARM::tLEApcrel:
Bits = 8;
Scale = 4;
break;
case ARM::LDR:
case ARM::LDRcp:
case ARM::t2LDRpci:
Bits = 12; // +-offset_12
NegOk = true;
break;
case ARM::tLDRpci:
case ARM::tLDRcp:
Bits = 8;
Scale = 4; // +(offset_8*4)
break;
case ARM::VLDRD:
case ARM::VLDRS:
Bits = 8;
Scale = 4; // +-(offset_8*4)
NegOk = true;
break;
}
// Remember that this is a user of a CP entry.
unsigned CPI = I->getOperand(op).getIndex();
MachineInstr *CPEMI = CPEMIs[CPI];
unsigned MaxOffs = ((1 << Bits)-1) * Scale;
CPUsers.push_back(CPUser(I, CPEMI, MaxOffs, NegOk, IsSoImm));
// Increment corresponding CPEntry reference count.
CPEntry *CPE = findConstPoolEntry(CPI, CPEMI);
assert(CPE && "Cannot find a corresponding CPEntry!");
CPE->RefCount++;
// Instructions can only use one CP entry, don't bother scanning the
// rest of the operands.
break;
}
}
// In thumb mode, if this block is a constpool island, we may need padding
// so it's aligned on 4 byte boundary.
if (isThumb &&
!MBB.empty() &&
MBB.begin()->getOpcode() == ARM::CONSTPOOL_ENTRY &&
((Offset%4) != 0 || HasInlineAsm))
MBBSize += 2;
BBSizes.push_back(MBBSize);
BBOffsets.push_back(Offset);
Offset += MBBSize;
}
}
/// GetOffsetOf - Return the current offset of the specified machine instruction
/// from the start of the function. This offset changes as stuff is moved
/// around inside the function.
unsigned ARMConstantIslands::GetOffsetOf(MachineInstr *MI) const {
MachineBasicBlock *MBB = MI->getParent();
// The offset is composed of two things: the sum of the sizes of all MBB's
// before this instruction's block, and the offset from the start of the block
// it is in.
unsigned Offset = BBOffsets[MBB->getNumber()];
// If we're looking for a CONSTPOOL_ENTRY in Thumb, see if this block has
// alignment padding, and compensate if so.
if (isThumb &&
MI->getOpcode() == ARM::CONSTPOOL_ENTRY &&
(Offset%4 != 0 || HasInlineAsm))
Offset += 2;
// Sum instructions before MI in MBB.
for (MachineBasicBlock::iterator I = MBB->begin(); ; ++I) {
assert(I != MBB->end() && "Didn't find MI in its own basic block?");
if (&*I == MI) return Offset;
Offset += TII->GetInstSizeInBytes(I);
}
}
/// CompareMBBNumbers - Little predicate function to sort the WaterList by MBB
/// ID.
static bool CompareMBBNumbers(const MachineBasicBlock *LHS,
const MachineBasicBlock *RHS) {
return LHS->getNumber() < RHS->getNumber();
}
/// UpdateForInsertedWaterBlock - When a block is newly inserted into the
/// machine function, it upsets all of the block numbers. Renumber the blocks
/// and update the arrays that parallel this numbering.
void ARMConstantIslands::UpdateForInsertedWaterBlock(MachineBasicBlock *NewBB) {
// Renumber the MBB's to keep them consequtive.
NewBB->getParent()->RenumberBlocks(NewBB);
// Insert a size into BBSizes to align it properly with the (newly
// renumbered) block numbers.
BBSizes.insert(BBSizes.begin()+NewBB->getNumber(), 0);
// Likewise for BBOffsets.
BBOffsets.insert(BBOffsets.begin()+NewBB->getNumber(), 0);
// Next, update WaterList. Specifically, we need to add NewMBB as having
// available water after it.
water_iterator IP =
std::lower_bound(WaterList.begin(), WaterList.end(), NewBB,
CompareMBBNumbers);
WaterList.insert(IP, NewBB);
}
/// Split the basic block containing MI into two blocks, which are joined by
/// an unconditional branch. Update data structures and renumber blocks to
/// account for this change and returns the newly created block.
MachineBasicBlock *ARMConstantIslands::SplitBlockBeforeInstr(MachineInstr *MI) {
MachineBasicBlock *OrigBB = MI->getParent();
MachineFunction &MF = *OrigBB->getParent();
// Create a new MBB for the code after the OrigBB.
MachineBasicBlock *NewBB =
MF.CreateMachineBasicBlock(OrigBB->getBasicBlock());
MachineFunction::iterator MBBI = OrigBB; ++MBBI;
MF.insert(MBBI, NewBB);
// Splice the instructions starting with MI over to NewBB.
NewBB->splice(NewBB->end(), OrigBB, MI, OrigBB->end());
// Add an unconditional branch from OrigBB to NewBB.
// Note the new unconditional branch is not being recorded.
// There doesn't seem to be meaningful DebugInfo available; this doesn't
// correspond to anything in the source.
unsigned Opc = isThumb ? (isThumb2 ? ARM::t2B : ARM::tB) : ARM::B;
BuildMI(OrigBB, DebugLoc(), TII->get(Opc)).addMBB(NewBB);
++NumSplit;
// Update the CFG. All succs of OrigBB are now succs of NewBB.
while (!OrigBB->succ_empty()) {
MachineBasicBlock *Succ = *OrigBB->succ_begin();
OrigBB->removeSuccessor(Succ);
NewBB->addSuccessor(Succ);
// This pass should be run after register allocation, so there should be no
// PHI nodes to update.
assert((Succ->empty() || !Succ->begin()->isPHI())
&& "PHI nodes should be eliminated by now!");
}
// OrigBB branches to NewBB.
OrigBB->addSuccessor(NewBB);
// Update internal data structures to account for the newly inserted MBB.
// This is almost the same as UpdateForInsertedWaterBlock, except that
// the Water goes after OrigBB, not NewBB.
MF.RenumberBlocks(NewBB);
// Insert a size into BBSizes to align it properly with the (newly
// renumbered) block numbers.
BBSizes.insert(BBSizes.begin()+NewBB->getNumber(), 0);
// Likewise for BBOffsets.
BBOffsets.insert(BBOffsets.begin()+NewBB->getNumber(), 0);
// Next, update WaterList. Specifically, we need to add OrigMBB as having
// available water after it (but not if it's already there, which happens
// when splitting before a conditional branch that is followed by an
// unconditional branch - in that case we want to insert NewBB).
water_iterator IP =
std::lower_bound(WaterList.begin(), WaterList.end(), OrigBB,
CompareMBBNumbers);
MachineBasicBlock* WaterBB = *IP;
if (WaterBB == OrigBB)
WaterList.insert(llvm::next(IP), NewBB);
else
WaterList.insert(IP, OrigBB);
NewWaterList.insert(OrigBB);
unsigned OrigBBI = OrigBB->getNumber();
unsigned NewBBI = NewBB->getNumber();
int delta = isThumb1 ? 2 : 4;
// Figure out how large the OrigBB is. As the first half of the original
// block, it cannot contain a tablejump. The size includes
// the new jump we added. (It should be possible to do this without
// recounting everything, but it's very confusing, and this is rarely
// executed.)
unsigned OrigBBSize = 0;
for (MachineBasicBlock::iterator I = OrigBB->begin(), E = OrigBB->end();
I != E; ++I)
OrigBBSize += TII->GetInstSizeInBytes(I);
BBSizes[OrigBBI] = OrigBBSize;
// ...and adjust BBOffsets for NewBB accordingly.
BBOffsets[NewBBI] = BBOffsets[OrigBBI] + BBSizes[OrigBBI];
// Figure out how large the NewMBB is. As the second half of the original
// block, it may contain a tablejump.
unsigned NewBBSize = 0;
for (MachineBasicBlock::iterator I = NewBB->begin(), E = NewBB->end();
I != E; ++I)
NewBBSize += TII->GetInstSizeInBytes(I);
// Set the size of NewBB in BBSizes. It does not include any padding now.
BBSizes[NewBBI] = NewBBSize;
MachineInstr* ThumbJTMI = prior(NewBB->end());
if (ThumbJTMI->getOpcode() == ARM::tBR_JTr) {
// We've added another 2-byte instruction before this tablejump, which
// means we will always need padding if we didn't before, and vice versa.
// The original offset of the jump instruction was:
unsigned OrigOffset = BBOffsets[OrigBBI] + BBSizes[OrigBBI] - delta;
if (OrigOffset%4 == 0) {
// We had padding before and now we don't. No net change in code size.
delta = 0;
} else {
// We didn't have padding before and now we do.
BBSizes[NewBBI] += 2;
delta = 4;
}
}
// All BBOffsets following these blocks must be modified.
if (delta)
AdjustBBOffsetsAfter(NewBB, delta);
return NewBB;
}
/// OffsetIsInRange - Checks whether UserOffset (the location of a constant pool
/// reference) is within MaxDisp of TrialOffset (a proposed location of a
/// constant pool entry).
bool ARMConstantIslands::OffsetIsInRange(unsigned UserOffset,
unsigned TrialOffset, unsigned MaxDisp,
bool NegativeOK, bool IsSoImm) {
// On Thumb offsets==2 mod 4 are rounded down by the hardware for
// purposes of the displacement computation; compensate for that here.
// Effectively, the valid range of displacements is 2 bytes smaller for such
// references.
unsigned TotalAdj = 0;
if (isThumb && UserOffset%4 !=0) {
UserOffset -= 2;
TotalAdj = 2;
}
// CPEs will be rounded up to a multiple of 4.
if (isThumb && TrialOffset%4 != 0) {
TrialOffset += 2;
TotalAdj += 2;
}
// In Thumb2 mode, later branch adjustments can shift instructions up and
// cause alignment change. In the worst case scenario this can cause the
// user's effective address to be subtracted by 2 and the CPE's address to
// be plus 2.
if (isThumb2 && TotalAdj != 4)
MaxDisp -= (4 - TotalAdj);
if (UserOffset <= TrialOffset) {
// User before the Trial.
if (TrialOffset - UserOffset <= MaxDisp)
return true;
// FIXME: Make use full range of soimm values.
} else if (NegativeOK) {
if (UserOffset - TrialOffset <= MaxDisp)
return true;
// FIXME: Make use full range of soimm values.
}
return false;
}
/// WaterIsInRange - Returns true if a CPE placed after the specified
/// Water (a basic block) will be in range for the specific MI.
bool ARMConstantIslands::WaterIsInRange(unsigned UserOffset,
MachineBasicBlock* Water, CPUser &U) {
unsigned MaxDisp = U.MaxDisp;
unsigned CPEOffset = BBOffsets[Water->getNumber()] +
BBSizes[Water->getNumber()];
// If the CPE is to be inserted before the instruction, that will raise
// the offset of the instruction.
if (CPEOffset < UserOffset)
UserOffset += U.CPEMI->getOperand(2).getImm();
return OffsetIsInRange(UserOffset, CPEOffset, MaxDisp, U.NegOk, U.IsSoImm);
}
/// CPEIsInRange - Returns true if the distance between specific MI and
/// specific ConstPool entry instruction can fit in MI's displacement field.
bool ARMConstantIslands::CPEIsInRange(MachineInstr *MI, unsigned UserOffset,
MachineInstr *CPEMI, unsigned MaxDisp,
bool NegOk, bool DoDump) {
unsigned CPEOffset = GetOffsetOf(CPEMI);
assert((CPEOffset%4 == 0 || HasInlineAsm) && "Misaligned CPE");
if (DoDump) {
DEBUG(errs() << "User of CPE#" << CPEMI->getOperand(0).getImm()
<< " max delta=" << MaxDisp
<< " insn address=" << UserOffset
<< " CPE address=" << CPEOffset
<< " offset=" << int(CPEOffset-UserOffset) << "\t" << *MI);
}
return OffsetIsInRange(UserOffset, CPEOffset, MaxDisp, NegOk);
}
#ifndef NDEBUG
/// BBIsJumpedOver - Return true of the specified basic block's only predecessor
/// unconditionally branches to its only successor.
static bool BBIsJumpedOver(MachineBasicBlock *MBB) {
if (MBB->pred_size() != 1 || MBB->succ_size() != 1)
return false;
MachineBasicBlock *Succ = *MBB->succ_begin();
MachineBasicBlock *Pred = *MBB->pred_begin();
MachineInstr *PredMI = &Pred->back();
if (PredMI->getOpcode() == ARM::B || PredMI->getOpcode() == ARM::tB
|| PredMI->getOpcode() == ARM::t2B)
return PredMI->getOperand(0).getMBB() == Succ;
return false;
}
#endif // NDEBUG
void ARMConstantIslands::AdjustBBOffsetsAfter(MachineBasicBlock *BB,
int delta) {
MachineFunction::iterator MBBI = BB; MBBI = llvm::next(MBBI);
for(unsigned i = BB->getNumber()+1, e = BB->getParent()->getNumBlockIDs();
i < e; ++i) {
BBOffsets[i] += delta;
// If some existing blocks have padding, adjust the padding as needed, a
// bit tricky. delta can be negative so don't use % on that.
if (!isThumb)
continue;
MachineBasicBlock *MBB = MBBI;
if (!MBB->empty() && !HasInlineAsm) {
// Constant pool entries require padding.
if (MBB->begin()->getOpcode() == ARM::CONSTPOOL_ENTRY) {
unsigned OldOffset = BBOffsets[i] - delta;
if ((OldOffset%4) == 0 && (BBOffsets[i]%4) != 0) {
// add new padding
BBSizes[i] += 2;
delta += 2;
} else if ((OldOffset%4) != 0 && (BBOffsets[i]%4) == 0) {
// remove existing padding
BBSizes[i] -= 2;
delta -= 2;
}
}
// Thumb1 jump tables require padding. They should be at the end;
// following unconditional branches are removed by AnalyzeBranch.
// tBR_JTr expands to a mov pc followed by .align 2 and then the jump
// table entries. So this code checks whether offset of tBR_JTr
// is aligned; if it is, the offset of the jump table following the
// instruction will not be aligned, and we need padding.
MachineInstr *ThumbJTMI = prior(MBB->end());
if (ThumbJTMI->getOpcode() == ARM::tBR_JTr) {
unsigned NewMIOffset = GetOffsetOf(ThumbJTMI);
unsigned OldMIOffset = NewMIOffset - delta;
if ((OldMIOffset%4) == 0 && (NewMIOffset%4) != 0) {
// remove existing padding
BBSizes[i] -= 2;
delta -= 2;
} else if ((OldMIOffset%4) != 0 && (NewMIOffset%4) == 0) {
// add new padding
BBSizes[i] += 2;
delta += 2;
}
}
if (delta==0)
return;
}
MBBI = llvm::next(MBBI);
}
}
/// DecrementOldEntry - find the constant pool entry with index CPI
/// and instruction CPEMI, and decrement its refcount. If the refcount
/// becomes 0 remove the entry and instruction. Returns true if we removed
/// the entry, false if we didn't.
bool ARMConstantIslands::DecrementOldEntry(unsigned CPI, MachineInstr *CPEMI) {
// Find the old entry. Eliminate it if it is no longer used.
CPEntry *CPE = findConstPoolEntry(CPI, CPEMI);
assert(CPE && "Unexpected!");
if (--CPE->RefCount == 0) {
RemoveDeadCPEMI(CPEMI);
CPE->CPEMI = NULL;
--NumCPEs;
return true;
}
return false;
}
/// LookForCPEntryInRange - see if the currently referenced CPE is in range;
/// if not, see if an in-range clone of the CPE is in range, and if so,
/// change the data structures so the user references the clone. Returns:
/// 0 = no existing entry found
/// 1 = entry found, and there were no code insertions or deletions
/// 2 = entry found, and there were code insertions or deletions
int ARMConstantIslands::LookForExistingCPEntry(CPUser& U, unsigned UserOffset)
{
MachineInstr *UserMI = U.MI;
MachineInstr *CPEMI = U.CPEMI;
// Check to see if the CPE is already in-range.
if (CPEIsInRange(UserMI, UserOffset, CPEMI, U.MaxDisp, U.NegOk, true)) {
DEBUG(errs() << "In range\n");
return 1;
}
// No. Look for previously created clones of the CPE that are in range.
unsigned CPI = CPEMI->getOperand(1).getIndex();
std::vector<CPEntry> &CPEs = CPEntries[CPI];
for (unsigned i = 0, e = CPEs.size(); i != e; ++i) {
// We already tried this one
if (CPEs[i].CPEMI == CPEMI)
continue;
// Removing CPEs can leave empty entries, skip
if (CPEs[i].CPEMI == NULL)
continue;
if (CPEIsInRange(UserMI, UserOffset, CPEs[i].CPEMI, U.MaxDisp, U.NegOk)) {
DEBUG(errs() << "Replacing CPE#" << CPI << " with CPE#"
<< CPEs[i].CPI << "\n");
// Point the CPUser node to the replacement
U.CPEMI = CPEs[i].CPEMI;
// Change the CPI in the instruction operand to refer to the clone.
for (unsigned j = 0, e = UserMI->getNumOperands(); j != e; ++j)
if (UserMI->getOperand(j).isCPI()) {
UserMI->getOperand(j).setIndex(CPEs[i].CPI);
break;
}
// Adjust the refcount of the clone...
CPEs[i].RefCount++;
// ...and the original. If we didn't remove the old entry, none of the
// addresses changed, so we don't need another pass.
return DecrementOldEntry(CPI, CPEMI) ? 2 : 1;
}
}
return 0;
}
/// getUnconditionalBrDisp - Returns the maximum displacement that can fit in
/// the specific unconditional branch instruction.
static inline unsigned getUnconditionalBrDisp(int Opc) {
switch (Opc) {
case ARM::tB:
return ((1<<10)-1)*2;
case ARM::t2B:
return ((1<<23)-1)*2;
default:
break;
}
return ((1<<23)-1)*4;
}
/// LookForWater - Look for an existing entry in the WaterList in which
/// we can place the CPE referenced from U so it's within range of U's MI.
/// Returns true if found, false if not. If it returns true, WaterIter
Last week, ARMConstantIslandPass was failing to converge for the MultiSource/Benchmarks/MiBench/automotive-susan test. The failure has since been masked by an unrelated change (just randomly), so I don't have a testcase for this now. Radar 7291928. The situation where this happened is that a constant pool entry (CPE) was placed at a lower address than the load that referenced it. There were in fact 2 CPEs placed at adjacent addresses and referenced by 2 loads that were close together in the code. The distance from the loads to the CPEs was right at the limit of what they could handle, so that only one of the CPEs could be placed within range. On every iteration, the first CPE was found to be out of range, causing a new CPE to be inserted. The second CPE had been in range but the newly inserted entry pushed it too far away. Thus the second CPE was also replaced by a new entry, which in turn pushed the first CPE out of range. Etc. Judging from some comments in the code, the initial implementation of this pass did not support CPEs placed _before_ their references. In the case where the CPE is placed at a higher address, the key to making the algorithm terminate is that new CPEs are only inserted at the end of a group of adjacent CPEs. This is implemented by removing a basic block from the "WaterList" once it has been used, and then adding the newly inserted CPE block to the list so that the next insertion will come after it. This avoids the ping-pong effect where CPEs are repeatedly moved to the beginning of a group of adjacent CPEs. This does not work when going backwards, however, because the entries at the end of an adjacent group of CPEs are closer than the CPEs earlier in the group. To make this pass terminate, we need to maintain a property that changes can only happen in some sort of monotonic fashion. The fix used here is to require that the CPE for a particular constant pool load can only move to lower addresses. This is a very simple change to the code and should not cause any significant degradation in the results. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83902 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-12 21:23:15 +00:00
/// is set to the WaterList entry. For Thumb, prefer water that will not
/// introduce padding to water that will. To ensure that this pass
/// terminates, the CPE location for a particular CPUser is only allowed to
/// move to a lower address, so search backward from the end of the list and
/// prefer the first water that is in range.
bool ARMConstantIslands::LookForWater(CPUser &U, unsigned UserOffset,
water_iterator &WaterIter) {
if (WaterList.empty())
return false;
bool FoundWaterThatWouldPad = false;
water_iterator IPThatWouldPad;
for (water_iterator IP = prior(WaterList.end()),
B = WaterList.begin();; --IP) {
MachineBasicBlock* WaterBB = *IP;
// Check if water is in range and is either at a lower address than the
// current "high water mark" or a new water block that was created since
// the previous iteration by inserting an unconditional branch. In the
// latter case, we want to allow resetting the high water mark back to
// this new water since we haven't seen it before. Inserting branches
// should be relatively uncommon and when it does happen, we want to be
// sure to take advantage of it for all the CPEs near that block, so that
// we don't insert more branches than necessary.
if (WaterIsInRange(UserOffset, WaterBB, U) &&
(WaterBB->getNumber() < U.HighWaterMark->getNumber() ||
NewWaterList.count(WaterBB))) {
unsigned WBBId = WaterBB->getNumber();
if (isThumb &&
(BBOffsets[WBBId] + BBSizes[WBBId])%4 != 0) {
// This is valid Water, but would introduce padding. Remember
// it in case we don't find any Water that doesn't do this.
if (!FoundWaterThatWouldPad) {
FoundWaterThatWouldPad = true;
IPThatWouldPad = IP;
}
} else {
WaterIter = IP;
return true;
}
}
if (IP == B)
break;
}
if (FoundWaterThatWouldPad) {
WaterIter = IPThatWouldPad;
return true;
}
return false;
}
/// CreateNewWater - No existing WaterList entry will work for
/// CPUsers[CPUserIndex], so create a place to put the CPE. The end of the
/// block is used if in range, and the conditional branch munged so control
/// flow is correct. Otherwise the block is split to create a hole with an
/// unconditional branch around it. In either case NewMBB is set to a
/// block following which the new island can be inserted (the WaterList
/// is not adjusted).
void ARMConstantIslands::CreateNewWater(unsigned CPUserIndex,
unsigned UserOffset,
MachineBasicBlock *&NewMBB) {
CPUser &U = CPUsers[CPUserIndex];
MachineInstr *UserMI = U.MI;
MachineInstr *CPEMI = U.CPEMI;
MachineBasicBlock *UserMBB = UserMI->getParent();
unsigned OffsetOfNextBlock = BBOffsets[UserMBB->getNumber()] +
BBSizes[UserMBB->getNumber()];
assert(OffsetOfNextBlock== BBOffsets[UserMBB->getNumber()+1]);
Fix another problem with ARM constant pools. Radar 7303551. When ARMConstantIslandPass cannot find any good locations (i.e., "water") to place constants, it falls back to inserting unconditional branches to make a place to put them. My recent change exposed a problem in this area. We may sometimes append to the same block more than one unconditional branch. The symptoms of this are that the generated assembly has a branch to an undefined label and running llc with -debug will cause a seg fault. This happens more easily since my change to prevent CPEs from moving from lower to higher addresses as the algorithm iterates, but it could have happened before. The end of the block may be in range for various constant pool references, but the insertion point for new CPEs is not right at the end of the block -- it is at the end of the CPEs that have already been placed at the end of the block. The insertion point could be out of range. When that happens, the fallback code will always append another unconditional branch if the end of the block is in range. The fix is to only append an unconditional branch if the block does not already end with one. I also removed a check to see if the constant pool load instruction is at the end of the block, since that is redundant with checking if the end of the block is in-range. There is more to be done here, but I think this fixes the immediate problem. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84172 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-15 05:10:36 +00:00
// If the block does not end in an unconditional branch already, and if the
// end of the block is within range, make new water there. (The addition
// below is for the unconditional branch we will be adding: 4 bytes on ARM +
// Thumb2, 2 on Thumb1. Possible Thumb1 alignment padding is allowed for
// inside OffsetIsInRange.
Fix another problem with ARM constant pools. Radar 7303551. When ARMConstantIslandPass cannot find any good locations (i.e., "water") to place constants, it falls back to inserting unconditional branches to make a place to put them. My recent change exposed a problem in this area. We may sometimes append to the same block more than one unconditional branch. The symptoms of this are that the generated assembly has a branch to an undefined label and running llc with -debug will cause a seg fault. This happens more easily since my change to prevent CPEs from moving from lower to higher addresses as the algorithm iterates, but it could have happened before. The end of the block may be in range for various constant pool references, but the insertion point for new CPEs is not right at the end of the block -- it is at the end of the CPEs that have already been placed at the end of the block. The insertion point could be out of range. When that happens, the fallback code will always append another unconditional branch if the end of the block is in range. The fix is to only append an unconditional branch if the block does not already end with one. I also removed a check to see if the constant pool load instruction is at the end of the block, since that is redundant with checking if the end of the block is in-range. There is more to be done here, but I think this fixes the immediate problem. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@84172 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-15 05:10:36 +00:00
if (BBHasFallthrough(UserMBB) &&
OffsetIsInRange(UserOffset, OffsetOfNextBlock + (isThumb1 ? 2: 4),
U.MaxDisp, U.NegOk, U.IsSoImm)) {
DEBUG(errs() << "Split at end of block\n");
if (&UserMBB->back() == UserMI)
assert(BBHasFallthrough(UserMBB) && "Expected a fallthrough BB!");
NewMBB = llvm::next(MachineFunction::iterator(UserMBB));
// Add an unconditional branch from UserMBB to fallthrough block.
// Record it for branch lengthening; this new branch will not get out of
// range, but if the preceding conditional branch is out of range, the
// targets will be exchanged, and the altered branch may be out of
// range, so the machinery has to know about it.
int UncondBr = isThumb ? ((isThumb2) ? ARM::t2B : ARM::tB) : ARM::B;
BuildMI(UserMBB, DebugLoc(), TII->get(UncondBr)).addMBB(NewMBB);
unsigned MaxDisp = getUnconditionalBrDisp(UncondBr);
ImmBranches.push_back(ImmBranch(&UserMBB->back(),
MaxDisp, false, UncondBr));
int delta = isThumb1 ? 2 : 4;
BBSizes[UserMBB->getNumber()] += delta;
AdjustBBOffsetsAfter(UserMBB, delta);
} else {
// What a big block. Find a place within the block to split it.
// This is a little tricky on Thumb1 since instructions are 2 bytes
// and constant pool entries are 4 bytes: if instruction I references
// island CPE, and instruction I+1 references CPE', it will
// not work well to put CPE as far forward as possible, since then
// CPE' cannot immediately follow it (that location is 2 bytes
// farther away from I+1 than CPE was from I) and we'd need to create
// a new island. So, we make a first guess, then walk through the
// instructions between the one currently being looked at and the
// possible insertion point, and make sure any other instructions
// that reference CPEs will be able to use the same island area;
// if not, we back up the insertion point.
// The 4 in the following is for the unconditional branch we'll be
// inserting (allows for long branch on Thumb1). Alignment of the
// island is handled inside OffsetIsInRange.
unsigned BaseInsertOffset = UserOffset + U.MaxDisp -4;
// This could point off the end of the block if we've already got
// constant pool entries following this block; only the last one is
// in the water list. Back past any possible branches (allow for a
// conditional and a maximally long unconditional).
if (BaseInsertOffset >= BBOffsets[UserMBB->getNumber()+1])
BaseInsertOffset = BBOffsets[UserMBB->getNumber()+1] -
(isThumb1 ? 6 : 8);
unsigned EndInsertOffset = BaseInsertOffset +
CPEMI->getOperand(2).getImm();
MachineBasicBlock::iterator MI = UserMI;
++MI;
unsigned CPUIndex = CPUserIndex+1;
unsigned NumCPUsers = CPUsers.size();
MachineInstr *LastIT = 0;
for (unsigned Offset = UserOffset+TII->GetInstSizeInBytes(UserMI);
Offset < BaseInsertOffset;
Offset += TII->GetInstSizeInBytes(MI),
MI = llvm::next(MI)) {
if (CPUIndex < NumCPUsers && CPUsers[CPUIndex].MI == MI) {
CPUser &U = CPUsers[CPUIndex];
if (!OffsetIsInRange(Offset, EndInsertOffset,
U.MaxDisp, U.NegOk, U.IsSoImm)) {
BaseInsertOffset -= (isThumb1 ? 2 : 4);
EndInsertOffset -= (isThumb1 ? 2 : 4);
}
// This is overly conservative, as we don't account for CPEMIs
// being reused within the block, but it doesn't matter much.
EndInsertOffset += CPUsers[CPUIndex].CPEMI->getOperand(2).getImm();
CPUIndex++;
}
// Remember the last IT instruction.
if (MI->getOpcode() == ARM::t2IT)
LastIT = MI;
}
DEBUG(errs() << "Split in middle of big block\n");
--MI;
// Avoid splitting an IT block.
if (LastIT) {
unsigned PredReg = 0;
ARMCC::CondCodes CC = llvm::getITInstrPredicate(MI, PredReg);
if (CC != ARMCC::AL)
MI = LastIT;
}
NewMBB = SplitBlockBeforeInstr(MI);
}
}
/// HandleConstantPoolUser - Analyze the specified user, checking to see if it
/// is out-of-range. If so, pick up the constant pool value and move it some
/// place in-range. Return true if we changed any addresses (thus must run
/// another pass of branch lengthening), false otherwise.
bool ARMConstantIslands::HandleConstantPoolUser(MachineFunction &MF,
unsigned CPUserIndex) {
CPUser &U = CPUsers[CPUserIndex];
MachineInstr *UserMI = U.MI;
MachineInstr *CPEMI = U.CPEMI;
unsigned CPI = CPEMI->getOperand(1).getIndex();
unsigned Size = CPEMI->getOperand(2).getImm();
// Compute this only once, it's expensive. The 4 or 8 is the value the
// hardware keeps in the PC.
unsigned UserOffset = GetOffsetOf(UserMI) + (isThumb ? 4 : 8);
// See if the current entry is within range, or there is a clone of it
// in range.
int result = LookForExistingCPEntry(U, UserOffset);
if (result==1) return false;
else if (result==2) return true;
// No existing clone of this CPE is within range.
// We will be generating a new clone. Get a UID for it.
unsigned ID = AFI->createConstPoolEntryUId();
Last week, ARMConstantIslandPass was failing to converge for the MultiSource/Benchmarks/MiBench/automotive-susan test. The failure has since been masked by an unrelated change (just randomly), so I don't have a testcase for this now. Radar 7291928. The situation where this happened is that a constant pool entry (CPE) was placed at a lower address than the load that referenced it. There were in fact 2 CPEs placed at adjacent addresses and referenced by 2 loads that were close together in the code. The distance from the loads to the CPEs was right at the limit of what they could handle, so that only one of the CPEs could be placed within range. On every iteration, the first CPE was found to be out of range, causing a new CPE to be inserted. The second CPE had been in range but the newly inserted entry pushed it too far away. Thus the second CPE was also replaced by a new entry, which in turn pushed the first CPE out of range. Etc. Judging from some comments in the code, the initial implementation of this pass did not support CPEs placed _before_ their references. In the case where the CPE is placed at a higher address, the key to making the algorithm terminate is that new CPEs are only inserted at the end of a group of adjacent CPEs. This is implemented by removing a basic block from the "WaterList" once it has been used, and then adding the newly inserted CPE block to the list so that the next insertion will come after it. This avoids the ping-pong effect where CPEs are repeatedly moved to the beginning of a group of adjacent CPEs. This does not work when going backwards, however, because the entries at the end of an adjacent group of CPEs are closer than the CPEs earlier in the group. To make this pass terminate, we need to maintain a property that changes can only happen in some sort of monotonic fashion. The fix used here is to require that the CPE for a particular constant pool load can only move to lower addresses. This is a very simple change to the code and should not cause any significant degradation in the results. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@83902 91177308-0d34-0410-b5e6-96231b3b80d8
2009-10-12 21:23:15 +00:00
// Look for water where we can place this CPE.
MachineBasicBlock *NewIsland = MF.CreateMachineBasicBlock();
MachineBasicBlock *NewMBB;
water_iterator IP;
if (LookForWater(U, UserOffset, IP)) {
DEBUG(errs() << "found water in range\n");
MachineBasicBlock *WaterBB = *IP;
// If the original WaterList entry was "new water" on this iteration,
// propagate that to the new island. This is just keeping NewWaterList
// updated to match the WaterList, which will be updated below.
if (NewWaterList.count(WaterBB)) {
NewWaterList.erase(WaterBB);
NewWaterList.insert(NewIsland);
}
// The new CPE goes before the following block (NewMBB).
NewMBB = llvm::next(MachineFunction::iterator(WaterBB));
} else {
// No water found.
DEBUG(errs() << "No water found\n");
CreateNewWater(CPUserIndex, UserOffset, NewMBB);
// SplitBlockBeforeInstr adds to WaterList, which is important when it is
// called while handling branches so that the water will be seen on the
// next iteration for constant pools, but in this context, we don't want
// it. Check for this so it will be removed from the WaterList.
// Also remove any entry from NewWaterList.
MachineBasicBlock *WaterBB = prior(MachineFunction::iterator(NewMBB));
IP = std::find(WaterList.begin(), WaterList.end(), WaterBB);
if (IP != WaterList.end())
NewWaterList.erase(WaterBB);
// We are adding new water. Update NewWaterList.
NewWaterList.insert(NewIsland);
}
// Remove the original WaterList entry; we want subsequent insertions in
// this vicinity to go after the one we're about to insert. This
// considerably reduces the number of times we have to move the same CPE
// more than once and is also important to ensure the algorithm terminates.
if (IP != WaterList.end())
WaterList.erase(IP);
// Okay, we know we can put an island before NewMBB now, do it!
MF.insert(NewMBB, NewIsland);
// Update internal data structures to account for the newly inserted MBB.
UpdateForInsertedWaterBlock(NewIsland);
// Decrement the old entry, and remove it if refcount becomes 0.
DecrementOldEntry(CPI, CPEMI);
// Now that we have an island to add the CPE to, clone the original CPE and
// add it to the island.
U.HighWaterMark = NewIsland;
U.CPEMI = BuildMI(NewIsland, DebugLoc(), TII->get(ARM::CONSTPOOL_ENTRY))
.addImm(ID).addConstantPoolIndex(CPI).addImm(Size);
CPEntries[CPI].push_back(CPEntry(U.CPEMI, ID, 1));
++NumCPEs;
BBOffsets[NewIsland->getNumber()] = BBOffsets[NewMBB->getNumber()];
// Compensate for .align 2 in thumb mode.
if (isThumb && (BBOffsets[NewIsland->getNumber()]%4 != 0 || HasInlineAsm))
Size += 2;
// Increase the size of the island block to account for the new entry.
BBSizes[NewIsland->getNumber()] += Size;
AdjustBBOffsetsAfter(NewIsland, Size);
// Finally, change the CPI in the instruction operand to be ID.
for (unsigned i = 0, e = UserMI->getNumOperands(); i != e; ++i)
if (UserMI->getOperand(i).isCPI()) {
UserMI->getOperand(i).setIndex(ID);
break;
}
DEBUG(errs() << " Moved CPE to #" << ID << " CPI=" << CPI
<< '\t' << *UserMI);
return true;
}
/// RemoveDeadCPEMI - Remove a dead constant pool entry instruction. Update
/// sizes and offsets of impacted basic blocks.
void ARMConstantIslands::RemoveDeadCPEMI(MachineInstr *CPEMI) {
MachineBasicBlock *CPEBB = CPEMI->getParent();
unsigned Size = CPEMI->getOperand(2).getImm();
CPEMI->eraseFromParent();
BBSizes[CPEBB->getNumber()] -= Size;
// All succeeding offsets have the current size value added in, fix this.
if (CPEBB->empty()) {
// In thumb1 mode, the size of island may be padded by two to compensate for
// the alignment requirement. Then it will now be 2 when the block is
// empty, so fix this.
// All succeeding offsets have the current size value added in, fix this.
if (BBSizes[CPEBB->getNumber()] != 0) {
Size += BBSizes[CPEBB->getNumber()];
BBSizes[CPEBB->getNumber()] = 0;
}
}
AdjustBBOffsetsAfter(CPEBB, -Size);
// An island has only one predecessor BB and one successor BB. Check if
// this BB's predecessor jumps directly to this BB's successor. This
// shouldn't happen currently.
assert(!BBIsJumpedOver(CPEBB) && "How did this happen?");
// FIXME: remove the empty blocks after all the work is done?
}
/// RemoveUnusedCPEntries - Remove constant pool entries whose refcounts
/// are zero.
bool ARMConstantIslands::RemoveUnusedCPEntries() {
unsigned MadeChange = false;
for (unsigned i = 0, e = CPEntries.size(); i != e; ++i) {
std::vector<CPEntry> &CPEs = CPEntries[i];
for (unsigned j = 0, ee = CPEs.size(); j != ee; ++j) {
if (CPEs[j].RefCount == 0 && CPEs[j].CPEMI) {
RemoveDeadCPEMI(CPEs[j].CPEMI);
CPEs[j].CPEMI = NULL;
MadeChange = true;
}
}
}
return MadeChange;
}
/// BBIsInRange - Returns true if the distance between specific MI and
/// specific BB can fit in MI's displacement field.
bool ARMConstantIslands::BBIsInRange(MachineInstr *MI,MachineBasicBlock *DestBB,
unsigned MaxDisp) {
unsigned PCAdj = isThumb ? 4 : 8;
unsigned BrOffset = GetOffsetOf(MI) + PCAdj;
unsigned DestOffset = BBOffsets[DestBB->getNumber()];
DEBUG(errs() << "Branch of destination BB#" << DestBB->getNumber()
<< " from BB#" << MI->getParent()->getNumber()
<< " max delta=" << MaxDisp
<< " from " << GetOffsetOf(MI) << " to " << DestOffset
<< " offset " << int(DestOffset-BrOffset) << "\t" << *MI);
if (BrOffset <= DestOffset) {
// Branch before the Dest.
if (DestOffset-BrOffset <= MaxDisp)
return true;
} else {
if (BrOffset-DestOffset <= MaxDisp)
return true;
}
return false;
}
/// FixUpImmediateBr - Fix up an immediate branch whose destination is too far
/// away to fit in its displacement field.
bool ARMConstantIslands::FixUpImmediateBr(MachineFunction &MF, ImmBranch &Br) {
MachineInstr *MI = Br.MI;
MachineBasicBlock *DestBB = MI->getOperand(0).getMBB();
// Check to see if the DestBB is already in-range.
if (BBIsInRange(MI, DestBB, Br.MaxDisp))
return false;
if (!Br.isCond)
return FixUpUnconditionalBr(MF, Br);
return FixUpConditionalBr(MF, Br);
}
/// FixUpUnconditionalBr - Fix up an unconditional branch whose destination is
/// too far away to fit in its displacement field. If the LR register has been
/// spilled in the epilogue, then we can use BL to implement a far jump.
/// Otherwise, add an intermediate branch instruction to a branch.
bool
ARMConstantIslands::FixUpUnconditionalBr(MachineFunction &MF, ImmBranch &Br) {
MachineInstr *MI = Br.MI;
MachineBasicBlock *MBB = MI->getParent();
if (!isThumb1)
llvm_unreachable("FixUpUnconditionalBr is Thumb1 only!");
// Use BL to implement far jump.
Br.MaxDisp = (1 << 21) * 2;
MI->setDesc(TII->get(ARM::tBfar));
BBSizes[MBB->getNumber()] += 2;
AdjustBBOffsetsAfter(MBB, 2);
HasFarJump = true;
++NumUBrFixed;
DEBUG(errs() << " Changed B to long jump " << *MI);
return true;
}
/// FixUpConditionalBr - Fix up a conditional branch whose destination is too
/// far away to fit in its displacement field. It is converted to an inverse
/// conditional branch + an unconditional branch to the destination.
bool
ARMConstantIslands::FixUpConditionalBr(MachineFunction &MF, ImmBranch &Br) {
MachineInstr *MI = Br.MI;
MachineBasicBlock *DestBB = MI->getOperand(0).getMBB();
// Add an unconditional branch to the destination and invert the branch
// condition to jump over it:
// blt L1
// =>
// bge L2
// b L1
// L2:
ARMCC::CondCodes CC = (ARMCC::CondCodes)MI->getOperand(1).getImm();
CC = ARMCC::getOppositeCondition(CC);
unsigned CCReg = MI->getOperand(2).getReg();
// If the branch is at the end of its MBB and that has a fall-through block,
// direct the updated conditional branch to the fall-through block. Otherwise,
// split the MBB before the next instruction.
MachineBasicBlock *MBB = MI->getParent();
MachineInstr *BMI = &MBB->back();
bool NeedSplit = (BMI != MI) || !BBHasFallthrough(MBB);
++NumCBrFixed;
if (BMI != MI) {
if (llvm::next(MachineBasicBlock::iterator(MI)) == prior(MBB->end()) &&
BMI->getOpcode() == Br.UncondBr) {
// Last MI in the BB is an unconditional branch. Can we simply invert the
// condition and swap destinations:
// beq L1
// b L2
// =>
// bne L2
// b L1
MachineBasicBlock *NewDest = BMI->getOperand(0).getMBB();
if (BBIsInRange(MI, NewDest, Br.MaxDisp)) {
DEBUG(errs() << " Invert Bcc condition and swap its destination with "
<< *BMI);
BMI->getOperand(0).setMBB(DestBB);
MI->getOperand(0).setMBB(NewDest);
MI->getOperand(1).setImm(CC);
return true;
}
}
}
if (NeedSplit) {
SplitBlockBeforeInstr(MI);
// No need for the branch to the next block. We're adding an unconditional
// branch to the destination.
int delta = TII->GetInstSizeInBytes(&MBB->back());
BBSizes[MBB->getNumber()] -= delta;
MachineBasicBlock* SplitBB = llvm::next(MachineFunction::iterator(MBB));
AdjustBBOffsetsAfter(SplitBB, -delta);
MBB->back().eraseFromParent();
// BBOffsets[SplitBB] is wrong temporarily, fixed below
}
MachineBasicBlock *NextBB = llvm::next(MachineFunction::iterator(MBB));
DEBUG(errs() << " Insert B to BB#" << DestBB->getNumber()
<< " also invert condition and change dest. to BB#"
<< NextBB->getNumber() << "\n");
// Insert a new conditional branch and a new unconditional branch.
// Also update the ImmBranch as well as adding a new entry for the new branch.
BuildMI(MBB, DebugLoc(), TII->get(MI->getOpcode()))
.addMBB(NextBB).addImm(CC).addReg(CCReg);
Br.MI = &MBB->back();
BBSizes[MBB->getNumber()] += TII->GetInstSizeInBytes(&MBB->back());
BuildMI(MBB, DebugLoc(), TII->get(Br.UncondBr)).addMBB(DestBB);
BBSizes[MBB->getNumber()] += TII->GetInstSizeInBytes(&MBB->back());
unsigned MaxDisp = getUnconditionalBrDisp(Br.UncondBr);
ImmBranches.push_back(ImmBranch(&MBB->back(), MaxDisp, false, Br.UncondBr));
// Remove the old conditional branch. It may or may not still be in MBB.
BBSizes[MI->getParent()->getNumber()] -= TII->GetInstSizeInBytes(MI);
MI->eraseFromParent();
// The net size change is an addition of one unconditional branch.
int delta = TII->GetInstSizeInBytes(&MBB->back());
AdjustBBOffsetsAfter(MBB, delta);
return true;
}
/// UndoLRSpillRestore - Remove Thumb push / pop instructions that only spills
/// LR / restores LR to pc. FIXME: This is done here because it's only possible
/// to do this if tBfar is not used.
bool ARMConstantIslands::UndoLRSpillRestore() {
bool MadeChange = false;
for (unsigned i = 0, e = PushPopMIs.size(); i != e; ++i) {
MachineInstr *MI = PushPopMIs[i];
// First two operands are predicates.
if (MI->getOpcode() == ARM::tPOP_RET &&
MI->getOperand(2).getReg() == ARM::PC &&
MI->getNumExplicitOperands() == 3) {
BuildMI(MI->getParent(), MI->getDebugLoc(), TII->get(ARM::tBX_RET));
MI->eraseFromParent();
MadeChange = true;
}
}
return MadeChange;
}
bool ARMConstantIslands::OptimizeThumb2Instructions(MachineFunction &MF) {
bool MadeChange = false;
// Shrink ADR and LDR from constantpool.
for (unsigned i = 0, e = CPUsers.size(); i != e; ++i) {
CPUser &U = CPUsers[i];
unsigned Opcode = U.MI->getOpcode();
unsigned NewOpc = 0;
unsigned Scale = 1;
unsigned Bits = 0;
switch (Opcode) {
default: break;
case ARM::t2LEApcrel:
if (isARMLowRegister(U.MI->getOperand(0).getReg())) {
NewOpc = ARM::tLEApcrel;
Bits = 8;
Scale = 4;
}
break;
case ARM::t2LDRpci:
if (isARMLowRegister(U.MI->getOperand(0).getReg())) {
NewOpc = ARM::tLDRpci;
Bits = 8;
Scale = 4;
}
break;
}
if (!NewOpc)
continue;
unsigned UserOffset = GetOffsetOf(U.MI) + 4;
unsigned MaxOffs = ((1 << Bits) - 1) * Scale;
// FIXME: Check if offset is multiple of scale if scale is not 4.
if (CPEIsInRange(U.MI, UserOffset, U.CPEMI, MaxOffs, false, true)) {
U.MI->setDesc(TII->get(NewOpc));
MachineBasicBlock *MBB = U.MI->getParent();
BBSizes[MBB->getNumber()] -= 2;
AdjustBBOffsetsAfter(MBB, -2);
++NumT2CPShrunk;
MadeChange = true;
}
}
MadeChange |= OptimizeThumb2Branches(MF);
MadeChange |= OptimizeThumb2JumpTables(MF);
return MadeChange;
}
bool ARMConstantIslands::OptimizeThumb2Branches(MachineFunction &MF) {
bool MadeChange = false;
for (unsigned i = 0, e = ImmBranches.size(); i != e; ++i) {
ImmBranch &Br = ImmBranches[i];
unsigned Opcode = Br.MI->getOpcode();
unsigned NewOpc = 0;
unsigned Scale = 1;
unsigned Bits = 0;
switch (Opcode) {
default: break;
case ARM::t2B:
NewOpc = ARM::tB;
Bits = 11;
Scale = 2;
break;
case ARM::t2Bcc: {
NewOpc = ARM::tBcc;
Bits = 8;
Scale = 2;
break;
}
}
if (NewOpc) {
unsigned MaxOffs = ((1 << (Bits-1))-1) * Scale;
MachineBasicBlock *DestBB = Br.MI->getOperand(0).getMBB();
if (BBIsInRange(Br.MI, DestBB, MaxOffs)) {
Br.MI->setDesc(TII->get(NewOpc));
MachineBasicBlock *MBB = Br.MI->getParent();
BBSizes[MBB->getNumber()] -= 2;
AdjustBBOffsetsAfter(MBB, -2);
++NumT2BrShrunk;
MadeChange = true;
}
}
Opcode = Br.MI->getOpcode();
if (Opcode != ARM::tBcc)
continue;
NewOpc = 0;
unsigned PredReg = 0;
ARMCC::CondCodes Pred = llvm::getInstrPredicate(Br.MI, PredReg);
if (Pred == ARMCC::EQ)
NewOpc = ARM::tCBZ;
else if (Pred == ARMCC::NE)
NewOpc = ARM::tCBNZ;
if (!NewOpc)
continue;
MachineBasicBlock *DestBB = Br.MI->getOperand(0).getMBB();
// Check if the distance is within 126. Subtract starting offset by 2
// because the cmp will be eliminated.
unsigned BrOffset = GetOffsetOf(Br.MI) + 4 - 2;
unsigned DestOffset = BBOffsets[DestBB->getNumber()];
if (BrOffset < DestOffset && (DestOffset - BrOffset) <= 126) {
MachineBasicBlock::iterator CmpMI = Br.MI; --CmpMI;
if (CmpMI->getOpcode() == ARM::tCMPzi8) {
unsigned Reg = CmpMI->getOperand(0).getReg();
Pred = llvm::getInstrPredicate(CmpMI, PredReg);
if (Pred == ARMCC::AL &&
CmpMI->getOperand(1).getImm() == 0 &&
isARMLowRegister(Reg)) {
MachineBasicBlock *MBB = Br.MI->getParent();
MachineInstr *NewBR =
BuildMI(*MBB, CmpMI, Br.MI->getDebugLoc(), TII->get(NewOpc))
.addReg(Reg).addMBB(DestBB, Br.MI->getOperand(0).getTargetFlags());
CmpMI->eraseFromParent();
Br.MI->eraseFromParent();
Br.MI = NewBR;
BBSizes[MBB->getNumber()] -= 2;
AdjustBBOffsetsAfter(MBB, -2);
++NumCBZ;
MadeChange = true;
}
}
}
}
return MadeChange;
}
/// OptimizeThumb2JumpTables - Use tbb / tbh instructions to generate smaller
/// jumptables when it's possible.
bool ARMConstantIslands::OptimizeThumb2JumpTables(MachineFunction &MF) {
bool MadeChange = false;
// FIXME: After the tables are shrunk, can we get rid some of the
// constantpool tables?
MachineJumpTableInfo *MJTI = MF.getJumpTableInfo();
if (MJTI == 0) return false;
const std::vector<MachineJumpTableEntry> &JT = MJTI->getJumpTables();
for (unsigned i = 0, e = T2JumpTables.size(); i != e; ++i) {
MachineInstr *MI = T2JumpTables[i];
const TargetInstrDesc &TID = MI->getDesc();
unsigned NumOps = TID.getNumOperands();
unsigned JTOpIdx = NumOps - (TID.isPredicable() ? 3 : 2);
MachineOperand JTOP = MI->getOperand(JTOpIdx);
unsigned JTI = JTOP.getIndex();
assert(JTI < JT.size());
bool ByteOk = true;
bool HalfWordOk = true;
unsigned JTOffset = GetOffsetOf(MI) + 4;
const std::vector<MachineBasicBlock*> &JTBBs = JT[JTI].MBBs;
for (unsigned j = 0, ee = JTBBs.size(); j != ee; ++j) {
MachineBasicBlock *MBB = JTBBs[j];
unsigned DstOffset = BBOffsets[MBB->getNumber()];
// Negative offset is not ok. FIXME: We should change BB layout to make
// sure all the branches are forward.
if (ByteOk && (DstOffset - JTOffset) > ((1<<8)-1)*2)
ByteOk = false;
unsigned TBHLimit = ((1<<16)-1)*2;
if (HalfWordOk && (DstOffset - JTOffset) > TBHLimit)
HalfWordOk = false;
if (!ByteOk && !HalfWordOk)
break;
}
if (ByteOk || HalfWordOk) {
MachineBasicBlock *MBB = MI->getParent();
unsigned BaseReg = MI->getOperand(0).getReg();
bool BaseRegKill = MI->getOperand(0).isKill();
if (!BaseRegKill)
continue;
unsigned IdxReg = MI->getOperand(1).getReg();
bool IdxRegKill = MI->getOperand(1).isKill();
// Scan backwards to find the instruction that defines the base
// register. Due to post-RA scheduling, we can't count on it
// immediately preceding the branch instruction.
MachineBasicBlock::iterator PrevI = MI;
MachineBasicBlock::iterator B = MBB->begin();
while (PrevI != B && !PrevI->definesRegister(BaseReg))
--PrevI;
// If for some reason we didn't find it, we can't do anything, so
// just skip this one.
if (!PrevI->definesRegister(BaseReg))
continue;
MachineInstr *AddrMI = PrevI;
bool OptOk = true;
// Examine the instruction that calculates the jumptable entry address.
// Make sure it only defines the base register and kills any uses
// other than the index register.
for (unsigned k = 0, eee = AddrMI->getNumOperands(); k != eee; ++k) {
const MachineOperand &MO = AddrMI->getOperand(k);
if (!MO.isReg() || !MO.getReg())
continue;
if (MO.isDef() && MO.getReg() != BaseReg) {
OptOk = false;
break;
}
if (MO.isUse() && !MO.isKill() && MO.getReg() != IdxReg) {
OptOk = false;
break;
}
}
if (!OptOk)
continue;
// Now scan back again to find the tLEApcrel or t2LEApcrelJT instruction
// that gave us the initial base register definition.
for (--PrevI; PrevI != B && !PrevI->definesRegister(BaseReg); --PrevI)
;
// The instruction should be a tLEApcrel or t2LEApcrelJT; we want
// to delete it as well.
MachineInstr *LeaMI = PrevI;
if ((LeaMI->getOpcode() != ARM::tLEApcrelJT &&
LeaMI->getOpcode() != ARM::t2LEApcrelJT) ||
LeaMI->getOperand(0).getReg() != BaseReg)
OptOk = false;
if (!OptOk)
continue;
unsigned Opc = ByteOk ? ARM::t2TBB : ARM::t2TBH;
MachineInstr *NewJTMI = BuildMI(MBB, MI->getDebugLoc(), TII->get(Opc))
.addReg(IdxReg, getKillRegState(IdxRegKill))
.addJumpTableIndex(JTI, JTOP.getTargetFlags())
.addImm(MI->getOperand(JTOpIdx+1).getImm());
// FIXME: Insert an "ALIGN" instruction to ensure the next instruction
// is 2-byte aligned. For now, asm printer will fix it up.
unsigned NewSize = TII->GetInstSizeInBytes(NewJTMI);
unsigned OrigSize = TII->GetInstSizeInBytes(AddrMI);
OrigSize += TII->GetInstSizeInBytes(LeaMI);
OrigSize += TII->GetInstSizeInBytes(MI);
AddrMI->eraseFromParent();
LeaMI->eraseFromParent();
MI->eraseFromParent();
int delta = OrigSize - NewSize;
BBSizes[MBB->getNumber()] -= delta;
AdjustBBOffsetsAfter(MBB, -delta);
++NumTBs;
MadeChange = true;
}
}
return MadeChange;
}
/// ReorderThumb2JumpTables - Adjust the function's block layout to ensure that
/// jump tables always branch forwards, since that's what tbb and tbh need.
bool ARMConstantIslands::ReorderThumb2JumpTables(MachineFunction &MF) {
bool MadeChange = false;
MachineJumpTableInfo *MJTI = MF.getJumpTableInfo();
if (MJTI == 0) return false;
const std::vector<MachineJumpTableEntry> &JT = MJTI->getJumpTables();
for (unsigned i = 0, e = T2JumpTables.size(); i != e; ++i) {
MachineInstr *MI = T2JumpTables[i];
const TargetInstrDesc &TID = MI->getDesc();
unsigned NumOps = TID.getNumOperands();
unsigned JTOpIdx = NumOps - (TID.isPredicable() ? 3 : 2);
MachineOperand JTOP = MI->getOperand(JTOpIdx);
unsigned JTI = JTOP.getIndex();
assert(JTI < JT.size());
// We prefer if target blocks for the jump table come after the jump
// instruction so we can use TB[BH]. Loop through the target blocks
// and try to adjust them such that that's true.
int JTNumber = MI->getParent()->getNumber();
const std::vector<MachineBasicBlock*> &JTBBs = JT[JTI].MBBs;
for (unsigned j = 0, ee = JTBBs.size(); j != ee; ++j) {
MachineBasicBlock *MBB = JTBBs[j];
int DTNumber = MBB->getNumber();
if (DTNumber < JTNumber) {
// The destination precedes the switch. Try to move the block forward
// so we have a positive offset.
MachineBasicBlock *NewBB =
AdjustJTTargetBlockForward(MBB, MI->getParent());
if (NewBB)
MJTI->ReplaceMBBInJumpTable(JTI, JTBBs[j], NewBB);
MadeChange = true;
}
}
}
return MadeChange;
}
MachineBasicBlock *ARMConstantIslands::
AdjustJTTargetBlockForward(MachineBasicBlock *BB, MachineBasicBlock *JTBB)
{
MachineFunction &MF = *BB->getParent();
// If the destination block is terminated by an unconditional branch,
// try to move it; otherwise, create a new block following the jump
// table that branches back to the actual target. This is a very simple
// heuristic. FIXME: We can definitely improve it.
MachineBasicBlock *TBB = 0, *FBB = 0;
SmallVector<MachineOperand, 4> Cond;
SmallVector<MachineOperand, 4> CondPrior;
MachineFunction::iterator BBi = BB;
MachineFunction::iterator OldPrior = prior(BBi);
// If the block terminator isn't analyzable, don't try to move the block
bool B = TII->AnalyzeBranch(*BB, TBB, FBB, Cond);
// If the block ends in an unconditional branch, move it. The prior block
// has to have an analyzable terminator for us to move this one. Be paranoid
// and make sure we're not trying to move the entry block of the function.
if (!B && Cond.empty() && BB != MF.begin() &&
!TII->AnalyzeBranch(*OldPrior, TBB, FBB, CondPrior)) {
BB->moveAfter(JTBB);
OldPrior->updateTerminator();
BB->updateTerminator();
// Update numbering to account for the block being moved.
MF.RenumberBlocks();
++NumJTMoved;
return NULL;
}
// Create a new MBB for the code after the jump BB.
MachineBasicBlock *NewBB =
MF.CreateMachineBasicBlock(JTBB->getBasicBlock());
MachineFunction::iterator MBBI = JTBB; ++MBBI;
MF.insert(MBBI, NewBB);
// Add an unconditional branch from NewBB to BB.
// There doesn't seem to be meaningful DebugInfo available; this doesn't
// correspond directly to anything in the source.
assert (isThumb2 && "Adjusting for TB[BH] but not in Thumb2?");
BuildMI(NewBB, DebugLoc(), TII->get(ARM::t2B)).addMBB(BB);
// Update internal data structures to account for the newly inserted MBB.
MF.RenumberBlocks(NewBB);
// Update the CFG.
NewBB->addSuccessor(BB);
JTBB->removeSuccessor(BB);
JTBB->addSuccessor(NewBB);
++NumJTInserted;
return NewBB;
}