2012-02-18 20:03:15 +08:00
|
|
|
//===-- X86InstrInfo.h - X86 Instruction Information ------------*- C++ -*-===//
|
2005-04-22 07:38:14 +08:00
|
|
|
//
|
2003-10-21 23:17:13 +08:00
|
|
|
// The LLVM Compiler Infrastructure
|
|
|
|
//
|
2007-12-30 04:36:04 +08:00
|
|
|
// This file is distributed under the University of Illinois Open Source
|
|
|
|
// License. See LICENSE.TXT for details.
|
2005-04-22 07:38:14 +08:00
|
|
|
//
|
2003-10-21 23:17:13 +08:00
|
|
|
//===----------------------------------------------------------------------===//
|
2002-10-26 06:55:53 +08:00
|
|
|
//
|
2003-01-15 06:00:31 +08:00
|
|
|
// This file contains the X86 implementation of the TargetInstrInfo class.
|
2002-10-26 06:55:53 +08:00
|
|
|
//
|
|
|
|
//===----------------------------------------------------------------------===//
|
|
|
|
|
2014-08-14 00:26:38 +08:00
|
|
|
#ifndef LLVM_LIB_TARGET_X86_X86INSTRINFO_H
|
|
|
|
#define LLVM_LIB_TARGET_X86_X86INSTRINFO_H
|
2002-10-26 06:55:53 +08:00
|
|
|
|
2014-03-19 14:53:25 +08:00
|
|
|
#include "MCTargetDesc/X86BaseInfo.h"
|
2016-08-12 06:07:33 +08:00
|
|
|
#include "X86InstrFMA3Info.h"
|
2002-10-26 06:55:53 +08:00
|
|
|
#include "X86RegisterInfo.h"
|
2009-01-06 01:59:02 +08:00
|
|
|
#include "llvm/ADT/DenseMap.h"
|
2012-03-18 02:46:09 +08:00
|
|
|
#include "llvm/Target/TargetInstrInfo.h"
|
2002-10-26 06:55:53 +08:00
|
|
|
|
2011-07-02 01:57:27 +08:00
|
|
|
#define GET_INSTRINFO_HEADER
|
|
|
|
#include "X86GenInstrInfo.inc"
|
|
|
|
|
2003-11-12 06:41:34 +08:00
|
|
|
namespace llvm {
|
2016-03-25 09:10:56 +08:00
|
|
|
class MachineInstrBuilder;
|
2006-09-08 14:48:29 +08:00
|
|
|
class X86RegisterInfo;
|
2014-06-11 06:34:31 +08:00
|
|
|
class X86Subtarget;
|
2003-11-12 06:41:34 +08:00
|
|
|
|
2006-10-21 01:42:20 +08:00
|
|
|
namespace X86 {
|
|
|
|
// X86 specific condition code. These correspond to X86_*_COND in
|
|
|
|
// X86InstrInfo.td. They must be kept in synch.
|
2016-03-24 05:45:37 +08:00
|
|
|
enum CondCode {
|
|
|
|
COND_A = 0,
|
|
|
|
COND_AE = 1,
|
|
|
|
COND_B = 2,
|
|
|
|
COND_BE = 3,
|
|
|
|
COND_E = 4,
|
|
|
|
COND_G = 5,
|
|
|
|
COND_GE = 6,
|
|
|
|
COND_L = 7,
|
|
|
|
COND_LE = 8,
|
|
|
|
COND_NE = 9,
|
|
|
|
COND_NO = 10,
|
|
|
|
COND_NP = 11,
|
|
|
|
COND_NS = 12,
|
|
|
|
COND_O = 13,
|
|
|
|
COND_P = 14,
|
|
|
|
COND_S = 15,
|
|
|
|
LAST_VALID_COND = COND_S,
|
Optimized FCMP_OEQ and FCMP_UNE for x86.
Where previously LLVM might emit code like this:
ucomisd %xmm1, %xmm0
setne %al
setp %cl
orb %al, %cl
jne .LBB4_2
it now emits this:
ucomisd %xmm1, %xmm0
jne .LBB4_2
jp .LBB4_2
It has fewer instructions and uses fewer registers, but it does
have more branches. And in the case that this code is followed by
a non-fallthrough edge, it may be followed by a jmp instruction,
resulting in three branch instructions in sequence. Some effort
is made to avoid this situation.
To achieve this, X86ISelLowering.cpp now recognizes FCMP_OEQ and
FCMP_UNE in lowered form, and replace them with code that emits
two branches, except in the case where it would require converting
a fall-through edge to an explicit branch.
Also, X86InstrInfo.cpp's branch analysis and transform code now
knows now to handle blocks with multiple conditional branches. It
uses loops instead of having fixed checks for up to two
instructions. It can now analyze and transform code generated
from FCMP_OEQ and FCMP_UNE.
llvm-svn: 57873
2008-10-21 11:29:32 +08:00
|
|
|
|
2016-03-24 05:45:37 +08:00
|
|
|
// Artificial condition codes. These are used by AnalyzeBranch
|
|
|
|
// to indicate a block terminated with two conditional branches that together
|
|
|
|
// form a compound condition. They occur in code using FCMP_OEQ or FCMP_UNE,
|
|
|
|
// which can't be represented on x86 with a single condition. These
|
|
|
|
// are never used in MachineInstrs and are inverses of one another.
|
|
|
|
COND_NE_OR_P,
|
|
|
|
COND_E_AND_NP,
|
Optimized FCMP_OEQ and FCMP_UNE for x86.
Where previously LLVM might emit code like this:
ucomisd %xmm1, %xmm0
setne %al
setp %cl
orb %al, %cl
jne .LBB4_2
it now emits this:
ucomisd %xmm1, %xmm0
jne .LBB4_2
jp .LBB4_2
It has fewer instructions and uses fewer registers, but it does
have more branches. And in the case that this code is followed by
a non-fallthrough edge, it may be followed by a jmp instruction,
resulting in three branch instructions in sequence. Some effort
is made to avoid this situation.
To achieve this, X86ISelLowering.cpp now recognizes FCMP_OEQ and
FCMP_UNE in lowered form, and replace them with code that emits
two branches, except in the case where it would require converting
a fall-through edge to an explicit branch.
Also, X86InstrInfo.cpp's branch analysis and transform code now
knows now to handle blocks with multiple conditional branches. It
uses loops instead of having fixed checks for up to two
instructions. It can now analyze and transform code generated
from FCMP_OEQ and FCMP_UNE.
llvm-svn: 57873
2008-10-21 11:29:32 +08:00
|
|
|
|
2016-03-24 05:45:37 +08:00
|
|
|
COND_INVALID
|
|
|
|
};
|
2011-03-05 14:31:54 +08:00
|
|
|
|
2016-03-24 05:45:37 +08:00
|
|
|
// Turn condition code into conditional branch opcode.
|
|
|
|
unsigned GetCondBranchFromCond(CondCode CC);
|
2011-03-05 14:31:54 +08:00
|
|
|
|
2016-03-24 05:45:37 +08:00
|
|
|
/// \brief Return a set opcode for the given condition and whether it has
|
|
|
|
/// a memory operand.
|
|
|
|
unsigned getSETFromCond(CondCode CC, bool HasMemoryOperand = false);
|
2014-06-24 05:55:36 +08:00
|
|
|
|
2016-03-24 05:45:37 +08:00
|
|
|
/// \brief Return a cmov opcode for the given condition, register size in
|
|
|
|
/// bytes, and operand type.
|
|
|
|
unsigned getCMovFromCond(CondCode CC, unsigned RegBytes,
|
|
|
|
bool HasMemoryOperand = false);
|
2014-06-17 07:58:24 +08:00
|
|
|
|
2016-03-24 05:45:37 +08:00
|
|
|
// Turn CMov opcode into condition code.
|
|
|
|
CondCode getCondFromCMovOpc(unsigned Opc);
|
2012-09-20 11:06:15 +08:00
|
|
|
|
2016-03-24 05:45:37 +08:00
|
|
|
/// GetOppositeBranchCondition - Return the inverse of the specified cond,
|
|
|
|
/// e.g. turning COND_E to COND_NE.
|
|
|
|
CondCode GetOppositeBranchCondition(CondCode CC);
|
2015-06-23 17:49:53 +08:00
|
|
|
} // end namespace X86;
|
2006-10-21 13:52:40 +08:00
|
|
|
|
2009-07-10 14:06:17 +08:00
|
|
|
|
2009-07-10 14:29:59 +08:00
|
|
|
/// isGlobalStubReference - Return true if the specified TargetFlag operand is
|
2009-07-10 14:06:17 +08:00
|
|
|
/// a reference to a stub for a global, not the global itself.
|
2009-07-10 14:29:59 +08:00
|
|
|
inline static bool isGlobalStubReference(unsigned char TargetFlag) {
|
|
|
|
switch (TargetFlag) {
|
2009-07-10 14:06:17 +08:00
|
|
|
case X86II::MO_DLLIMPORT: // dllimport stub.
|
|
|
|
case X86II::MO_GOTPCREL: // rip-relative GOT reference.
|
|
|
|
case X86II::MO_GOT: // normal GOT reference.
|
|
|
|
case X86II::MO_DARWIN_NONLAZY_PIC_BASE: // Normal $non_lazy_ptr ref.
|
|
|
|
case X86II::MO_DARWIN_NONLAZY: // Normal $non_lazy_ptr ref.
|
|
|
|
return true;
|
|
|
|
default:
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
}
|
2009-07-10 15:33:30 +08:00
|
|
|
|
|
|
|
/// isGlobalRelativeToPICBase - Return true if the specified global value
|
|
|
|
/// reference is relative to a 32-bit PIC base (X86ISD::GlobalBaseReg). If this
|
|
|
|
/// is true, the addressing mode has the PIC base register added in (e.g. EBX).
|
|
|
|
inline static bool isGlobalRelativeToPICBase(unsigned char TargetFlag) {
|
|
|
|
switch (TargetFlag) {
|
|
|
|
case X86II::MO_GOTOFF: // isPICStyleGOT: local global.
|
|
|
|
case X86II::MO_GOT: // isPICStyleGOT: other global.
|
|
|
|
case X86II::MO_PIC_BASE_OFFSET: // Darwin local global.
|
|
|
|
case X86II::MO_DARWIN_NONLAZY_PIC_BASE: // Darwin/32 external global.
|
2010-06-03 12:07:48 +08:00
|
|
|
case X86II::MO_TLVP: // ??? Pretty sure..
|
2009-07-10 15:33:30 +08:00
|
|
|
return true;
|
|
|
|
default:
|
|
|
|
return false;
|
|
|
|
}
|
|
|
|
}
|
2011-03-05 14:31:54 +08:00
|
|
|
|
2008-06-28 19:07:54 +08:00
|
|
|
inline static bool isScale(const MachineOperand &MO) {
|
2008-10-03 23:45:36 +08:00
|
|
|
return MO.isImm() &&
|
2008-06-28 19:07:54 +08:00
|
|
|
(MO.getImm() == 1 || MO.getImm() == 2 ||
|
|
|
|
MO.getImm() == 4 || MO.getImm() == 8);
|
|
|
|
}
|
|
|
|
|
2016-06-30 08:01:54 +08:00
|
|
|
inline static bool isLeaMem(const MachineInstr &MI, unsigned Op) {
|
|
|
|
if (MI.getOperand(Op).isFI())
|
|
|
|
return true;
|
|
|
|
return Op + X86::AddrSegmentReg <= MI.getNumOperands() &&
|
|
|
|
MI.getOperand(Op + X86::AddrBaseReg).isReg() &&
|
|
|
|
isScale(MI.getOperand(Op + X86::AddrScaleAmt)) &&
|
|
|
|
MI.getOperand(Op + X86::AddrIndexReg).isReg() &&
|
|
|
|
(MI.getOperand(Op + X86::AddrDisp).isImm() ||
|
|
|
|
MI.getOperand(Op + X86::AddrDisp).isGlobal() ||
|
|
|
|
MI.getOperand(Op + X86::AddrDisp).isCPI() ||
|
|
|
|
MI.getOperand(Op + X86::AddrDisp).isJTI());
|
2008-06-28 19:07:54 +08:00
|
|
|
}
|
|
|
|
|
2016-06-30 08:01:54 +08:00
|
|
|
inline static bool isMem(const MachineInstr &MI, unsigned Op) {
|
|
|
|
if (MI.getOperand(Op).isFI())
|
|
|
|
return true;
|
|
|
|
return Op + X86::AddrNumOperands <= MI.getNumOperands() &&
|
|
|
|
MI.getOperand(Op + X86::AddrSegmentReg).isReg() && isLeaMem(MI, Op);
|
2009-04-09 05:14:34 +08:00
|
|
|
}
|
|
|
|
|
2014-03-31 14:53:13 +08:00
|
|
|
class X86InstrInfo final : public X86GenInstrInfo {
|
2014-06-11 06:34:31 +08:00
|
|
|
X86Subtarget &Subtarget;
|
2002-10-26 06:55:53 +08:00
|
|
|
const X86RegisterInfo RI;
|
2011-03-05 14:31:54 +08:00
|
|
|
|
2012-06-01 13:34:01 +08:00
|
|
|
/// RegOp2MemOpTable3Addr, RegOp2MemOpTable0, RegOp2MemOpTable1,
|
|
|
|
/// RegOp2MemOpTable2, RegOp2MemOpTable3 - Load / store folding opcode maps.
|
2008-01-07 09:35:02 +08:00
|
|
|
///
|
* Combines Alignment, AuxInfo, and TB_NOT_REVERSABLE flag into a
single field (Flags), which is a bitwise OR of items from the TB_*
enum. This makes it easier to add new information in the future.
* Gives every static array an equivalent layout: { RegOp, MemOp, Flags }
* Adds a helper function, AddTableEntry, to avoid duplication of the
insertion code.
* Renames TB_NOT_REVERSABLE to TB_NO_REVERSE.
* Adds TB_NO_FORWARD, which is analogous to TB_NO_REVERSE, except that
it prevents addition of the Reg->Mem entry. (This is going to be used
by Native Client, in the next CL).
Patch by David Meyer
llvm-svn: 139311
2011-09-09 02:35:57 +08:00
|
|
|
typedef DenseMap<unsigned,
|
2016-05-01 01:59:49 +08:00
|
|
|
std::pair<uint16_t, uint16_t> > RegOp2MemOpTableType;
|
* Combines Alignment, AuxInfo, and TB_NOT_REVERSABLE flag into a
single field (Flags), which is a bitwise OR of items from the TB_*
enum. This makes it easier to add new information in the future.
* Gives every static array an equivalent layout: { RegOp, MemOp, Flags }
* Adds a helper function, AddTableEntry, to avoid duplication of the
insertion code.
* Renames TB_NOT_REVERSABLE to TB_NO_REVERSE.
* Adds TB_NO_FORWARD, which is analogous to TB_NO_REVERSE, except that
it prevents addition of the Reg->Mem entry. (This is going to be used
by Native Client, in the next CL).
Patch by David Meyer
llvm-svn: 139311
2011-09-09 02:35:57 +08:00
|
|
|
RegOp2MemOpTableType RegOp2MemOpTable2Addr;
|
|
|
|
RegOp2MemOpTableType RegOp2MemOpTable0;
|
|
|
|
RegOp2MemOpTableType RegOp2MemOpTable1;
|
|
|
|
RegOp2MemOpTableType RegOp2MemOpTable2;
|
2012-05-31 17:20:20 +08:00
|
|
|
RegOp2MemOpTableType RegOp2MemOpTable3;
|
2014-12-18 20:28:22 +08:00
|
|
|
RegOp2MemOpTableType RegOp2MemOpTable4;
|
2011-03-05 14:31:54 +08:00
|
|
|
|
2008-01-07 09:35:02 +08:00
|
|
|
/// MemOp2RegOpTable - Load / store unfolding opcode map.
|
|
|
|
///
|
* Combines Alignment, AuxInfo, and TB_NOT_REVERSABLE flag into a
single field (Flags), which is a bitwise OR of items from the TB_*
enum. This makes it easier to add new information in the future.
* Gives every static array an equivalent layout: { RegOp, MemOp, Flags }
* Adds a helper function, AddTableEntry, to avoid duplication of the
insertion code.
* Renames TB_NOT_REVERSABLE to TB_NO_REVERSE.
* Adds TB_NO_FORWARD, which is analogous to TB_NO_REVERSE, except that
it prevents addition of the Reg->Mem entry. (This is going to be used
by Native Client, in the next CL).
Patch by David Meyer
llvm-svn: 139311
2011-09-09 02:35:57 +08:00
|
|
|
typedef DenseMap<unsigned,
|
2016-05-01 01:59:49 +08:00
|
|
|
std::pair<uint16_t, uint16_t> > MemOp2RegOpTableType;
|
* Combines Alignment, AuxInfo, and TB_NOT_REVERSABLE flag into a
single field (Flags), which is a bitwise OR of items from the TB_*
enum. This makes it easier to add new information in the future.
* Gives every static array an equivalent layout: { RegOp, MemOp, Flags }
* Adds a helper function, AddTableEntry, to avoid duplication of the
insertion code.
* Renames TB_NOT_REVERSABLE to TB_NO_REVERSE.
* Adds TB_NO_FORWARD, which is analogous to TB_NO_REVERSE, except that
it prevents addition of the Reg->Mem entry. (This is going to be used
by Native Client, in the next CL).
Patch by David Meyer
llvm-svn: 139311
2011-09-09 02:35:57 +08:00
|
|
|
MemOp2RegOpTableType MemOp2RegOpTable;
|
|
|
|
|
2012-06-23 12:58:41 +08:00
|
|
|
static void AddTableEntry(RegOp2MemOpTableType &R2MTable,
|
|
|
|
MemOp2RegOpTableType &M2RTable,
|
2016-05-01 01:59:49 +08:00
|
|
|
uint16_t RegOp, uint16_t MemOp, uint16_t Flags);
|
2010-03-26 01:25:00 +08:00
|
|
|
|
2013-11-19 08:57:56 +08:00
|
|
|
virtual void anchor();
|
|
|
|
|
2015-06-16 02:44:21 +08:00
|
|
|
bool AnalyzeBranchImpl(MachineBasicBlock &MBB, MachineBasicBlock *&TBB,
|
|
|
|
MachineBasicBlock *&FBB,
|
|
|
|
SmallVectorImpl<MachineOperand> &Cond,
|
|
|
|
SmallVectorImpl<MachineInstr *> &CondBranches,
|
|
|
|
bool AllowModify) const;
|
|
|
|
|
2002-10-26 06:55:53 +08:00
|
|
|
public:
|
2014-06-11 06:34:31 +08:00
|
|
|
explicit X86InstrInfo(X86Subtarget &STI);
|
2002-10-26 06:55:53 +08:00
|
|
|
|
2003-01-15 06:00:31 +08:00
|
|
|
/// getRegisterInfo - TargetInstrInfo is a superset of MRegister info. As
|
2002-10-26 06:55:53 +08:00
|
|
|
/// such, whenever a client has an instance of instruction info, it should
|
|
|
|
/// always be able to get register info as well (through this method).
|
|
|
|
///
|
2014-03-09 15:58:15 +08:00
|
|
|
const X86RegisterInfo &getRegisterInfo() const { return RI; }
|
2002-10-26 06:55:53 +08:00
|
|
|
|
2015-02-02 00:56:04 +08:00
|
|
|
/// getSPAdjust - This returns the stack pointer adjustment made by
|
|
|
|
/// this instruction. For x86, we need to handle more complex call
|
|
|
|
/// sequences involving PUSHes.
|
2016-06-30 08:01:54 +08:00
|
|
|
int getSPAdjust(const MachineInstr &MI) const override;
|
2015-02-02 00:56:04 +08:00
|
|
|
|
2010-01-13 08:30:23 +08:00
|
|
|
/// isCoalescableExtInstr - Return true if the instruction is a "coalescable"
|
|
|
|
/// extension instruction. That is, it's like a copy where it's legal for the
|
|
|
|
/// source to overlap the destination. e.g. X86::MOVSX64rr32. If this returns
|
|
|
|
/// true, then it's expected the pre-extension value is available as a subreg
|
|
|
|
/// of the result register. This also returns the sub-register index in
|
|
|
|
/// SubIdx.
|
2014-03-09 15:44:38 +08:00
|
|
|
bool isCoalescableExtInstr(const MachineInstr &MI,
|
|
|
|
unsigned &SrcReg, unsigned &DstReg,
|
|
|
|
unsigned &SubIdx) const override;
|
2010-01-12 08:09:37 +08:00
|
|
|
|
2016-06-30 08:01:54 +08:00
|
|
|
unsigned isLoadFromStackSlot(const MachineInstr &MI,
|
2014-03-09 15:44:38 +08:00
|
|
|
int &FrameIndex) const override;
|
2009-11-13 08:29:53 +08:00
|
|
|
/// isLoadFromStackSlotPostFE - Check for post-frame ptr elimination
|
|
|
|
/// stack locations as well. This uses a heuristic so it isn't
|
|
|
|
/// reliable for correctness.
|
2016-06-30 08:01:54 +08:00
|
|
|
unsigned isLoadFromStackSlotPostFE(const MachineInstr &MI,
|
2014-03-09 15:44:38 +08:00
|
|
|
int &FrameIndex) const override;
|
2009-11-13 04:55:29 +08:00
|
|
|
|
2016-06-30 08:01:54 +08:00
|
|
|
unsigned isStoreToStackSlot(const MachineInstr &MI,
|
2014-03-09 15:44:38 +08:00
|
|
|
int &FrameIndex) const override;
|
2009-11-13 08:29:53 +08:00
|
|
|
/// isStoreToStackSlotPostFE - Check for post-frame ptr elimination
|
|
|
|
/// stack locations as well. This uses a heuristic so it isn't
|
|
|
|
/// reliable for correctness.
|
2016-06-30 08:01:54 +08:00
|
|
|
unsigned isStoreToStackSlotPostFE(const MachineInstr &MI,
|
2014-03-09 15:44:38 +08:00
|
|
|
int &FrameIndex) const override;
|
2008-04-01 04:40:39 +08:00
|
|
|
|
2016-06-30 08:01:54 +08:00
|
|
|
bool isReallyTriviallyReMaterializable(const MachineInstr &MI,
|
2014-03-09 15:44:38 +08:00
|
|
|
AliasAnalysis *AA) const override;
|
2008-04-01 04:40:39 +08:00
|
|
|
void reMaterialize(MachineBasicBlock &MBB, MachineBasicBlock::iterator MI,
|
2009-07-16 17:20:10 +08:00
|
|
|
unsigned DestReg, unsigned SubIdx,
|
2016-06-30 08:01:54 +08:00
|
|
|
const MachineInstr &Orig,
|
2014-03-09 15:44:38 +08:00
|
|
|
const TargetRegisterInfo &TRI) const override;
|
2008-04-01 04:40:39 +08:00
|
|
|
|
2013-06-11 04:43:49 +08:00
|
|
|
/// Given an operand within a MachineInstr, insert preceding code to put it
|
|
|
|
/// into the right format for a particular kind of LEA instruction. This may
|
|
|
|
/// involve using an appropriate super-register instead (with an implicit use
|
|
|
|
/// of the original) or creating a new virtual register and inserting COPY
|
|
|
|
/// instructions to get the data into the right class.
|
|
|
|
///
|
|
|
|
/// Reference parameters are set to indicate how caller should add this
|
|
|
|
/// operand to the LEA instruction.
|
2016-06-30 08:01:54 +08:00
|
|
|
bool classifyLEAReg(MachineInstr &MI, const MachineOperand &Src,
|
|
|
|
unsigned LEAOpcode, bool AllowSP, unsigned &NewSrc,
|
|
|
|
bool &isKill, bool &isUndef,
|
2016-08-09 09:47:26 +08:00
|
|
|
MachineOperand &ImplicitOp, LiveVariables *LV) const;
|
2013-06-11 04:43:49 +08:00
|
|
|
|
2005-01-02 10:37:07 +08:00
|
|
|
/// convertToThreeAddress - This method must be implemented by targets that
|
|
|
|
/// set the M_CONVERTIBLE_TO_3_ADDR flag. When this flag is set, the target
|
|
|
|
/// may be able to convert a two-address instruction into a true
|
|
|
|
/// three-address instruction on demand. This allows the X86 target (for
|
|
|
|
/// example) to convert ADD and SHL instructions into LEA instructions if they
|
|
|
|
/// would require register copies due to two-addressness.
|
|
|
|
///
|
|
|
|
/// This method returns a null pointer if the transformation cannot be
|
|
|
|
/// performed, otherwise it returns the new instruction.
|
|
|
|
///
|
2014-03-09 15:44:38 +08:00
|
|
|
MachineInstr *convertToThreeAddress(MachineFunction::iterator &MFI,
|
2016-06-30 08:01:54 +08:00
|
|
|
MachineInstr &MI,
|
2014-03-09 15:44:38 +08:00
|
|
|
LiveVariables *LV) const override;
|
2005-01-02 10:37:07 +08:00
|
|
|
|
2015-09-29 04:33:22 +08:00
|
|
|
/// Returns true iff the routine could find two commutable operands in the
|
|
|
|
/// given machine instruction.
|
|
|
|
/// The 'SrcOpIdx1' and 'SrcOpIdx2' are INPUT and OUTPUT arguments. Their
|
|
|
|
/// input values can be re-defined in this method only if the input values
|
|
|
|
/// are not pre-defined, which is designated by the special value
|
|
|
|
/// 'CommuteAnyOperandIndex' assigned to it.
|
|
|
|
/// If both of indices are pre-defined and refer to some operands, then the
|
|
|
|
/// method simply returns true if the corresponding operands are commutable
|
|
|
|
/// and returns false otherwise.
|
Teach the code generator that shrd/shld is commutable if it has an immediate.
This allows us to generate this:
foo:
mov %EAX, DWORD PTR [%ESP + 4]
mov %EDX, DWORD PTR [%ESP + 8]
shld %EDX, %EDX, 2
shl %EAX, 2
ret
instead of this:
foo:
mov %EAX, DWORD PTR [%ESP + 4]
mov %ECX, DWORD PTR [%ESP + 8]
mov %EDX, %EAX
shrd %EDX, %ECX, 30
shl %EAX, 2
ret
Note the magically transmogrifying immediate.
llvm-svn: 19686
2005-01-19 15:11:01 +08:00
|
|
|
///
|
2015-09-29 04:33:22 +08:00
|
|
|
/// For example, calling this method this way:
|
|
|
|
/// unsigned Op1 = 1, Op2 = CommuteAnyOperandIndex;
|
|
|
|
/// findCommutedOpIndices(MI, Op1, Op2);
|
|
|
|
/// can be interpreted as a query asking to find an operand that would be
|
|
|
|
/// commutable with the operand#1.
|
2016-06-30 08:01:54 +08:00
|
|
|
bool findCommutedOpIndices(MachineInstr &MI, unsigned &SrcOpIdx1,
|
2014-04-03 07:57:49 +08:00
|
|
|
unsigned &SrcOpIdx2) const override;
|
|
|
|
|
2015-11-07 03:47:25 +08:00
|
|
|
/// Returns true if the routine could find two commutable operands
|
2016-08-12 06:07:33 +08:00
|
|
|
/// in the given FMA instruction \p MI. Otherwise, returns false.
|
2015-11-07 03:47:25 +08:00
|
|
|
///
|
|
|
|
/// \p SrcOpIdx1 and \p SrcOpIdx2 are INPUT and OUTPUT arguments.
|
|
|
|
/// The output indices of the commuted operands are returned in these
|
|
|
|
/// arguments. Also, the input values of these arguments may be preset either
|
|
|
|
/// to indices of operands that must be commuted or be equal to a special
|
|
|
|
/// value 'CommuteAnyOperandIndex' which means that the corresponding
|
|
|
|
/// operand index is not set and this method is free to pick any of
|
|
|
|
/// available commutable operands.
|
2016-08-12 06:07:33 +08:00
|
|
|
/// The parameter \p FMA3Group keeps the reference to the group of relative
|
|
|
|
/// FMA3 opcodes including register/memory forms of 132/213/231 opcodes.
|
2015-11-07 03:47:25 +08:00
|
|
|
///
|
|
|
|
/// For example, calling this method this way:
|
|
|
|
/// unsigned Idx1 = 1, Idx2 = CommuteAnyOperandIndex;
|
2016-08-12 06:07:33 +08:00
|
|
|
/// findFMA3CommutedOpIndices(MI, Idx1, Idx2, FMA3Group);
|
2015-11-07 03:47:25 +08:00
|
|
|
/// can be interpreted as a query asking if the operand #1 can be swapped
|
|
|
|
/// with any other available operand (e.g. operand #2, operand #3, etc.).
|
|
|
|
///
|
|
|
|
/// The returned FMA opcode may differ from the opcode in the given MI.
|
|
|
|
/// For example, commuting the operands #1 and #3 in the following FMA
|
|
|
|
/// FMA213 #1, #2, #3
|
|
|
|
/// results into instruction with adjusted opcode:
|
|
|
|
/// FMA231 #3, #2, #1
|
2016-08-12 06:07:33 +08:00
|
|
|
bool findFMA3CommutedOpIndices(const MachineInstr &MI,
|
2016-07-23 15:16:53 +08:00
|
|
|
unsigned &SrcOpIdx1,
|
2016-08-12 06:07:33 +08:00
|
|
|
unsigned &SrcOpIdx2,
|
|
|
|
const X86InstrFMA3Group &FMA3Group) const;
|
|
|
|
|
|
|
|
/// Returns an adjusted FMA opcode that must be used in FMA instruction that
|
|
|
|
/// performs the same computations as the given \p MI but which has the
|
|
|
|
/// operands \p SrcOpIdx1 and \p SrcOpIdx2 commuted.
|
|
|
|
/// It may return 0 if it is unsafe to commute the operands.
|
|
|
|
/// Note that a machine instruction (instead of its opcode) is passed as the
|
|
|
|
/// first parameter to make it possible to analyze the instruction's uses and
|
|
|
|
/// commute the first operand of FMA even when it seems unsafe when you look
|
|
|
|
/// at the opcode. For example, it is Ok to commute the first operand of
|
|
|
|
/// VFMADD*SD_Int, if ONLY the lowest 64-bit element of the result is used.
|
|
|
|
///
|
|
|
|
/// The returned FMA opcode may differ from the opcode in the given \p MI.
|
|
|
|
/// For example, commuting the operands #1 and #3 in the following FMA
|
|
|
|
/// FMA213 #1, #2, #3
|
|
|
|
/// results into instruction with adjusted opcode:
|
|
|
|
/// FMA231 #3, #2, #1
|
|
|
|
unsigned getFMA3OpcodeToCommuteOperands(const MachineInstr &MI,
|
|
|
|
unsigned SrcOpIdx1,
|
|
|
|
unsigned SrcOpIdx2,
|
|
|
|
const X86InstrFMA3Group &FMA3Group) const;
|
2015-11-07 03:47:25 +08:00
|
|
|
|
2006-10-21 01:42:20 +08:00
|
|
|
// Branch analysis.
|
2016-02-23 10:46:52 +08:00
|
|
|
bool isUnpredicatedTerminator(const MachineInstr &MI) const override;
|
X86: Fold tail calls into conditional branches where possible (PR26302)
When branching to a block that immediately tail calls, it is possible to fold
the call directly into the branch if the call is direct and there is no stack
adjustment, saving one byte.
Example:
define void @f(i32 %x, i32 %y) {
entry:
%p = icmp eq i32 %x, %y
br i1 %p, label %bb1, label %bb2
bb1:
tail call void @foo()
ret void
bb2:
tail call void @bar()
ret void
}
before:
f:
movl 4(%esp), %eax
cmpl 8(%esp), %eax
jne .LBB0_2
jmp foo
.LBB0_2:
jmp bar
after:
f:
movl 4(%esp), %eax
cmpl 8(%esp), %eax
jne bar
.LBB0_1:
jmp foo
I don't expect any significant size savings from this (on a Clang bootstrap I
saw 288 bytes), but it does make the code a little tighter.
This patch only does 32-bit, but 64-bit would work similarly.
Differential Revision: https://reviews.llvm.org/D24108
llvm-svn: 280832
2016-09-08 01:52:14 +08:00
|
|
|
bool isUnconditionalTailCall(const MachineInstr &MI) const override;
|
|
|
|
bool canMakeTailCallConditional(SmallVectorImpl<MachineOperand> &Cond,
|
|
|
|
const MachineInstr &TailCall) const override;
|
|
|
|
void replaceBranchWithTailCall(MachineBasicBlock &MBB,
|
|
|
|
SmallVectorImpl<MachineOperand> &Cond,
|
|
|
|
const MachineInstr &TailCall) const override;
|
|
|
|
|
2016-07-15 22:41:04 +08:00
|
|
|
bool analyzeBranch(MachineBasicBlock &MBB, MachineBasicBlock *&TBB,
|
2014-03-09 15:44:38 +08:00
|
|
|
MachineBasicBlock *&FBB,
|
|
|
|
SmallVectorImpl<MachineOperand> &Cond,
|
|
|
|
bool AllowModify) const override;
|
2015-06-16 02:44:14 +08:00
|
|
|
|
2016-06-30 08:01:54 +08:00
|
|
|
bool getMemOpBaseRegImmOfs(MachineInstr &LdSt, unsigned &BaseReg,
|
2016-03-10 00:00:35 +08:00
|
|
|
int64_t &Offset,
|
2015-06-16 02:44:14 +08:00
|
|
|
const TargetRegisterInfo *TRI) const override;
|
2016-07-15 22:41:04 +08:00
|
|
|
bool analyzeBranchPredicate(MachineBasicBlock &MBB,
|
2015-06-16 02:44:21 +08:00
|
|
|
TargetInstrInfo::MachineBranchPredicate &MBP,
|
|
|
|
bool AllowModify = false) const override;
|
|
|
|
|
2016-09-15 04:43:16 +08:00
|
|
|
unsigned removeBranch(MachineBasicBlock &MBB,
|
2016-09-15 01:23:48 +08:00
|
|
|
int *BytesRemoved = nullptr) const override;
|
2016-09-15 01:24:15 +08:00
|
|
|
unsigned insertBranch(MachineBasicBlock &MBB, MachineBasicBlock *TBB,
|
2015-06-12 03:30:37 +08:00
|
|
|
MachineBasicBlock *FBB, ArrayRef<MachineOperand> Cond,
|
2016-09-15 01:23:48 +08:00
|
|
|
const DebugLoc &DL,
|
|
|
|
int *BytesAdded = nullptr) const override;
|
2015-06-12 03:30:37 +08:00
|
|
|
bool canInsertSelect(const MachineBasicBlock&, ArrayRef<MachineOperand> Cond,
|
2014-03-09 15:44:38 +08:00
|
|
|
unsigned, unsigned, int&, int&, int&) const override;
|
2016-06-12 23:39:02 +08:00
|
|
|
void insertSelect(MachineBasicBlock &MBB, MachineBasicBlock::iterator MI,
|
|
|
|
const DebugLoc &DL, unsigned DstReg,
|
|
|
|
ArrayRef<MachineOperand> Cond, unsigned TrueReg,
|
|
|
|
unsigned FalseReg) const override;
|
|
|
|
void copyPhysReg(MachineBasicBlock &MBB, MachineBasicBlock::iterator MI,
|
|
|
|
const DebugLoc &DL, unsigned DestReg, unsigned SrcReg,
|
2014-03-09 15:44:38 +08:00
|
|
|
bool KillSrc) const override;
|
|
|
|
void storeRegToStackSlot(MachineBasicBlock &MBB,
|
|
|
|
MachineBasicBlock::iterator MI,
|
|
|
|
unsigned SrcReg, bool isKill, int FrameIndex,
|
|
|
|
const TargetRegisterClass *RC,
|
|
|
|
const TargetRegisterInfo *TRI) const override;
|
2008-01-02 05:11:32 +08:00
|
|
|
|
2014-03-09 15:58:15 +08:00
|
|
|
void storeRegToAddr(MachineFunction &MF, unsigned SrcReg, bool isKill,
|
|
|
|
SmallVectorImpl<MachineOperand> &Addr,
|
|
|
|
const TargetRegisterClass *RC,
|
|
|
|
MachineInstr::mmo_iterator MMOBegin,
|
|
|
|
MachineInstr::mmo_iterator MMOEnd,
|
|
|
|
SmallVectorImpl<MachineInstr*> &NewMIs) const;
|
2008-01-02 05:11:32 +08:00
|
|
|
|
2014-03-09 15:44:38 +08:00
|
|
|
void loadRegFromStackSlot(MachineBasicBlock &MBB,
|
|
|
|
MachineBasicBlock::iterator MI,
|
|
|
|
unsigned DestReg, int FrameIndex,
|
|
|
|
const TargetRegisterClass *RC,
|
|
|
|
const TargetRegisterInfo *TRI) const override;
|
2008-01-02 05:11:32 +08:00
|
|
|
|
2014-03-09 15:58:15 +08:00
|
|
|
void loadRegFromAddr(MachineFunction &MF, unsigned DestReg,
|
|
|
|
SmallVectorImpl<MachineOperand> &Addr,
|
|
|
|
const TargetRegisterClass *RC,
|
|
|
|
MachineInstr::mmo_iterator MMOBegin,
|
|
|
|
MachineInstr::mmo_iterator MMOEnd,
|
|
|
|
SmallVectorImpl<MachineInstr*> &NewMIs) const;
|
2011-09-29 13:10:54 +08:00
|
|
|
|
2016-06-30 08:01:54 +08:00
|
|
|
bool expandPostRAPseudo(MachineInstr &MI) const override;
|
2011-09-29 13:10:54 +08:00
|
|
|
|
2016-11-24 02:33:49 +08:00
|
|
|
/// Check whether the target can fold a load that feeds a subreg operand
|
|
|
|
/// (or a subreg operand that feeds a store).
|
|
|
|
bool isSubregFoldable() const override { return true; }
|
|
|
|
|
2008-01-07 09:35:02 +08:00
|
|
|
/// foldMemoryOperand - If this target supports it, fold a load or store of
|
|
|
|
/// the specified stack slot into the specified machine instruction for the
|
|
|
|
/// specified operand(s). If this is possible, the target should perform the
|
|
|
|
/// folding and return true, otherwise it should return false. If it folds
|
|
|
|
/// the instruction, it is likely that the MachineInstruction the iterator
|
|
|
|
/// references has been changed.
|
2016-06-30 08:01:54 +08:00
|
|
|
MachineInstr *
|
|
|
|
foldMemoryOperandImpl(MachineFunction &MF, MachineInstr &MI,
|
|
|
|
ArrayRef<unsigned> Ops,
|
|
|
|
MachineBasicBlock::iterator InsertPt, int FrameIndex,
|
|
|
|
LiveIntervals *LIS = nullptr) const override;
|
2008-01-07 09:35:02 +08:00
|
|
|
|
|
|
|
/// foldMemoryOperand - Same as the previous version except it allows folding
|
|
|
|
/// of any load and store from / to any address, not just from a specific
|
|
|
|
/// stack slot.
|
2016-06-30 08:01:54 +08:00
|
|
|
MachineInstr *foldMemoryOperandImpl(
|
|
|
|
MachineFunction &MF, MachineInstr &MI, ArrayRef<unsigned> Ops,
|
|
|
|
MachineBasicBlock::iterator InsertPt, MachineInstr &LoadMI,
|
|
|
|
LiveIntervals *LIS = nullptr) const override;
|
2008-01-07 09:35:02 +08:00
|
|
|
|
|
|
|
/// unfoldMemoryOperand - Separate a single instruction which folded a load or
|
|
|
|
/// a store or a load and a store into two or more instruction. If this is
|
|
|
|
/// possible, returns true as well as the new instructions by reference.
|
2016-06-30 08:01:54 +08:00
|
|
|
bool
|
|
|
|
unfoldMemoryOperand(MachineFunction &MF, MachineInstr &MI, unsigned Reg,
|
|
|
|
bool UnfoldLoad, bool UnfoldStore,
|
|
|
|
SmallVectorImpl<MachineInstr *> &NewMIs) const override;
|
2008-01-07 09:35:02 +08:00
|
|
|
|
2014-03-09 15:44:38 +08:00
|
|
|
bool unfoldMemoryOperand(SelectionDAG &DAG, SDNode *N,
|
|
|
|
SmallVectorImpl<SDNode*> &NewNodes) const override;
|
2008-01-07 09:35:02 +08:00
|
|
|
|
|
|
|
/// getOpcodeAfterMemoryUnfold - Returns the opcode of the would be new
|
|
|
|
/// instruction after load / store are unfolded from an instruction of the
|
|
|
|
/// specified opcode. It returns zero if the specified unfolding is not
|
2009-10-31 06:18:41 +08:00
|
|
|
/// possible. If LoadRegIndex is non-null, it is filled in with the operand
|
|
|
|
/// index of the operand which will hold the register holding the loaded
|
|
|
|
/// value.
|
2014-03-09 15:44:38 +08:00
|
|
|
unsigned getOpcodeAfterMemoryUnfold(unsigned Opc,
|
|
|
|
bool UnfoldLoad, bool UnfoldStore,
|
2014-04-28 12:05:08 +08:00
|
|
|
unsigned *LoadRegIndex = nullptr) const override;
|
2011-03-05 14:31:54 +08:00
|
|
|
|
2010-01-22 11:34:51 +08:00
|
|
|
/// areLoadsFromSameBasePtr - This is used by the pre-regalloc scheduler
|
|
|
|
/// to determine if two loads are loading from the same base address. It
|
|
|
|
/// should only return true if the base pointers are the same and the
|
|
|
|
/// only differences between the two addresses are the offset. It also returns
|
|
|
|
/// the offsets by reference.
|
2014-03-09 15:44:38 +08:00
|
|
|
bool areLoadsFromSameBasePtr(SDNode *Load1, SDNode *Load2, int64_t &Offset1,
|
|
|
|
int64_t &Offset2) const override;
|
2010-01-22 11:34:51 +08:00
|
|
|
|
|
|
|
/// shouldScheduleLoadsNear - This is a used by the pre-regalloc scheduler to
|
2011-04-15 13:18:47 +08:00
|
|
|
/// determine (in conjunction with areLoadsFromSameBasePtr) if two loads should
|
2010-01-22 11:34:51 +08:00
|
|
|
/// be scheduled togther. On some targets if two loads are loading from
|
|
|
|
/// addresses in the same cache line, it's better if they are scheduled
|
|
|
|
/// together. This function takes two integers that represent the load offsets
|
|
|
|
/// from the common base address. It returns true if it decides it's desirable
|
|
|
|
/// to schedule the two loads together. "NumLoads" is the number of loads that
|
|
|
|
/// have already been scheduled after Load1.
|
2014-03-09 15:44:38 +08:00
|
|
|
bool shouldScheduleLoadsNear(SDNode *Load1, SDNode *Load2,
|
|
|
|
int64_t Offset1, int64_t Offset2,
|
|
|
|
unsigned NumLoads) const override;
|
2010-01-22 11:34:51 +08:00
|
|
|
|
2016-11-29 04:11:54 +08:00
|
|
|
bool shouldScheduleAdjacent(const MachineInstr &First,
|
|
|
|
const MachineInstr &Second) const override;
|
2013-06-23 17:00:28 +08:00
|
|
|
|
2014-03-09 15:44:38 +08:00
|
|
|
void getNoopForMachoTarget(MCInst &NopInst) const override;
|
2010-04-27 07:37:21 +08:00
|
|
|
|
2014-03-09 15:44:38 +08:00
|
|
|
bool
|
2016-09-15 04:43:16 +08:00
|
|
|
reverseBranchCondition(SmallVectorImpl<MachineOperand> &Cond) const override;
|
Teach the code generator that shrd/shld is commutable if it has an immediate.
This allows us to generate this:
foo:
mov %EAX, DWORD PTR [%ESP + 4]
mov %EDX, DWORD PTR [%ESP + 8]
shld %EDX, %EDX, 2
shl %EAX, 2
ret
instead of this:
foo:
mov %EAX, DWORD PTR [%ESP + 4]
mov %ECX, DWORD PTR [%ESP + 8]
mov %EDX, %EAX
shrd %EDX, %ECX, 30
shl %EAX, 2
ret
Note the magically transmogrifying immediate.
llvm-svn: 19686
2005-01-19 15:11:01 +08:00
|
|
|
|
2009-02-07 01:17:30 +08:00
|
|
|
/// isSafeToMoveRegClassDefs - Return true if it's safe to move a machine
|
|
|
|
/// instruction that defines the specified register class.
|
2014-03-09 15:44:38 +08:00
|
|
|
bool isSafeToMoveRegClassDefs(const TargetRegisterClass *RC) const override;
|
2008-10-27 15:14:50 +08:00
|
|
|
|
2014-05-20 16:55:50 +08:00
|
|
|
/// isSafeToClobberEFLAGS - Return true if it's safe insert an instruction tha
|
|
|
|
/// would clobber the EFLAGS condition register. Note the result may be
|
|
|
|
/// conservative. If it cannot definitely determine the safety after visiting
|
|
|
|
/// a few instructions in each direction it assumes it's not safe.
|
|
|
|
bool isSafeToClobberEFLAGS(MachineBasicBlock &MBB,
|
|
|
|
MachineBasicBlock::iterator I) const;
|
|
|
|
|
2015-08-27 04:36:52 +08:00
|
|
|
/// True if MI has a condition code def, e.g. EFLAGS, that is
|
|
|
|
/// not marked dead.
|
2016-06-30 08:01:54 +08:00
|
|
|
bool hasLiveCondCodeDef(MachineInstr &MI) const;
|
2015-08-27 04:36:52 +08:00
|
|
|
|
2008-09-30 08:58:23 +08:00
|
|
|
/// getGlobalBaseReg - Return a virtual register initialized with the
|
|
|
|
/// the global base register value. Output instructions required to
|
|
|
|
/// initialize the register in the function entry block, if necessary.
|
2008-09-24 02:22:58 +08:00
|
|
|
///
|
2008-09-30 08:58:23 +08:00
|
|
|
unsigned getGlobalBaseReg(MachineFunction *MF) const;
|
2008-09-24 02:22:58 +08:00
|
|
|
|
2011-09-28 06:57:18 +08:00
|
|
|
std::pair<uint16_t, uint16_t>
|
2016-06-30 08:01:54 +08:00
|
|
|
getExecutionDomain(const MachineInstr &MI) const override;
|
2010-03-30 07:24:21 +08:00
|
|
|
|
2016-06-30 08:01:54 +08:00
|
|
|
void setExecutionDomain(MachineInstr &MI, unsigned Domain) const override;
|
2010-03-26 01:25:00 +08:00
|
|
|
|
2014-03-09 15:44:38 +08:00
|
|
|
unsigned
|
2016-06-30 08:01:54 +08:00
|
|
|
getPartialRegUpdateClearance(const MachineInstr &MI, unsigned OpNum,
|
|
|
|
const TargetRegisterInfo *TRI) const override;
|
|
|
|
unsigned getUndefRegClearance(const MachineInstr &MI, unsigned &OpNum,
|
2014-03-09 15:44:38 +08:00
|
|
|
const TargetRegisterInfo *TRI) const override;
|
2016-06-30 08:01:54 +08:00
|
|
|
void breakPartialRegDependency(MachineInstr &MI, unsigned OpNum,
|
2014-03-09 15:44:38 +08:00
|
|
|
const TargetRegisterInfo *TRI) const override;
|
2011-11-15 09:15:30 +08:00
|
|
|
|
2016-06-30 08:01:54 +08:00
|
|
|
MachineInstr *foldMemoryOperandImpl(MachineFunction &MF, MachineInstr &MI,
|
implement rdar://6653118 - fastisel should fold loads where possible.
Since mem2reg isn't run at -O0, we get a ton of reloads from the stack,
for example, before, this code:
int foo(int x, int y, int z) {
return x+y+z;
}
used to compile into:
_foo: ## @foo
subq $12, %rsp
movl %edi, 8(%rsp)
movl %esi, 4(%rsp)
movl %edx, (%rsp)
movl 8(%rsp), %edx
movl 4(%rsp), %esi
addl %edx, %esi
movl (%rsp), %edx
addl %esi, %edx
movl %edx, %eax
addq $12, %rsp
ret
Now we produce:
_foo: ## @foo
subq $12, %rsp
movl %edi, 8(%rsp)
movl %esi, 4(%rsp)
movl %edx, (%rsp)
movl 8(%rsp), %edx
addl 4(%rsp), %edx ## Folded load
addl (%rsp), %edx ## Folded load
movl %edx, %eax
addq $12, %rsp
ret
Fewer instructions and less register use = faster compiles.
llvm-svn: 113102
2010-09-05 10:18:34 +08:00
|
|
|
unsigned OpNum,
|
2015-02-28 20:04:00 +08:00
|
|
|
ArrayRef<MachineOperand> MOs,
|
2015-06-09 04:09:58 +08:00
|
|
|
MachineBasicBlock::iterator InsertPt,
|
2014-10-21 06:14:22 +08:00
|
|
|
unsigned Size, unsigned Alignment,
|
|
|
|
bool AllowCommute) const;
|
2010-10-20 02:58:51 +08:00
|
|
|
|
2014-03-09 15:44:38 +08:00
|
|
|
bool isHighLatencyDef(int opc) const override;
|
2011-03-05 16:00:22 +08:00
|
|
|
|
2015-06-13 11:42:11 +08:00
|
|
|
bool hasHighOperandLatency(const TargetSchedModel &SchedModel,
|
2010-10-20 02:58:51 +08:00
|
|
|
const MachineRegisterInfo *MRI,
|
2016-06-30 08:01:54 +08:00
|
|
|
const MachineInstr &DefMI, unsigned DefIdx,
|
|
|
|
const MachineInstr &UseMI,
|
2014-03-09 15:44:38 +08:00
|
|
|
unsigned UseIdx) const override;
|
2016-06-30 08:01:54 +08:00
|
|
|
|
2015-06-11 04:32:21 +08:00
|
|
|
bool useMachineCombiner() const override {
|
|
|
|
return true;
|
|
|
|
}
|
2015-09-21 23:09:11 +08:00
|
|
|
|
|
|
|
bool isAssociativeAndCommutative(const MachineInstr &Inst) const override;
|
|
|
|
|
|
|
|
bool hasReassociableOperands(const MachineInstr &Inst,
|
|
|
|
const MachineBasicBlock *MBB) const override;
|
|
|
|
|
|
|
|
void setSpecialOperandAttr(MachineInstr &OldMI1, MachineInstr &OldMI2,
|
|
|
|
MachineInstr &NewMI1,
|
|
|
|
MachineInstr &NewMI2) const override;
|
2015-06-11 04:32:21 +08:00
|
|
|
|
2012-07-07 01:36:20 +08:00
|
|
|
/// analyzeCompare - For a comparison instruction, return the source registers
|
|
|
|
/// in SrcReg and SrcReg2 if having two register operands, and the value it
|
|
|
|
/// compares against in CmpValue. Return true if the comparison instruction
|
|
|
|
/// can be analyzed.
|
2016-06-30 08:01:54 +08:00
|
|
|
bool analyzeCompare(const MachineInstr &MI, unsigned &SrcReg,
|
2014-03-09 15:44:38 +08:00
|
|
|
unsigned &SrcReg2, int &CmpMask,
|
|
|
|
int &CmpValue) const override;
|
2012-07-07 01:36:20 +08:00
|
|
|
|
|
|
|
/// optimizeCompareInstr - Check if there exists an earlier instruction that
|
|
|
|
/// operates on the same source operands and sets flags in the same way as
|
|
|
|
/// Compare; remove Compare if possible.
|
2016-06-30 08:01:54 +08:00
|
|
|
bool optimizeCompareInstr(MachineInstr &CmpInstr, unsigned SrcReg,
|
2014-03-09 15:44:38 +08:00
|
|
|
unsigned SrcReg2, int CmpMask, int CmpValue,
|
|
|
|
const MachineRegisterInfo *MRI) const override;
|
2012-07-07 01:36:20 +08:00
|
|
|
|
2012-08-02 08:56:42 +08:00
|
|
|
/// optimizeLoadInstr - Try to remove the load by folding it to a register
|
|
|
|
/// operand at the use. We fold the load instructions if and only if the
|
2012-08-03 03:37:32 +08:00
|
|
|
/// def and use are in the same BB. We only look at one load and see
|
|
|
|
/// whether it can be folded into MI. FoldAsLoadDefReg is the virtual register
|
|
|
|
/// defined by the load we are trying to fold. DefMI returns the machine
|
|
|
|
/// instruction that defines FoldAsLoadDefReg, and the function returns
|
|
|
|
/// the machine instruction generated due to folding.
|
2016-06-30 08:01:54 +08:00
|
|
|
MachineInstr *optimizeLoadInstr(MachineInstr &MI,
|
2014-03-09 15:44:38 +08:00
|
|
|
const MachineRegisterInfo *MRI,
|
|
|
|
unsigned &FoldAsLoadDefReg,
|
|
|
|
MachineInstr *&DefMI) const override;
|
2012-08-02 08:56:42 +08:00
|
|
|
|
2015-08-06 08:44:07 +08:00
|
|
|
std::pair<unsigned, unsigned>
|
|
|
|
decomposeMachineOperandsTargetFlags(unsigned TF) const override;
|
|
|
|
|
|
|
|
ArrayRef<std::pair<unsigned, const char *>>
|
|
|
|
getSerializableDirectMachineOperandTargetFlags() const override;
|
|
|
|
|
2016-09-01 09:03:22 +08:00
|
|
|
bool isTailCall(const MachineInstr &Inst) const override;
|
|
|
|
|
2015-09-29 04:33:22 +08:00
|
|
|
protected:
|
|
|
|
/// Commutes the operands in the given instruction by changing the operands
|
|
|
|
/// order and/or changing the instruction's opcode and/or the immediate value
|
|
|
|
/// operand.
|
|
|
|
///
|
|
|
|
/// The arguments 'CommuteOpIdx1' and 'CommuteOpIdx2' specify the operands
|
|
|
|
/// to be commuted.
|
|
|
|
///
|
|
|
|
/// Do not call this method for a non-commutable instruction or
|
|
|
|
/// non-commutable operands.
|
|
|
|
/// Even though the instruction is commutable, the method may still
|
|
|
|
/// fail to commute the operands, null pointer is returned in such cases.
|
2016-06-30 08:01:54 +08:00
|
|
|
MachineInstr *commuteInstructionImpl(MachineInstr &MI, bool NewMI,
|
2015-09-29 04:33:22 +08:00
|
|
|
unsigned CommuteOpIdx1,
|
|
|
|
unsigned CommuteOpIdx2) const override;
|
|
|
|
|
2008-01-07 09:35:02 +08:00
|
|
|
private:
|
2016-06-30 08:01:54 +08:00
|
|
|
MachineInstr *convertToThreeAddressWithLEA(unsigned MIOpc,
|
|
|
|
MachineFunction::iterator &MFI,
|
|
|
|
MachineInstr &MI,
|
|
|
|
LiveVariables *LV) const;
|
2009-12-11 14:01:48 +08:00
|
|
|
|
2015-11-05 04:48:09 +08:00
|
|
|
/// Handles memory folding for special case instructions, for instance those
|
|
|
|
/// requiring custom manipulation of the address.
|
2016-06-30 08:01:54 +08:00
|
|
|
MachineInstr *foldMemoryOperandCustom(MachineFunction &MF, MachineInstr &MI,
|
2015-11-05 04:48:09 +08:00
|
|
|
unsigned OpNum,
|
|
|
|
ArrayRef<MachineOperand> MOs,
|
|
|
|
MachineBasicBlock::iterator InsertPt,
|
|
|
|
unsigned Size, unsigned Align) const;
|
|
|
|
|
2009-11-13 04:55:29 +08:00
|
|
|
/// isFrameOperand - Return true and the FrameIndex if the specified
|
|
|
|
/// operand and follow operands form a reference to the stack frame.
|
2016-06-30 08:01:54 +08:00
|
|
|
bool isFrameOperand(const MachineInstr &MI, unsigned int Op,
|
2009-11-13 04:55:29 +08:00
|
|
|
int &FrameIndex) const;
|
2016-09-22 11:00:50 +08:00
|
|
|
|
|
|
|
/// Returns true iff the routine could find two commutable operands in the
|
|
|
|
/// given machine instruction with 3 vector inputs.
|
|
|
|
/// The 'SrcOpIdx1' and 'SrcOpIdx2' are INPUT and OUTPUT arguments. Their
|
|
|
|
/// input values can be re-defined in this method only if the input values
|
|
|
|
/// are not pre-defined, which is designated by the special value
|
|
|
|
/// 'CommuteAnyOperandIndex' assigned to it.
|
|
|
|
/// If both of indices are pre-defined and refer to some operands, then the
|
|
|
|
/// method simply returns true if the corresponding operands are commutable
|
|
|
|
/// and returns false otherwise.
|
|
|
|
///
|
|
|
|
/// For example, calling this method this way:
|
|
|
|
/// unsigned Op1 = 1, Op2 = CommuteAnyOperandIndex;
|
|
|
|
/// findThreeSrcCommutedOpIndices(MI, Op1, Op2);
|
|
|
|
/// can be interpreted as a query asking to find an operand that would be
|
|
|
|
/// commutable with the operand#1.
|
|
|
|
bool findThreeSrcCommutedOpIndices(const MachineInstr &MI,
|
|
|
|
unsigned &SrcOpIdx1,
|
|
|
|
unsigned &SrcOpIdx2) const;
|
2002-10-26 06:55:53 +08:00
|
|
|
};
|
|
|
|
|
2015-06-23 17:49:53 +08:00
|
|
|
} // End llvm namespace
|
2003-11-12 06:41:34 +08:00
|
|
|
|
2002-10-26 06:55:53 +08:00
|
|
|
#endif
|