path: root/lib/Target/X86/MCTargetDesc/X86BaseInfo.h

//===-- X86BaseInfo.h - Top level definitions for X86 -------- --*- 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 small standalone helper functions and enum definitions for
// the X86 target useful for the compiler back-end and the MC libraries.
// As such, it deliberately does not include references to LLVM core
// code gen types, passes, etc..
//
//===----------------------------------------------------------------------===//

#ifndef X86BASEINFO_H
#define X86BASEINFO_H

#include "X86MCTargetDesc.h"
#include "llvm/Support/DataTypes.h"
#include "llvm/Support/ErrorHandling.h"
#include "llvm/MC/MCInstrInfo.h"

namespace llvm {

namespace X86 {
  // Enums for memory operand decoding.  Each memory operand is represented with
  // a 5 operand sequence in the form:
  //   [BaseReg, ScaleAmt, IndexReg, Disp, Segment]
  // These enums help decode this.
  enum {
    AddrBaseReg = 0,
    AddrScaleAmt = 1,
    AddrIndexReg = 2,
    AddrDisp = 3,

    /// AddrSegmentReg - The operand # of the segment in the memory operand.
    AddrSegmentReg = 4,

    /// AddrNumOperands - Total number of operands in a memory reference.
    AddrNumOperands = 5
  };
} // end namespace X86;

/// X86II - This namespace holds all of the target specific flags that
/// instruction info tracks.
///
namespace X86II {
  /// Target Operand Flag enum.
  enum TOF {
    //===------------------------------------------------------------------===//
    // X86 Specific MachineOperand flags.

    MO_NO_FLAG,

    /// MO_GOT_ABSOLUTE_ADDRESS - On a symbol operand, this represents a
    /// relocation of:
    ///    SYMBOL_LABEL + [. - PICBASELABEL]
    MO_GOT_ABSOLUTE_ADDRESS,

    /// MO_PIC_BASE_OFFSET - On a symbol operand this indicates that the
    /// immediate should get the value of the symbol minus the PIC base label:
    ///    SYMBOL_LABEL - PICBASELABEL
    MO_PIC_BASE_OFFSET,

    /// MO_GOT - On a symbol operand this indicates that the immediate is the
    /// offset to the GOT entry for the symbol name from the base of the GOT.
    ///
    /// See the X86-64 ELF ABI supplement for more details.
    ///    SYMBOL_LABEL @GOT
    MO_GOT,

    /// MO_GOTOFF - On a symbol operand this indicates that the immediate is
    /// the offset to the location of the symbol name from the base of the GOT.
    ///
    /// See the X86-64 ELF ABI supplement for more details.
    ///    SYMBOL_LABEL @GOTOFF
    MO_GOTOFF,

    /// MO_GOTPCREL - On a symbol operand this indicates that the immediate is
    /// offset to the GOT entry for the symbol name from the current code
    /// location.
    ///
    /// See the X86-64 ELF ABI supplement for more details.
    ///    SYMBOL_LABEL @GOTPCREL
    MO_GOTPCREL,

    /// MO_PLT - On a symbol operand this indicates that the immediate is
    /// offset to the PLT entry of symbol name from the current code location.
    ///
    /// See the X86-64 ELF ABI supplement for more details.
    ///    SYMBOL_LABEL @PLT
    MO_PLT,

    /// MO_TLSGD - On a symbol operand this indicates that the immediate is
    /// the offset of the GOT entry with the TLS index structure that contains
    /// the module number and variable offset for the symbol. Used in the
    /// general dynamic TLS access model.
    ///
    /// See 'ELF Handling for Thread-Local Storage' for more details.
    ///    SYMBOL_LABEL @TLSGD
    MO_TLSGD,

    /// MO_TLSLD - On a symbol operand this indicates that the immediate is
    /// the offset of the GOT entry with the TLS index for the module that
    /// contains the symbol. When this index is passed to a call to
    /// __tls_get_addr, the function will return the base address of the TLS
    /// block for the symbol. Used in the x86-64 local dynamic TLS access model.
    ///
    /// See 'ELF Handling for Thread-Local Storage' for more details.
    ///    SYMBOL_LABEL @TLSLD
    MO_TLSLD,

    /// MO_TLSLDM - On a symbol operand this indicates that the immediate is
    /// the offset of the GOT entry with the TLS index for the module that
    /// contains the symbol. When this index is passed to a call to
    /// ___tls_get_addr, the function will return the base address of the TLS
    /// block for the symbol. Used in the IA32 local dynamic TLS access model.
    ///
    /// See 'ELF Handling for Thread-Local Storage' for more details.
    ///    SYMBOL_LABEL @TLSLDM
    MO_TLSLDM,

    /// MO_GOTTPOFF - On a symbol operand this indicates that the immediate is
    /// the offset of the GOT entry with the thread-pointer offset for the
    /// symbol. Used in the x86-64 initial exec TLS access model.
    ///
    /// See 'ELF Handling for Thread-Local Storage' for more details.
    ///    SYMBOL_LABEL @GOTTPOFF
    MO_GOTTPOFF,

    /// MO_INDNTPOFF - On a symbol operand this indicates that the immediate is
    /// the absolute address of the GOT entry with the negative thread-pointer
    /// offset for the symbol. Used in the non-PIC IA32 initial exec TLS access
    /// model.
    ///
    /// See 'ELF Handling for Thread-Local Storage' for more details.
    ///    SYMBOL_LABEL @INDNTPOFF
    MO_INDNTPOFF,

    /// MO_TPOFF - On a symbol operand this indicates that the immediate is
    /// the thread-pointer offset for the symbol. Used in the x86-64 local
    /// exec TLS access model.
    ///
    /// See 'ELF Handling for Thread-Local Storage' for more details.
    ///    SYMBOL_LABEL @TPOFF
    MO_TPOFF,

    /// MO_DTPOFF - On a symbol operand this indicates that the immediate is
    /// the offset of the GOT entry with the TLS offset of the symbol. Used
    /// in the local dynamic TLS access model.
    ///
    /// See 'ELF Handling for Thread-Local Storage' for more details.
    ///    SYMBOL_LABEL @DTPOFF
    MO_DTPOFF,

    /// MO_NTPOFF - On a symbol operand this indicates that the immediate is
    /// the negative thread-pointer offset for the symbol. Used in the IA32
    /// local exec TLS access model.
    ///
    /// See 'ELF Handling for Thread-Local Storage' for more details.
    ///    SYMBOL_LABEL @NTPOFF
    MO_NTPOFF,

    /// MO_GOTNTPOFF - On a symbol operand this indicates that the immediate is
    /// the offset of the GOT entry with the negative thread-pointer offset for
    /// the symbol. Used in the PIC IA32 initial exec TLS access model.
    ///
    /// See 'ELF Handling for Thread-Local Storage' for more details.
    ///    SYMBOL_LABEL @GOTNTPOFF
    MO_GOTNTPOFF,

    /// MO_DLLIMPORT - On a symbol operand "FOO", this indicates that the
    /// reference is actually to the "__imp_FOO" symbol.  This is used for
    /// dllimport linkage on windows.
    MO_DLLIMPORT,

    /// MO_DARWIN_STUB - On a symbol operand "FOO", this indicates that the
    /// reference is actually to the "FOO$stub" symbol.  This is used for calls
    /// and jumps to external functions on Tiger and earlier.
    MO_DARWIN_STUB,

    /// MO_DARWIN_NONLAZY - On a symbol operand "FOO", this indicates that the
    /// reference is actually to the "FOO$non_lazy_ptr" symbol, which is a
    /// non-PIC-base-relative reference to a non-hidden dyld lazy pointer stub.
    MO_DARWIN_NONLAZY,

    /// MO_DARWIN_NONLAZY_PIC_BASE - On a symbol operand "FOO", this indicates
    /// that the reference is actually to "FOO$non_lazy_ptr - PICBASE", which is
    /// a PIC-base-relative reference to a non-hidden dyld lazy pointer stub.
    MO_DARWIN_NONLAZY_PIC_BASE,

    /// MO_DARWIN_HIDDEN_NONLAZY_PIC_BASE - On a symbol operand "FOO", this
    /// indicates that the reference is actually to "FOO$non_lazy_ptr -PICBASE",
    /// which is a PIC-base-relative reference to a hidden dyld lazy pointer
    /// stub.
    MO_DARWIN_HIDDEN_NONLAZY_PIC_BASE,

    /// MO_TLVP - On a symbol operand this indicates that the immediate is
    /// some TLS offset.
    ///
    /// This is the TLS offset for the Darwin TLS mechanism.
    MO_TLVP,

    /// MO_TLVP_PIC_BASE - On a symbol operand this indicates that the immediate
    /// is some TLS offset from the picbase.
    ///
    /// This is the 32-bit TLS offset for Darwin TLS in PIC mode.
    MO_TLVP_PIC_BASE,

    /// MO_SECREL - On a symbol operand this indicates that the immediate is
    /// the offset from beginning of section.
    ///
    /// This is the TLS offset for the COFF/Windows TLS mechanism.
    MO_SECREL
  };

  enum {
    //===------------------------------------------------------------------===//
    // Instruction encodings.  These are the standard/most common forms for X86
    // instructions.
    //

    // PseudoFrm - This represents an instruction that is a pseudo instruction
    // or one that has not been implemented yet.  It is illegal to code generate
    // it, but tolerated for intermediate implementation stages.
    Pseudo         = 0,

    /// Raw - This form is for instructions that don't have any operands, so
    /// they are just a fixed opcode value, like 'leave'.
    RawFrm         = 1,

    /// AddRegFrm - This form is used for instructions like 'push r32' that have
    /// their one register operand added to their opcode.
    AddRegFrm      = 2,

    /// MRMDestReg - This form is used for instructions that use the Mod/RM byte
    /// to specify a destination, which in this case is a register.
    ///
    MRMDestReg     = 3,

    /// MRMDestMem - This form is used for instructions that use the Mod/RM byte
    /// to specify a destination, which in this case is memory.
    ///
    MRMDestMem     = 4,

    /// MRMSrcReg - This form is used for instructions that use the Mod/RM byte
    /// to specify a source, which in this case is a register.
    ///
    MRMSrcReg      = 5,

    /// MRMSrcMem - This form is used for instructions that use the Mod/RM byte
    /// to specify a source, which in this case is memory.
    ///
    MRMSrcMem      = 6,

    /// MRM[0-7][rm] - These forms are used to represent instructions that use
    /// a Mod/RM byte, and use the middle field to hold extended opcode
    /// information.  In the intel manual these are represented as /0, /1, ...
    ///

    // First, instructions that operate on a register r/m operand...
    MRM0r = 16,  MRM1r = 17,  MRM2r = 18,  MRM3r = 19, // Format /0 /1 /2 /3
    MRM4r = 20,  MRM5r = 21,  MRM6r = 22,  MRM7r = 23, // Format /4 /5 /6 /7

    // Next, instructions that operate on a memory r/m operand...
    MRM0m = 24,  MRM1m = 25,  MRM2m = 26,  MRM3m = 27, // Format /0 /1 /2 /3
    MRM4m = 28,  MRM5m = 29,  MRM6m = 30,  MRM7m = 31, // Format /4 /5 /6 /7

    // MRMInitReg - This form is used for instructions whose source and
    // destinations are the same register.
    MRMInitReg = 32,

    //// MRM_XX - A mod/rm byte of exactly 0xXX.
    MRM_C1 = 33, MRM_C2 = 34, MRM_C3 = 35, MRM_C4 = 36,
    MRM_C8 = 37, MRM_C9 = 38, MRM_CA = 39, MRM_CB = 40,
    MRM_E8 = 41, MRM_F0 = 42, MRM_F8 = 45, MRM_F9 = 46,
    MRM_D0 = 47, MRM_D1 = 48, MRM_D4 = 49, MRM_D5 = 50,
    MRM_D6 = 51, MRM_D8 = 52, MRM_D9 = 53, MRM_DA = 54,
    MRM_DB = 55, MRM_DC = 56, MRM_DD = 57, MRM_DE = 58,
    MRM_DF = 59,

    /// RawFrmImm8 - This is used for the ENTER instruction, which has two
    /// immediates, the first of which is a 16-bit immediate (specified by
    /// the imm encoding) and the second is a 8-bit fixed value.
    RawFrmImm8 = 43,

    /// RawFrmImm16 - This is used for CALL FAR instructions, which have two
    /// immediates, the first of which is a 16 or 32-bit immediate (specified by
    /// the imm encoding) and the second is a 16-bit fixed value.  In the AMD
    /// manual, this operand is described as pntr16:32 and pntr16:16
    RawFrmImm16 = 44,

    FormMask       = 63,

    //===------------------------------------------------------------------===//
    // Actual flags...

    // OpSize - Set if this instruction requires an operand size prefix (0x66),
    // which most often indicates that the instruction operates on 16 bit data
    // instead of 32 bit data.
    OpSize      = 1 << 6,

    // AsSize - Set if this instruction requires an operand size prefix (0x67),
    // which most often indicates that the instruction address 16 bit address
    // instead of 32 bit address (or 32 bit address in 64 bit mode).
    AdSize      = 1 << 7,

    //===------------------------------------------------------------------===//
    // Op0Mask - There are several prefix bytes that are used to form two byte
    // opcod