assembler-x64.h

Go to the documentation of this file.

// Copyright (c) 1994-2006 Sun Microsystems Inc.
// All Rights Reserved.
//
// Redistribution and use in source and binary forms, with or without
// modification, are permitted provided that the following conditions are
// met:
//
// - Redistributions of source code must retain the above copyright notice,
// this list of conditions and the following disclaimer.
//
// - Redistribution in binary form must reproduce the above copyright
// notice, this list of conditions and the following disclaimer in the
// documentation and/or other materials provided with the distribution.
//
// - Neither the name of Sun Microsystems or the names of contributors may
// be used to endorse or promote products derived from this software without
// specific prior written permission.
//
// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS
// IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
// THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
// PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
// CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
// EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
// PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
// PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
// LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
// NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
// SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 
// The original source code covered by the above license above has been
// modified significantly by Google Inc.
// Copyright 2012 the V8 project authors. All rights reserved.
 
// A lightweight X64 Assembler.
 
#ifndef V8_CODEGEN_X64_ASSEMBLER_X64_H_
#define V8_CODEGEN_X64_ASSEMBLER_X64_H_
 
#include <deque>
#include <map>
#include <memory>
#include <vector>
 
#include "src/base/export-template.h"
#include "src/codegen/assembler.h"
#include "src/codegen/cpu-features.h"
#include "src/codegen/label.h"
#include "src/codegen/x64/builtin-jump-table-info-x64.h"
#include "src/codegen/x64/constants-x64.h"
#include "src/codegen/x64/fma-instr.h"
#include "src/codegen/x64/register-x64.h"
#include "src/codegen/x64/sse-instr.h"
#include "src/objects/smi.h"
 
#if defined(V8_OS_WIN_X64)
#include "src/diagnostics/unwinding-info-win64.h"
#endif
 
namespace v8 {
namespace internal {
 
class SafepointTableBuilder;
class MaglevSafepointTableBuilder;
 
// Utility functions
 
enum Condition : int {
  overflow = 0,
  no_overflow = 1,
  below = 2,
  above_equal = 3,
  equal = 4,
  not_equal = 5,
  below_equal = 6,
  above = 7,
  negative = 8,
  positive = 9,
  parity_even = 10,
  parity_odd = 11,
  less = 12,
  greater_equal = 13,
  less_equal = 14,
  greater = 15,
 
  // aliases
  carry = below,
  not_carry = above_equal,
  zero = equal,
  not_zero = not_equal,
  sign = negative,
  not_sign = positive,
 
  // Unified cross-platform condition names/aliases.
  kEqual = equal,
  kNotEqual = not_equal,
  kLessThan = less,
  kGreaterThan = greater,
  kLessThanEqual = less_equal,
  kGreaterThanEqual = greater_equal,
  kUnsignedLessThan = below,
  kUnsignedGreaterThan = above,
  kUnsignedLessThanEqual = below_equal,
  kUnsignedGreaterThanEqual = above_equal,
  kOverflow = overflow,
  kNoOverflow = no_overflow,
  kZero = equal,
  kNotZero = not_equal,
};
 
// Returns the equivalent of !cc.
inline Condition NegateCondition(Condition cc) {
  return static_cast<Condition>(cc ^ 1);
}
 
enum RoundingMode {
  kRoundToNearest = 0x0,
  kRoundDown = 0x1,
  kRoundUp = 0x2,
  kRoundToZero = 0x3
};
 
// -----------------------------------------------------------------------------
// Machine instruction Immediates
 
class Immediate {
 public:
  explicit constexpr Immediate(int32_t value) : value_(value) {}
  explicit constexpr Immediate(int32_t value, RelocInfo::Mode rmode)
      : value_(value), rmode_(rmode) {}
  explicit Immediate(Tagged<Smi> value)
      : value_(static_cast<int32_t>(static_cast<intptr_t>(value.ptr()))) {
    DCHECK(SmiValuesAre31Bits());  // Only available for 31-bit SMI.
  }
 
  int32_t value() const { return value_; }
  RelocInfo::Mode rmode() const { return rmode_; }
 
 private:
  const int32_t value_;
  const RelocInfo::Mode rmode_ = RelocInfo::NO_INFO;
 
  friend class Assembler;
};
ASSERT_TRIVIALLY_COPYABLE(Immediate);
static_assert(sizeof(Immediate) <= kSystemPointerSize,
              "Immediate must be small enough to pass it by value");
 
class Immediate64 {
 public:
  explicit constexpr Immediate64(int64_t value) : value_(value) {}
  explicit constexpr Immediate64(int64_t value, RelocInfo::Mode rmode)
      : value_(value), rmode_(rmode) {}
  explicit constexpr Immediate64(Address value, RelocInfo::Mode rmode)
      : value_(static_cast<int64_t>(value)), rmode_(rmode) {}
 
 private:
  const int64_t value_;
  const RelocInfo::Mode rmode_ = RelocInfo::NO_INFO;
 
  friend class Assembler;
};
 
// -----------------------------------------------------------------------------
// Machine instruction Operands
 
enum ScaleFactor : int8_t {
  times_1 = 0,
  times_2 = 1,
  times_4 = 2,
  times_8 = 3,
  times_int_size = times_4,
 
  times_half_system_pointer_size = times_4,
  times_system_pointer_size = times_8,
  times_tagged_size = (kTaggedSize == 8) ? times_8 : times_4,
  times_external_pointer_size = V8_ENABLE_SANDBOX_BOOL ? times_4 : times_8,
};
 
class V8_EXPORT_PRIVATE Operand {
 public:
  struct LabelOperand {
    // The first two fields are shared in {LabelOperand} and {MemoryOperand},
    // but cannot be pulled out of the union, because otherwise the compiler
    // introduces additional padding between them and the union, increasing the
    // size unnecessarily.
    bool is_label_operand = true;
    uint8_t rex = 0;  // REX prefix, always zero for label operands.
 
    int8_t addend;  // Used for rip + offset + addend operands.
    Label* label;
  };
 
  struct MemoryOperand {
    bool is_label_operand = false;
    uint8_t rex = 0;  // REX prefix.
 
    // Register (1 byte) + SIB (0 or 1 byte) + displacement (0, 1, or 4 byte).
    uint8_t buf[6] = {0};
    // Number of bytes of buf in use.
    // We must keep {len} and {buf} together for the compiler to elide the
    // stack canary protection code.
    size_t len = 1;
  };
 
  // Assert that the shared {is_label_operand} and {rex} fields have the same
  // type and offset in both union variants.
  static_assert(std::is_same<decltype(LabelOperand::is_label_operand),
                             decltype(MemoryOperand::is_label_operand)>::value);
  static_assert(offsetof(LabelOperand, is_label_operand) ==
                offsetof(MemoryOperand, is_label_operand));
  static_assert(std::is_same<decltype(LabelOperand::rex),
                             decltype(MemoryOperand::rex)>::value);
  static_assert(offsetof(LabelOperand, rex) == offsetof(MemoryOperand, rex));
 
  static_assert(sizeof(MemoryOperand::len) == kSystemPointerSize,
                "Length must have native word size to avoid spurious reloads "
                "after writing it.");
  static_assert(offsetof(MemoryOperand, len) % kSystemPointerSize == 0,
                "Length must be aligned for fast access.");
 
  // [base + disp/r]
  V8_INLINE constexpr Operand(Register base, int32_t disp) {
    if (base == rsp || base == r12) {
      // SIB byte is needed to encode (rsp + offset) or (r12 + offset).
      set_sib(times_1, rsp, base);
    }
 
    if (disp == 0 && base != rbp && base != r13) {
      set_modrm(0, base);
    } else if (is_int8(disp)) {
      set_modrm(1, base);
      set_disp8(disp);
    } else {
      set_modrm(2, base);
      set_disp32(disp);
    }
  }
 
  // [base + index*scale + disp/r]
  V8_INLINE Operand(Register base, Register index, ScaleFactor scale,
                    int32_t disp) {
    DCHECK(index != rsp);
    set_sib(scale, index, base);
    if (disp == 0 && base != rbp && base != r13) {
      // This call to set_modrm doesn't overwrite the REX.B (or REX.X) bits
      // possibly set by set_sib.
      set_modrm(0, rsp);
    } else if (is_int8(disp)) {
      set_modrm(1, rsp);
      set_disp8(disp);
    } else {
      set_modrm(2, rsp);
      set_disp32(disp);
    }
  }
 
  // [index*scale + disp/r]
  V8_INLINE Operand(Register index, ScaleFactor scale, int32_t disp) {
    // The encoding generated by this constructor is longer than the
    // {register, displacement} constructor above. Hence only use this if the
    // scale factor is >1.
    // We could dynamically check this and do what the other constructor does,
    // but that adds unnecessary checks to a very commonly used constructor in
    // the Assembler, which should be as fast as possible.
    DCHECK_NE(ScaleFactor::times_1, scale);
 
    DCHECK(index != rsp);
    set_modrm(0, rsp);
    set_sib(scale, index, rbp);
    set_disp32(disp);
  }
 
  // Offset from existing memory operand.
  // Offset is added to existing displacement as 32-bit signed values and
  // this must not overflow.
  Operand(Operand base, int32_t offset);
 
  // [rip + disp/r]
  V8_INLINE explicit Operand(Label* label, int addend = 0) {
    DCHECK_NOT_NULL(label);
    DCHECK(addend == 0 || (is_int8(addend) && label->is_bound()));
    label_ = {};
    label_.label = label;
    label_.addend = addend;
  }
 
  Operand(const Operand&) V8_NOEXCEPT = default;
  Operand& operator=(const Operand&) V8_NOEXCEPT = default;
 
  V8_INLINE constexpr bool is_label_operand() const {
    // Since this field is in the common initial sequence of {label_} and
    // {memory_}, the access is valid regardless of the active union member.
    return memory_.is_label_operand;
  }
 
  V8_INLINE constexpr uint8_t rex() const {
    // Since both fields are in the common initial sequence of {label_} and
    // {memory_}, the access is valid regardless of the active union member.
    // Label operands always have a REX prefix of zero.
    V8_ASSUME(!memory_.is_label_operand || memory_.rex == 0);
    return memory_.rex;
  }
 
  V8_INLINE const MemoryOperand& memory() const {
    DCHECK(!is_label_operand());
    return memory_;
  }
 
  V8_INLINE const LabelOperand& label() const {
    DCHECK(is_label_operand());
    return label_;
  }
 
  // Checks whether either base or index register is the given register.
  // Does not check the "reg" part of the Operand.
  bool AddressUsesRegister(Register reg) const;
 
 private:
  V8_INLINE constexpr void set_modrm(int mod, Register rm_reg) {
    DCHECK(!is_label_operand());
    DCHECK(is_uint2(mod));
    memory_.buf[0] = mod << 6 | rm_reg.low_bits();
    // Set REX.B to the high bit of rm.code().
    memory_.rex |= rm_reg.high_bit();
  }
 
  V8_INLINE constexpr void set_sib(ScaleFactor scale, Register index,
                                   Register base) {
    V8_ASSUME(memory_.len == 1);
    DCHECK(is_uint2(scale));
    // Use SIB with no index register only for base rsp or r12. Otherwise we
    // would skip the SIB byte entirely.
    DCHECK(index != rsp || base == rsp || base == r12);
    memory_.buf[1] = (scale << 6) | (index.low_bits() << 3) | base.low_bits();
    memory_.rex |= index.high_bit() << 1 | base.high_bit();
    memory_.len = 2;
  }
 
  V8_INLINE constexpr void set_disp8(int disp) {
    V8_ASSUME(memory_.len == 1 || memory_.len == 2);
    DCHECK(is_int8(disp));
    memory_.buf[memory_.len] = disp;
    memory_.len += sizeof(int8_t);
  }
 
  V8_INLINE void set_disp32(int disp) {
    V8_ASSUME(memory_.len == 1 || memory_.len == 2);
    Address p = reinterpret_cast<Address>(&memory_.buf[memory_.len]);
    WriteUnalignedValue(p, disp);
    memory_.len += sizeof(int32_t);
  }
 
  union {
    LabelOperand label_;
    MemoryOperand memory_ = {};
  };
};
 
class V8_EXPORT_PRIVATE Operand256 : public Operand {
 public:
  // [base + disp/r]
  V8_INLINE Operand256(Register base, int32_t disp) : Operand(base, disp) {}
 
  // [base + index*scale + disp/r]
  V8_INLINE Operand256(Register base, Register index, ScaleFactor scale,
                       int32_t disp)
      : Operand(base, index, scale, disp) {}
 
  // [index*scale + disp/r]
  V8_INLINE Operand256(Register index, ScaleFactor scale, int32_t disp)
      : Operand(index, scale, disp) {}
 
  Operand256(const Operand256&) V8_NOEXCEPT = default;
  Operand256& operator=(const Operand256&) V8_NOEXCEPT = default;
 
 private:
  friend class Operand;
};
 
ASSERT_TRIVIALLY_COPYABLE(Operand);
static_assert(sizeof(Operand) <= 2 * kSystemPointerSize,
              "Operand must be small enough to pass it by value");
 
// Support DCHECK_NE in shared code. On x64, an {Operand} is never an alias
// for a register.
inline bool operator!=(Operand op, XMMRegister r) { return true; }
 
#define ASSEMBLER_INSTRUCTION_LIST(V) \
  V(add)                              \
  V(and)                              \
  V(cmp)                              \
  V(cmpxchg)                          \
  V(dec)                              \
  V(idiv)                             \
  V(div)                              \
  V(imul)                             \
  V(inc)                              \
  V(lea)                              \
  V(mov)                              \
  V(movzxb)                           \
  V(movzxw)                           \
  V(not )                             \
  V(or)                               \
  V(repmovs)                          \
  V(sbb)                              \
  V(sub)                              \
  V(test)                             \
  V(xchg)                             \
  V(xor)                              \
  V(aligned_cmp)                      \
  V(aligned_test)
 
// Shift instructions on operands/registers with kInt32Size and kInt64Size.
#define SHIFT_INSTRUCTION_LIST(V) \
  V(rol, 0x0)                     \
  V(ror, 0x1)                     \
  V(rcl, 0x2)                     \
  V(rcr, 0x3)                     \
  V(shl, 0x4)                     \
  V(shr, 0x5)                     \
  V(sar, 0x7)
 
// Partial Constant Pool
// Different from complete constant pool (like arm does), partial constant pool
// only takes effects for shareable constants in order to reduce code size.
// Partial constant pool does not emit constant pool entries at the end of each
// code object. Instead, it keeps the first shareable constant inlined in the
// instructions and uses rip-relative memory loadings for the same constants in
// subsequent instructions. These rip-relative memory loadings will target at
// the position of the first inlined constant. For example:
//
//  REX.W movq r10,0x7f9f75a32c20   ; 10 bytes
//  …
//  REX.W movq r10,0x7f9f75a32c20   ; 10 bytes
//  …
//
// turns into
//
//  REX.W movq r10,0x7f9f75a32c20   ; 10 bytes
//  …
//  REX.W movq r10,[rip+0xffffff96] ; 7 bytes
//  …
 
class ConstPool {
 public:
  explicit ConstPool(Assembler* assm) : assm_(assm) {}
  // Returns true when partial constant pool is valid for this entry.
  bool TryRecordEntry(intptr_t data, RelocInfo::Mode mode);
  bool IsEmpty() const { return entries_.empty(); }
 
  void PatchEntries();
  // Discard any pending pool entries.
  void Clear();
 
 private:
  // Adds a shared entry to entries_. Returns true if this is not the first time
  // we add this entry, false otherwise.
  bool AddSharedEntry(uint64_t data, int offset);
 
  // Check if the instruction is a rip-relative move.
  bool IsMoveRipRelative(Address instr);
 
  Assembler* assm_;
 
  // Values, pc offsets of entries.
  std::multimap<uint64_t, int> entries_;
 
  // Number of bytes taken up by the displacement of rip-relative addressing.
  static constexpr int kRipRelativeDispSize = 4;  // 32-bit displacement.
  // Distance between the address of the displacement in the rip-relative move
  // instruction and the head address of the instruction.
  static constexpr int kMoveRipRelativeDispOffset =
      3;  // REX Opcode ModRM Displacement
  // Distance between the address of the imm64 in the 'movq reg, imm64'
  // instruction and the head address of the instruction.
  static constexpr int kMoveImm64Offset = 2;  // REX Opcode imm64
  // A mask for rip-relative move instruction.
  static constexpr uint32_t kMoveRipRelativeMask = 0x00C7FFFB;
  // The bits for a rip-relative move instruction after mask.
  static constexpr uint32_t kMoveRipRelativeInstr = 0x00058B48;
};
 
class V8_EXPORT_PRIVATE Assembler : public AssemblerBase {
 private:
  // We check before assembling an instruction that there is sufficient
  // space to write an instruction and its relocation information.
  // The relocation writer's position must be kGap bytes above the end of
  // the generated instructions. This leaves enough space for the
  // longest possible x64 instruction, 15 bytes, and the longest possible
  // relocation information encoding, RelocInfoWriter::kMaxLength == 16.
  // (There is a 15 byte limit on x64 instruction length that rules out some
  // otherwise valid instructions.)
  // This allows for a single, fast space check per instruction.
  static constexpr int kGap = 32;
  static_assert(AssemblerBase::kMinimalBufferSize >= 2 * kGap);
 
 public:
  // Create an assembler. Instructions and relocation information are emitted
  // into a buffer, with the instructions starting from the beginning and the
  // relocation information starting from the end of the buffer. See CodeDesc
  // for a detailed comment on the layout (globals.h).
  //
  // If the provided buffer is nullptr, the assembler allocates and grows its
  // own buffer. Otherwise it takes ownership of the provided buffer.
  explicit Assembler(const AssemblerOptions&,
                     std::unique_ptr<AssemblerBuffer> = {});
  // For compatibility with assemblers that require a zone.
  Assembler(const MaybeAssemblerZone&, const AssemblerOptions& options,
            std::unique_ptr<AssemblerBuffer> buffer = {})
      : Assembler(options, std::move(buffer)) {}
 
  ~Assembler() override = default;
 
  // GetCode emits any pending (non-emitted) code and fills the descriptor desc.
  static constexpr int kNoHandlerTable = 0;
  static constexpr SafepointTableBuilderBase* kNoSafepointTable = nullptr;
 
  void GetCode(LocalIsolate* isolate, CodeDesc* desc,
               SafepointTableBuilderBase* safepoint_table_builder,
               int handler_table_offset);
 
  // Convenience wrapper for allocating with an Isolate.
  void GetCode(Isolate* isolate, CodeDesc* desc);
  // Convenience wrapper for code without safepoint or handler tables.
  void GetCode(LocalIsolate* isolate, CodeDesc* desc) {
    GetCode(isolate, desc, kNoSafepointTable, kNoHandlerTable);
  }
 
  void FinalizeJumpOptimizationInfo();
 
  // Unused on this architecture.
  void MaybeEmitOutOfLineConstantPool() {}
 
  // Read/Modify the code target in the relative branch/call instruction at pc.
  // On the x64 architecture, we use relative jumps with a 32-bit displacement
  // to jump to other InstructionStream objects in the InstructionStream space
  // in the heap. Jumps to C functions are done indirectly through a 64-bit
  // register holding the absolute address of the target. These functions
  // convert between absolute Addresses of InstructionStream objects and the
  // relative displacements stored in the code. The isolate argument is unused
  // (and may be nullptr) when skipping flushing.
  static inline Address target_address_at(Address pc, Address constant_pool);
  static inline void set_target_address_at(
      Address pc, Address constant_pool, Address target,
      WritableJitAllocation* writable_jit_allocation = nullptr,
      ICacheFlushMode icache_flush_mode = FLUSH_ICACHE_IF_NEEDED);
  static inline int32_t relative_target_offset(Address target, Address pc);
 
  // During code generation builtin targets in PC-relative call/jump
  // instructions are temporarily encoded as builtin ID until the generated
  // code is moved into the code space.
  static inline Builtin target_builtin_at(Address pc);
 
  // Get the size of the special target encoded at 'instruction_payload'.
  inline static int deserialization_special_target_size(
      Address instruction_payload);
 
  // This sets the internal reference at the pc.
  inline static void deserialization_set_target_internal_reference_at(
      Address pc, Address target, WritableJitAllocation& jit_allocation,
      RelocInfo::Mode mode = RelocInfo::INTERNAL_REFERENCE);
 
  inline DirectHandle<Code> code_target_object_handle_at(Address pc);
  inline DirectHandle<HeapObject> compressed_embedded_object_handle_at(
      Address pc);
 
  // Read/modify the uint32 constant used at pc.
  static inline uint32_t uint32_constant_at(Address pc, Address constant_pool);
  static inline void set_uint32_constant_at(
      Address pc, Address constant_pool, uint32_t new_constant,
      WritableJitAllocation* jit_allocation = nullptr,
      ICacheFlushMode icache_flush_mode = FLUSH_ICACHE_IF_NEEDED);
 
  // Number of bytes taken up by the branch target in the code.
  static constexpr int kSpecialTargetSize = 4;  // 32-bit displacement.
 
  // One byte opcode for test eax,0xXXXXXXXX.
  static constexpr uint8_t kTestEaxByte = 0xA9;
  // One byte opcode for test al, 0xXX.
  static constexpr uint8_t kTestAlByte = 0xA8;
  // One byte opcode for nop.
  static constexpr uint8_t kNopByte = 0x90;
 
  // One byte prefix for a short conditional jump.
  static constexpr uint8_t kJccShortPrefix = 0x70;
  static constexpr uint8_t kJncShortOpcode = kJccShortPrefix | not_carry;
  static constexpr uint8_t kJcShortOpcode = kJccShortPrefix | carry;
  static constexpr uint8_t kJnzShortOpcode = kJccShortPrefix | not_zero;
  static constexpr uint8_t kJzShortOpcode = kJccShortPrefix | zero;
 
  // VEX prefix encodings.
  enum SIMDPrefix { kNoPrefix = 0x0, k66 = 0x1, kF3 = 0x2, kF2 = 0x3 };
  enum VectorLength { kL128 = 0x0, kL256 = 0x4, kLIG = kL128, kLZ = kL128 };
  enum VexW { kW0 = 0x0, kW1 = 0x80, kWIG = kW0 };
  enum LeadingOpcode { k0F = 0x1, k0F38 = 0x2, k0F3A = 0x3 };
 
  // ---------------------------------------------------------------------------
  // InstructionStream generation
  //
  // Function names correspond one-to-one to x64 instruction mnemonics.
  // Unless specified otherwise, instructions operate on 64-bit operands.
  //
  // If we need versions of an assembly instruction that operate on different
  // width arguments, we add a single-letter suffix specifying the width.
  // This is done for the following instructions: mov, cmp, inc, dec,
  // add, sub, and test.
  // There are no versions of these instructions without the suffix.
  // - Instructions on 8-bit (byte) operands/registers have a trailing 'b'.
  // - Instructions on 16-bit (word) operands/registers have a trailing 'w'.
  // - Instructions on 32-bit (doubleword) operands/registers use 'l'.
  // - Instructions on 64-bit (quadword) operands/registers use 'q'.
  // - Instructions on operands/registers with pointer size use 'p'.
 
#define DECLARE_INSTRUCTION(instruction)    \
  template <typename... Ps>                 \
  void instruction##_tagged(Ps... ps) {     \
    emit_##instruction(ps..., kTaggedSize); \
  }                                         \
                                            \
  template <typename... Ps>                 \
  void instruction##l(Ps... ps) {           \
    emit_##instruction(ps..., kInt32Size);  \
  }                                         \
                                            \
  template <typename... Ps>                 \
  void instruction##q(Ps... ps) {           \
    emit_##instruction(ps..., kInt64Size);  \
  }
  ASSEMBLER_INSTRUCTION_LIST(DECLARE_INSTRUCTION)
#undef DECLARE_INSTRUCTION
 
  // Insert the smallest number of nop instructions
  // possible to align the pc offset to a multiple
  // of m, where m must be a power of 2.
  void Align(int m);
  // Insert the smallest number of zero bytes possible to align the pc offset
  // to a multiple of m. m must be a power of 2 (>= 2).
  void DataAlign(int m);
  void Nop(int bytes = 1);
 
  // Intel CPUs with the Skylake microarchitecture suffer from a performance
  // regression by the JCC erratum. To mitigate the performance impact, we align
  // jcc instructions so that they will not cross or end at 32-byte boundaries.
  // {inst_size} is the total size of the instructions which we will avoid to
  // cross or end at the boundaries. For example, aaaabbbb is a fused jcc
  // instructions, e.g., cmpq+jmp. In the fused case we have:
  // ...aaaabbbbbb
  //    ^         ^
  //    |         pc_offset + inst_size
  //    pc_offset
  // And in the non-fused case:
  // ...bbbb
  //    ^   ^
  //    |   pc_offset + inst_size
  //    pc_offset
  void AlignForJCCErratum(int inst_size);
 
  void emit_trace_instruction(Immediate markid);
 
  // Aligns code to something that's optimal for a jump target for the platform.
  void CodeTargetAlign();
  void LoopHeaderAlign();
 
  // Stack
  void pushfq();
  void popfq();
 
  void pushq(Immediate value);
  // Push a 32 bit integer, and guarantee that it is actually pushed as a
  // 32 bit value, the normal push will optimize the 8 bit case.
  static constexpr int kPushq32InstrSize = 5;
  void pushq_imm32(int32_t imm32);
  void pushq(Register src);
  void pushq(Operand src);
 
  void popq(Register dst);
  void popq(Operand dst);
 
  void incsspq(Register number_of_words);
 
  void leave();
 
  // Moves
  void movb(Register dst, Operand src);
  void movb(Register dst, Immediate imm);
  void movb(Operand dst, Register src);
  void movb(Operand dst, Immediate imm);
 
  // Move the low 16 bits of a 64-bit register value to a 16-bit
  // memory location.
  void movw(Register dst, Operand src);
  void movw(Operand dst, Register src);
  void movw(Operand dst, Immediate imm);
 
  // Move the offset of the label location relative to the current
  // position (after the move) to the destination.
  void movl(Operand dst, Label* src);
 
  // Load a heap number into a register.
  // The heap number will not be allocated and embedded into the code right
  // away. Instead, we emit the load of a dummy object. Later, when calling
  // Assembler::GetCode, the heap number will be allocated and the code will be
  // patched by replacing the dummy with the actual object. The RelocInfo for
  // the embedded object gets already recorded correctly when emitting the dummy
  // move.
  void movq_heap_number(Register dst, double value);
 
  // Loads a 64-bit immediate into a register, potentially using the constant
  // pool.
  void movq(Register dst, int64_t value) { movq(dst, Immediate64(value)); }
  void movq(Register dst, uint64_t value) {
    movq(dst, Immediate64(static_cast<int64_t>(value)));
  }
 
  // Loads a 64-bit immediate into a register without using the constant pool.
  void movq_imm64(Register dst, int64_t value);
 
  void movsxbl(Register dst, Register src);
  void movsxbl(Register dst, Operand src);
  void movsxbq(Register dst, Register src);
  void movsxbq(Register dst, Operand src);
  void movsxwl(Register dst, Register src);
  void movsxwl(Register dst, Operand src);
  void movsxwq(Register dst, Register src);
  void movsxwq(Register dst, Operand src);
  void movsxlq(Register dst, Register src);
  void movsxlq(Register dst, Operand src);
 
  // Repeated moves.
  void repmovsb();
  void repmovsw();
  void repmovsl() { emit_repmovs(kInt32Size); }
  void repmovsq() { emit_repmovs(kInt64Size); }
 
  // Repeated store of doublewords (fill (E)CX bytes at ES:[(E)DI] with EAX).
  void repstosl();
  // Repeated store of quadwords (fill RCX quadwords at [RDI] with RAX).
  void repstosq();
 
  // Instruction to load from an immediate 64-bit pointer into RAX.
  void load_rax(Address value, RelocInfo::Mode rmode);
  void load_rax(ExternalReference ext);
 
  // Conditional moves.
  void cmovq(Condition cc, Register dst, Register src);
  void cmovq(Condition cc, Register dst, Operand src);
  void cmovl(Condition cc, Register dst, Register src);
  void cmovl(Condition cc, Register dst, Operand src);
 
  void cmpb(Register dst, Immediate src) {
    immediate_arithmetic_op_8(0x7, dst, src);
  }
 
  // Used for JCC erratum performance mitigation.
  void aligned_cmpb(Register dst, Immediate src) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* cmp */ 4 + /* jcc */ 6
    const int kMaxMacroFusionLength = 10;
    AlignForJCCErratum(kMaxMacroFusionLength);
    cmpb(dst, src);
  }
 
  void cmpb_al(Immediate src);
 
  void cmpb(Register dst, Register src) { arithmetic_op_8(0x3A, dst, src); }
 
  // Used for JCC erratum performance mitigation.
  void aligned_cmpb(Register dst, Register src) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* cmp */ 3 + /* jcc */ 6
    const int kMaxMacroFusionLength = 9;
    AlignForJCCErratum(kMaxMacroFusionLength);
    cmpb(dst, src);
  }
 
  void cmpb(Register dst, Operand src) { arithmetic_op_8(0x3A, dst, src); }
 
  // Used for JCC erratum performance mitigation.
  void aligned_cmpb(Register dst, Operand src) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* cmp */ 8 + /* jcc */ 6
    const int kMaxMacroFusionLength = 14;
    AlignForJCCErratum(kMaxMacroFusionLength);
    cmpb(dst, src);
  }
 
  void cmpb(Operand dst, Register src) { arithmetic_op_8(0x38, src, dst); }
 
  // Used for JCC erratum performance mitigation.
  void aligned_cmpb(Operand dst, Register src) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* cmp */ 8 + /* jcc */ 6
    const int kMaxMacroFusionLength = 14;
    AlignForJCCErratum(kMaxMacroFusionLength);
    cmpb(dst, src);
  }
 
  void cmpb(Operand dst, Immediate src) {
    immediate_arithmetic_op_8(0x7, dst, src);
  }
 
  // Used for JCC erratum performance mitigation.
  void aligned_cmpb(Operand dst, Immediate src) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // cmp can not be fused when comparing MEM-IMM, so we would not align this
    // instruction.
    cmpb(dst, src);
  }
 
  void cmpw(Operand dst, Immediate src) {
    immediate_arithmetic_op_16(0x7, dst, src);
  }
 
  // Used for JCC erratum performance mitigation.
  void aligned_cmpw(Operand dst, Immediate src) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // cmp can not be fused when comparing MEM-IMM, so we would not align this
    // instruction.
    cmpw(dst, src);
  }
 
  void cmpw(Register dst, Immediate src) {
    immediate_arithmetic_op_16(0x7, dst, src);
  }
 
  // Used for JCC erratum performance mitigation.
  void aligned_cmpw(Register dst, Immediate src) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* cmp */ 6 + /* jcc */ 6
    const int kMaxMacroFusionLength = 12;
    AlignForJCCErratum(kMaxMacroFusionLength);
    cmpw(dst, src);
  }
 
  void cmpw(Register dst, Operand src) { arithmetic_op_16(0x3B, dst, src); }
 
  // Used for JCC erratum performance mitigation.
  void aligned_cmpw(Register dst, Operand src) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* cmp */ 9 + /* jcc */ 6
    const int kMaxMacroFusionLength = 15;
    AlignForJCCErratum(kMaxMacroFusionLength);
    cmpw(dst, src);
  }
 
  void cmpw(Register dst, Register src) { arithmetic_op_16(0x3B, dst, src); }
 
  // Used for JCC erratum performance mitigation.
  void aligned_cmpw(Register dst, Register src) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* cmp */ 4 + /* jcc */ 6
    const int kMaxMacroFusionLength = 10;
    AlignForJCCErratum(kMaxMacroFusionLength);
    cmpw(dst, src);
  }
 
  void cmpw(Operand dst, Register src) { arithmetic_op_16(0x39, src, dst); }
 
  // Used for JCC erratum performance mitigation.
  void aligned_cmpw(Operand dst, Register src) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* cmp */ 9 + /* jcc */ 6
    const int kMaxMacroFusionLength = 15;
    AlignForJCCErratum(kMaxMacroFusionLength);
    cmpw(dst, src);
  }
 
  void testb(Register reg, Operand op) { testb(op, reg); }
 
  // Used for JCC erratum performance mitigation.
  void aligned_testb(Register reg, Operand op) { aligned_testb(op, reg); }
 
  void testw(Register reg, Operand op) { testw(op, reg); }
 
  // Used for JCC erratum performance mitigation.
  void aligned_testw(Register reg, Operand op) { aligned_testw(op, reg); }
 
  void andb(Register dst, Immediate src) {
    immediate_arithmetic_op_8(0x4, dst, src);
  }
 
  void decb(Register dst);
  void decb(Operand dst);
 
  // Lock prefix.
  void lock();
 
  void xchgb(Register reg, Operand op);
  void xchgw(Register reg, Operand op);
 
  void xaddb(Operand dst, Register src);
  void xaddw(Operand dst, Register src);
  void xaddl(Operand dst, Register src);
  void xaddq(Operand dst, Register src);
 
  void negb(Register reg);
  void negw(Register reg);
  void negl(Register reg);
  void negq(Register reg);
  void negb(Operand op);
  void negw(Operand op);
  void negl(Operand op);
  void negq(Operand op);
 
  void cmpxchgb(Operand dst, Register src);
  void cmpxchgw(Operand dst, Register src);
 
  // Sign-extends rax into rdx:rax.
  void cqo();
  // Sign-extends eax into edx:eax.
  void cdq();
 
  // Multiply eax by src, put the result in edx:eax.
  void mull(Register src);
  void mull(Operand src);
  // Multiply rax by src, put the result in rdx:rax.
  void mulq(Register src);
  void mulq(Operand src);
 
#define DECLARE_SHIFT_INSTRUCTION(instruction, subcode)                     \
  void instruction##l(Register dst, Immediate imm8) {                       \
    shift(dst, imm8, subcode, kInt32Size);                                  \
  }                                                                         \
                                                                            \
  void instruction##q(Register dst, Immediate imm8) {                       \
    shift(dst, imm8, subcode, kInt64Size);                                  \
  }                                                                         \
                                                                            \
  void instruction##l(Operand dst, Immediate imm8) {                        \
    shift(dst, imm8, subcode, kInt32Size);                                  \
  }                                                                         \
                                                                            \
  void instruction##q(Operand dst, Immediate imm8) {                        \
    shift(dst, imm8, subcode, kInt64Size);                                  \
  }                                                                         \
                                                                            \
  void instruction##l_cl(Register dst) { shift(dst, subcode, kInt32Size); } \
                                                                            \
  void instruction##q_cl(Register dst) { shift(dst, subcode, kInt64Size); } \
                                                                            \
  void instruction##l_cl(Operand dst) { shift(dst, subcode, kInt32Size); }  \
                                                                            \
  void instruction##q_cl(Operand dst) { shift(dst, subcode, kInt64Size); }
  SHIFT_INSTRUCTION_LIST(DECLARE_SHIFT_INSTRUCTION)
#undef DECLARE_SHIFT_INSTRUCTION
 
  // Shifts dst:src left by cl bits, affecting only dst.
  void shld(Register dst, Register src);
 
  // Shifts src:dst right by cl bits, affecting only dst.
  void shrd(Register dst, Register src);
 
  void store_rax(Address dst, RelocInfo::Mode mode);
  void store_rax(ExternalReference ref);
 
  void subb(Register dst, Immediate src) {
    immediate_arithmetic_op_8(0x5, dst, src);
  }
 
  void sub_sp_32(uint32_t imm);
 
  void testb(Register dst, Register src);
  // Used for JCC erratum performance mitigation.
  void aligned_testb(Register dst, Register src) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* test */ 3 + /* jcc */ 6
    const int kMaxMacroFusionLength = 9;
    AlignForJCCErratum(kMaxMacroFusionLength);
    testb(dst, src);
  }
 
  void testb(Register reg, Immediate mask);
  // Used for JCC erratum performance mitigation.
  void aligned_testb(Register reg, Immediate mask) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* test */ 4 + /* jcc */ 6
    const int kMaxMacroFusionLength = 10;
    AlignForJCCErratum(kMaxMacroFusionLength);
    testb(reg, mask);
  }
 
  void testb(Operand op, Immediate mask);
  // Used for JCC erratum performance mitigation.
  void aligned_testb(Operand op, Immediate mask) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // test can not be fused when comparing MEM-IMM, so we would not align this
    // instruction.
    testb(op, mask);
  }
 
  void testb(Operand op, Register reg);
  // Used for JCC erratum performance mitigation.
  void aligned_testb(Operand op, Register reg) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* test */ 8 + /* jcc */ 6
    const int kMaxMacroFusionLength = 14;
    AlignForJCCErratum(kMaxMacroFusionLength);
    testb(op, reg);
  }
 
  void testw(Register dst, Register src);
  // Used for JCC erratum performance mitigation.
  void aligned_testw(Register dst, Register src) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* test */ 4 + /* jcc */ 6
    const int kMaxMacroFusionLength = 10;
    AlignForJCCErratum(kMaxMacroFusionLength);
    testw(dst, src);
  }
 
  void testw(Register reg, Immediate mask);
  // Used for JCC erratum performance mitigation.
  void aligned_testw(Register reg, Immediate mask) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* test */ 6 + /* jcc */ 6
    const int kMaxMacroFusionLength = 12;
    AlignForJCCErratum(kMaxMacroFusionLength);
    testw(reg, mask);
  }
 
  void testw(Operand op, Immediate mask);
  // Used for JCC erratum performance mitigation.
  void aligned_testw(Operand op, Immediate mask) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // test can not be fused when comparing MEM-IMM, so we would not align this
    // instruction.
    testw(op, mask);
  }
 
  void testw(Operand op, Register reg);
  // Used for JCC erratum performance mitigation.
  void aligned_testw(Operand op, Register reg) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* test */ 9 + /* jcc */ 6
    const int kMaxMacroFusionLength = 15;
    AlignForJCCErratum(kMaxMacroFusionLength);
    testw(op, reg);
  }
 
  // Bit operations.
  void bswapl(Register dst);
  void bswapq(Register dst);
  void btq(Operand dst, Register src);
  void btsq(Operand dst, Register src);
  void btsq(Register dst, Immediate imm8);
  void btrq(Register dst, Immediate imm8);
  void bsrq(Register dst, Register src);
  void bsrq(Register dst, Operand src);
  void bsrl(Register dst, Register src);
  void bsrl(Register dst, Operand src);
  void bsfq(Register dst, Register src);
  void bsfq(Register dst, Operand src);
  void bsfl(Register dst, Register src);
  void bsfl(Register dst, Operand src);
 
  // Miscellaneous
  void clc();
  void cld();
  void cpuid();
  void hlt();
  void int3();
  void nop();
  void ret(int imm16);
  void ud2();
  void setcc(Condition cc, Register reg);
  void endbr64();
 
  void pblendw(XMMRegister dst, Operand src, uint8_t mask);
  void pblendw(XMMRegister dst, XMMRegister src, uint8_t mask);
  void palignr(XMMRegister dst, Operand src, uint8_t mask);
  void palignr(XMMRegister dst, XMMRegister src, uint8_t mask);
 
  void vpermq(YMMRegister dst, Operand src, uint8_t imm8) {
    vinstr(0x0, dst, ymm0, src, k66, k0F3A, kW1, AVX2);
    emit(imm8);
  }
  void vpermq(YMMRegister dst, YMMRegister src, uint8_t imm8) {
    vinstr(0x0, dst, ymm0, src, k66, k0F3A, kW1, AVX2);
    emit(imm8);
  }
 
  // Label operations & relative jumps (PPUM Appendix D)
  //
  // Takes a branch opcode (cc) and a label (L) and generates
  // either a backward branch or a forward branch and links it
  // to the label fixup chain. Usage:
  //
  // Label L;    // unbound label
  // j(cc, &L);  // forward branch to unbound label
  // bind(&L);   // bind label to the current pc
  // j(cc, &L);  // backward branch to bound label
  // bind(&L);   // illegal: a label may be bound only once
  //
  // Note: The same Label can be used for forward and backward branches
  // but it may be bound only once.
 
  void bind(Label* L);  // binds an unbound label L to the current code position
 
  // Calls
  // Call near relative 32-bit displacement, relative to next instruction.
  void call(Label* L);
 
  // Explicitly emit a near call / near jump. The displacement is relative to
  // the next instructions (which starts at
  // {pc_offset() + kIntraSegmentJmpInstrSize}).
  static constexpr int kIntraSegmentJmpInstrSize = 5;
  void near_call(intptr_t disp, RelocInfo::Mode rmode);
  void near_call(Builtin buitin, RelocInfo::Mode rmode);
  void near_jmp(intptr_t disp, RelocInfo::Mode rmode);
  void near_j(Condition cc, intptr_t disp, RelocInfo::Mode rmode);
 
  void call(Handle<Code> target,
            RelocInfo::Mode rmode = RelocInfo::CODE_TARGET);
 
  // Call near absolute indirect, address in register
  void call(Register adr);
 
  // Jumps
  // Jump short or near relative.
  // Use a 32-bit signed displacement.
  // Unconditional jump to L
  void jmp(Label* L, Label::Distance distance = Label::kFar);
  // Used for JCC erratum performance mitigation.
  void aligned_jmp(Label* L, Label::Distance distance = Label::kFar) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    const int kInstLength = distance == Label::kFar ? 6 : 2;
    AlignForJCCErratum(kInstLength);
    jmp(L, distance);
  }
  void jmp(Handle<Code> target, RelocInfo::Mode rmode);
 
  // Jump near absolute indirect (r64)
  // With notrack, add an optional prefix to disable CET IBT enforcement for
  // this jump.
  void jmp(Register adr, bool notrack = false);
  void jmp(Operand src, bool notrack = false);
 
  // Unconditional jump relative to the current address. Low-level routine,
  // use with caution!
  void jmp_rel(int offset);
 
  // Conditional jumps
  void j(Condition cc, Label* L, Label::Distance distance = Label::kFar);
  // Used for JCC erratum performance mitigation.
  void aligned_j(Condition cc, Label* L,
                 Label::Distance distance = Label::kFar) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    const int kInstLength = distance == Label::kFar ? 6 : 2;
    AlignForJCCErratum(kInstLength);
    j(cc, L, distance);
  }
  void j(Condition cc, Address entry, RelocInfo::Mode rmode);
  void j(Condition cc, Handle<Code> target, RelocInfo::Mode rmode);
 
  // Floating-point operations
  void fld(int i);
 
  void fld1();
  void fldz();
  void fldpi();
  void fldln2();
 
  void fld_s(Operand adr);
  void fld_d(Operand adr);
 
  void fstp_s(Operand adr);
  void fstp_d(Operand adr);
  void fstp(int index);
 
  void fild_s(Operand adr);
  void fild_d(Operand adr);
 
  void fist_s(Operand adr);
 
  void fistp_s(Operand adr);
  void fistp_d(Operand adr);
 
  void fisttp_s(Operand adr);
  void fisttp_d(Operand adr);
 
  void fabs();
  void fchs();
 
  void fadd(int i);
  void fsub(int i);
  void fmul(int i);
  void fdiv(int i);
 
  void fisub_s(Operand adr);
 
  void faddp(int i = 1);
  void fsubp(int i = 1);
  void fsubrp(int i = 1);
  void fmulp(int i = 1);
  void fdivp(int i = 1);
  void fprem();
  void fprem1();
 
  void fxch(int i = 1);
  void fincstp();
  void ffree(int i = 0);
 
  void ftst();
  void fucomp(int i);
  void fucompp();
  void fucomi(int i);
  void fucomip();
 
  void fcompp();
  void fnstsw_ax();
  void fwait();
  void fnclex();
 
  void fsin();
  void fcos();
  void fptan();
  void fyl2x();
  void f2xm1();
  void fscale();
  void fninit();
 
  void frndint();
 
  void sahf();
 
  void ucomiss(XMMRegister dst, XMMRegister src);
  void ucomiss(XMMRegister dst, Operand src);
  void movaps(XMMRegister dst, XMMRegister src);
  void movaps(XMMRegister dst, Operand src);
 
  // Don't use this unless it's important to keep the
  // top half of the destination register unchanged.
  // Use movaps when moving float values and movd for integer
  // values in xmm registers.
  void movss(XMMRegister dst, XMMRegister src);
 
  void movss(XMMRegister dst, Operand src);
  void movss(Operand dst, XMMRegister src);
 
  void movlps(XMMRegister dst, Operand src);
  void movlps(Operand dst, XMMRegister src);
 
  void movhps(XMMRegister dst, Operand src);
  void movhps(Operand dst, XMMRegister src);
 
  void shufps(XMMRegister dst, XMMRegister src, uint8_t imm8);
 
  void cvttss2si(Register dst, Operand src);
  void cvttss2si(Register dst, XMMRegister src);
  void cvtlsi2ss(XMMRegister dst, Operand src);
  void cvtlsi2ss(XMMRegister dst, Register src);
 
  void movmskps(Register dst, XMMRegister src);
 
  void vinstr(uint8_t op, XMMRegister dst, XMMRegister src1, XMMRegister src2,
              SIMDPrefix pp, LeadingOpcode m, VexW w, CpuFeature feature = AVX);
  void vinstr(uint8_t op, XMMRegister dst, XMMRegister src1, Operand src2,
              SIMDPrefix pp, LeadingOpcode m, VexW w, CpuFeature feature = AVX);
 
  template <typename Reg1, typename Reg2, typename Op>
  void vinstr(uint8_t op, Reg1 dst, Reg2 src1, Op src2, SIMDPrefix pp,
              LeadingOpcode m, VexW w, CpuFeature feature = AVX2);
 
  // SSE instructions
  void sse_instr(XMMRegister dst, XMMRegister src, uint8_t escape,
                 uint8_t opcode);
  void sse_instr(XMMRegister dst, Operand src, uint8_t escape, uint8_t opcode);
#define DECLARE_SSE_INSTRUCTION(instruction, escape, opcode) \
  void instruction(XMMRegister dst, XMMRegister src) {       \
    sse_instr(dst, src, 0x##escape, 0x##opcode);             \
  }                                                          \
  void instruction(XMMRegister dst, Operand src) {           \
    sse_instr(dst, src, 0x##escape, 0x##opcode);             \
  }
 
  SSE_UNOP_INSTRUCTION_LIST(DECLARE_SSE_INSTRUCTION)
  SSE_BINOP_INSTRUCTION_LIST(DECLARE_SSE_INSTRUCTION)
#undef DECLARE_SSE_INSTRUCTION
 
  // SSE instructions with prefix and SSE2 instructions
  void sse2_instr(XMMRegister dst, XMMRegister src, uint8_t prefix,
                  uint8_t escape, uint8_t opcode);
  void sse2_instr(XMMRegister dst, Operand src, uint8_t prefix, uint8_t escape,
                  uint8_t opcode);
#define DECLARE_SSE2_INSTRUCTION(instruction, prefix, escape, opcode) \
  void instruction(XMMRegister dst, XMMRegister src) {                \
    sse2_instr(dst, src, 0x##prefix, 0x##escape, 0x##opcode);         \
  }                                                                   \
  void instruction(XMMRegister dst, Operand src) {                    \
    sse2_instr(dst, src, 0x##prefix, 0x##escape, 0x##opcode);         \
  }
 
  // These SSE instructions have the same encoding as the SSE2 instructions.
  SSE_INSTRUCTION_LIST_SS(DECLARE_SSE2_INSTRUCTION)
  SSE2_INSTRUCTION_LIST(DECLARE_SSE2_INSTRUCTION)
  SSE2_INSTRUCTION_LIST_SD(DECLARE_SSE2_INSTRUCTION)
  SSE2_UNOP_INSTRUCTION_LIST(DECLARE_SSE2_INSTRUCTION)
#undef DECLARE_SSE2_INSTRUCTION
 
  void sse2_instr(XMMRegister reg, uint8_t imm8, uint8_t prefix, uint8_t escape,
                  uint8_t opcode, int extension) {
    XMMRegister ext_reg = XMMRegister::from_code(extension);
    sse2_instr(ext_reg, reg, prefix, escape, opcode);
    emit(imm8);
  }
 
#define DECLARE_SSE2_SHIFT_IMM(instruction, prefix, escape, opcode, extension) \
  void instruction(XMMRegister reg, uint8_t imm8) {                            \
    sse2_instr(reg, imm8, 0x##prefix, 0x##escape, 0x##opcode, 0x##extension);  \
  }
  SSE2_INSTRUCTION_LIST_SHIFT_IMM(DECLARE_SSE2_SHIFT_IMM)
#undef DECLARE_SSE2_SHIFT_IMM
 
#define DECLARE_SSE2_AVX_INSTRUCTION(instruction, prefix, escape, opcode)    \
  void v##instruction(XMMRegister dst, XMMRegister src1, XMMRegister src2) { \
    vinstr(0x##opcode, dst, src1, src2, k##prefix, k##escape, kW0);          \
  }                                                                          \
  void v##instruction(XMMRegister dst, XMMRegister src1, Operand src2) {     \
    vinstr(0x##opcode, dst, src1, src2, k##prefix, k##escape, kW0);          \
  }
 
#define DECLARE_SSE2_PD_AVX_INSTRUCTION(instruction, prefix, escape, opcode) \
  DECLARE_SSE2_AVX_INSTRUCTION(instruction, prefix, escape, opcode)          \
  void v##instruction(YMMRegister dst, YMMRegister src1, YMMRegister src2) { \
    vinstr(0x##opcode, dst, src1, src2, k##prefix, k##escape, kW0, AVX);     \
  }                                                                          \
  void v##instruction(YMMRegister dst, YMMRegister src1, Operand src2) {     \
    vinstr(0x##opcode, dst, src1, src2, k##prefix, k##escape, kW0, AVX);     \
  }
 
  SSE2_INSTRUCTION_LIST_PD(DECLARE_SSE2_PD_AVX_INSTRUCTION)
#undef DECLARE_SSE2_PD_AVX_INSTRUCTION
 
#define DECLARE_SSE2_PI_AVX_INSTRUCTION(instruction, prefix, escape, opcode) \
  DECLARE_SSE2_AVX_INSTRUCTION(instruction, prefix, escape, opcode)          \
  void v##instruction(YMMRegister dst, YMMRegister src1, YMMRegister src2) { \
    vinstr(0x##opcode, dst, src1, src2, k##prefix, k##escape, kW0, AVX2);    \
  }                                                                          \
  void v##instruction(YMMRegister dst, YMMRegister src1, Operand src2) {     \
    vinstr(0x##opcode, dst, src1, src2, k##prefix, k##escape, kW0, AVX2);    \
  }
 
  SSE2_INSTRUCTION_LIST_PI(DECLARE_SSE2_PI_AVX_INSTRUCTION)
#undef DECLARE_SSE2_PI_AVX_INSTRUCTION
 
#define DECLARE_SSE2_SHIFT_AVX_INSTRUCTION(instruction, prefix, escape,      \
                                           opcode)                           \
  DECLARE_SSE2_AVX_INSTRUCTION(instruction, prefix, escape, opcode)          \
  void v##instruction(YMMRegister dst, YMMRegister src1, XMMRegister src2) { \
    vinstr(0x##opcode, dst, src1, src2, k##prefix, k##escape, kW0, AVX2);    \
  }                                                                          \
  void v##instruction(YMMRegister dst, YMMRegister src1, Operand src2) {     \
    vinstr(0x##opcode, dst, src1, src2, k##prefix, k##escape, kW0, AVX2);    \
  }
 
  SSE2_INSTRUCTION_LIST_SHIFT(DECLARE_SSE2_SHIFT_AVX_INSTRUCTION)
#undef DECLARE_SSE2_SHIFT_AVX_INSTRUCTION
#undef DECLARE_SSE2_AVX_INSTRUCTION
 
#define DECLARE_SSE2_UNOP_AVX_INSTRUCTION(instruction, prefix, escape, opcode) \
  void v##instruction(XMMRegister dst, XMMRegister src) {                      \
    vpd(0x##opcode, dst, xmm0, src);                                           \
  }                                                                            \
  void v##instruction(XMMRegister dst, Operand src) {                          \
    vpd(0x##opcode, dst, xmm0, src);                                           \
  }
 
  SSE2_UNOP_INSTRUCTION_LIST(DECLARE_SSE2_UNOP_AVX_INSTRUCTION)
#undef DECLARE_SSE2_UNOP_AVX_INSTRUCTION
 
#define DECLARE_SSE2_UNOP_AVX_YMM_INSTRUCTION(                 \
    instruction, opcode, DSTRegister, SRCRegister, MemOperand) \
  void v##instruction(DSTRegister dst, SRCRegister src) {      \
    vpd(0x##opcode, dst, ymm0, src);                           \
  }                                                            \
  void v##instruction(DSTRegister dst, MemOperand src) {       \
    vpd(0x##opcode, dst, ymm0, src);                           \
  }
  DECLARE_SSE2_UNOP_AVX_YMM_INSTRUCTION(sqrtpd, 51, YMMRegister, YMMRegister,
                                        Operand)
  DECLARE_SSE2_UNOP_AVX_YMM_INSTRUCTION(cvtpd2ps, 5A, XMMRegister, YMMRegister,
                                        Operand256)
  DECLARE_SSE2_UNOP_AVX_YMM_INSTRUCTION(cvtps2dq, 5B, YMMRegister, YMMRegister,
                                        Operand)
  DECLARE_SSE2_UNOP_AVX_YMM_INSTRUCTION(cvttpd2dq, E6, XMMRegister, YMMRegister,
                                        Operand256)
#undef DECLARE_SSE2_UNOP_AVX_YMM_INSTRUCTION
 
  // SSE3
  void lddqu(XMMRegister dst, Operand src);
  void movddup(XMMRegister dst, Operand src);
  void movddup(XMMRegister dst, XMMRegister src);
  void movshdup(XMMRegister dst, XMMRegister src);
 
  // SSSE3
  void ssse3_instr(XMMRegister dst, XMMRegister src, uint8_t prefix,
                   uint8_t escape1, uint8_t escape2, uint8_t opcode);
  void ssse3_instr(XMMRegister dst, Operand src, uint8_t prefix,
                   uint8_t escape1, uint8_t escape2, uint8_t opcode);
 
#define DECLARE_SSSE3_INSTRUCTION(instruction, prefix, escape1, escape2,     \
                                  opcode)                                    \
  void instruction(XMMRegister dst, XMMRegister src) {                       \
    ssse3_instr(dst, src, 0x##prefix, 0x##escape1, 0x##escape2, 0x##opcode); \
  }                                                                          \
  void instruction(XMMRegister dst, Operand src) {                           \
    ssse3_instr(dst, src, 0x##prefix, 0x##escape1, 0x##escape2, 0x##opcode); \
  }
 
  SSSE3_INSTRUCTION_LIST(DECLARE_SSSE3_INSTRUCTION)
  SSSE3_UNOP_INSTRUCTION_LIST(DECLARE_SSSE3_INSTRUCTION)
#undef DECLARE_SSSE3_INSTRUCTION
 
  // SSE4
  void sse4_instr(Register dst, XMMRegister src, uint8_t prefix,
                  uint8_t escape1, uint8_t escape2, uint8_t opcode,
                  int8_t imm8);
  void sse4_instr(Operand dst, XMMRegister src, uint8_t prefix, uint8_t escape1,
                  uint8_t escape2, uint8_t opcode, int8_t imm8);
  void sse4_instr(XMMRegister dst, Register src, uint8_t prefix,
                  uint8_t escape1, uint8_t escape2, uint8_t opcode,
                  int8_t imm8);
  void sse4_instr(XMMRegister dst, XMMRegister src, uint8_t prefix,
                  uint8_t escape1, uint8_t escape2, uint8_t opcode);
  void sse4_instr(XMMRegister dst, Operand src, uint8_t prefix, uint8_t escape1,
                  uint8_t escape2, uint8_t opcode);
#define DECLARE_SSE4_INSTRUCTION(instruction, prefix, escape1, escape2,     \
                                 opcode)                                    \
  void instruction(XMMRegister dst, XMMRegister src) {                      \
    sse4_instr(dst, src, 0x##prefix, 0x##escape1, 0x##escape2, 0x##opcode); \
  }                                                                         \
  void instruction(XMMRegister dst, Operand src) {                          \
    sse4_instr(dst, src, 0x##prefix, 0x##escape1, 0x##escape2, 0x##opcode); \
  }
 
  SSE4_INSTRUCTION_LIST(DECLARE_SSE4_INSTRUCTION)
  SSE4_UNOP_INSTRUCTION_LIST(DECLARE_SSE4_INSTRUCTION)
  DECLARE_SSE4_INSTRUCTION(pblendvb, 66, 0F, 38, 10)
  DECLARE_SSE4_INSTRUCTION(blendvps, 66, 0F, 38, 14)
  DECLARE_SSE4_INSTRUCTION(blendvpd, 66, 0F, 38, 15)
#undef DECLARE_SSE4_INSTRUCTION
 
#define DECLARE_SSE4_EXTRACT_INSTRUCTION(instruction, prefix, escape1,     \
                                         escape2, opcode)                  \
  void instruction(Register dst, XMMRegister src, uint8_t imm8) {          \
    sse4_instr(dst, src, 0x##prefix, 0x##escape1, 0x##escape2, 0x##opcode, \
               imm8);                                                      \
  }                                                                        \
  void instruction(Operand dst, XMMRegister src, uint8_t imm8) {           \
    sse4_instr(dst, src, 0x##prefix, 0x##escape1, 0x##escape2, 0x##opcode, \
               imm8);                                                      \
  }
 
  SSE4_EXTRACT_INSTRUCTION_LIST(DECLARE_SSE4_EXTRACT_INSTRUCTION)
#undef DECLARE_SSE4_EXTRACT_INSTRUCTION
 
  // SSE4.2
  void sse4_2_instr(XMMRegister dst, XMMRegister src, uint8_t prefix,
                    uint8_t escape1, uint8_t escape2, uint8_t opcode);
  void sse4_2_instr(XMMRegister dst, Operand src, uint8_t prefix,
                    uint8_t escape1, uint8_t escape2, uint8_t opcode);
#define DECLARE_SSE4_2_INSTRUCTION(instruction, prefix, escape1, escape2,     \
                                   opcode)                                    \
  void instruction(XMMRegister dst, XMMRegister src) {                        \
    sse4_2_instr(dst, src, 0x##prefix, 0x##escape1, 0x##escape2, 0x##opcode); \
  }                                                                           \
  void instruction(XMMRegister dst, Operand src) {                            \
    sse4_2_instr(dst, src, 0x##prefix, 0x##escape1, 0x##escape2, 0x##opcode); \
  }
 
  SSE4_2_INSTRUCTION_LIST(DECLARE_SSE4_2_INSTRUCTION)
#undef DECLARE_SSE4_2_INSTRUCTION
 
#define DECLARE_SSE34_AVX_INSTRUCTION(instruction, prefix, escape1, escape2,  \
                                      opcode)                                 \
  void v##instruction(XMMRegister dst, XMMRegister src1, XMMRegister src2) {  \
    vinstr(0x##opcode, dst, src1, src2, k##prefix, k##escape1##escape2, kW0); \
  }                                                                           \
  void v##instruction(XMMRegister dst, XMMRegister src1, Operand src2) {      \
    vinstr(0x##opcode, dst, src1, src2, k##prefix, k##escape1##escape2, kW0); \
  }                                                                           \
  void v##instruction(YMMRegister dst, YMMRegister src1, YMMRegister src2) {  \
    vinstr(0x##opcode, dst, src1, src2, k##prefix, k##escape1##escape2, kW0,  \
           AVX2);                                                             \
  }                                                                           \
  void v##instruction(YMMRegister dst, YMMRegister src1, Operand src2) {      \
    vinstr(0x##opcode, dst, src1, src2, k##prefix, k##escape1##escape2, kW0,  \
           AVX2);                                                             \
  }
 
  SSSE3_INSTRUCTION_LIST(DECLARE_SSE34_AVX_INSTRUCTION)
  SSE4_INSTRUCTION_LIST(DECLARE_SSE34_AVX_INSTRUCTION)
  SSE4_2_INSTRUCTION_LIST(DECLARE_SSE34_AVX_INSTRUCTION)
#undef DECLARE_SSE34_AVX_INSTRUCTION
 
#define DECLARE_SSSE3_UNOP_AVX_INSTRUCTION(instruction, prefix, escape1,     \
                                           escape2, opcode)                  \
  void v##instruction(XMMRegister dst, XMMRegister src) {                    \
    vinstr(0x##opcode, dst, xmm0, src, k##prefix, k##escape1##escape2, kW0); \
  }                                                                          \
  void v##instruction(XMMRegister dst, Operand src) {                        \
    vinstr(0x##opcode, dst, xmm0, src, k##prefix, k##escape1##escape2, kW0); \
  }                                                                          \
  void v##instruction(YMMRegister dst, YMMRegister src) {                    \
    vinstr(0x##opcode, dst, ymm0, src, k##prefix, k##escape1##escape2, kW0); \
  }                                                                          \
  void v##instruction(YMMRegister dst, Operand src) {                        \
    vinstr(0x##opcode, dst, ymm0, src, k##prefix, k##escape1##escape2, kW0); \
  }
 
  SSSE3_UNOP_INSTRUCTION_LIST(DECLARE_SSSE3_UNOP_AVX_INSTRUCTION)
#undef DECLARE_SSSE3_UNOP_AVX_INSTRUCTION
 
  void vpblendvb(XMMRegister dst, XMMRegister src1, XMMRegister src2,
                 XMMRegister mask) {
    vinstr(0x4C, dst, src1, src2, k66, k0F3A, kW0);
    // The mask operand is encoded in bits[7:4] of the immediate byte.
    emit(mask.code() << 4);
  }
  void vpblendvb(YMMRegister dst, YMMRegister src1, YMMRegister src2,
                 YMMRegister mask) {
    vinstr(0x4C, dst, src1, src2, k66, k0F3A, kW0, AVX2);
    // The mask operand is encoded in bits[7:4] of the immediate byte.
    emit(mask.code() << 4);
  }
 
  void vblendvps(XMMRegister dst, XMMRegister src1, XMMRegister src2,
                 XMMRegister mask) {
    vinstr(0x4A, dst, src1, src2, k66, k0F3A, kW0);
    // The mask operand is encoded in bits[7:4] of the immediate byte.
    emit(mask.code() << 4);
  }
  void vblendvps(YMMRegister dst, YMMRegister src1, YMMRegister src2,
                 YMMRegister mask) {
    vinstr(0x4A, dst, src1, src2, k66, k0F3A, kW0, AVX);
    // The mask operand is encoded in bits[7:4] of the immediate byte.
    emit(mask.code() << 4);
  }
 
  void vblendvpd(XMMRegister dst, XMMRegister src1, XMMRegister src2,
                 XMMRegister mask) {
    vinstr(0x4B, dst, src1, src2, k66, k0F3A, kW0);
    // The mask operand is encoded in bits[7:4] of the immediate byte.
    emit(mask.code() << 4);
  }
  void vblendvpd(YMMRegister dst, YMMRegister src1, YMMRegister src2,
                 YMMRegister mask) {
    vinstr(0x4B, dst, src1, src2, k66, k0F3A, kW0, AVX);
    // The mask operand is encoded in bits[7:4] of the immediate byte.
    emit(mask.code() << 4);
  }
 
#define DECLARE_SSE4_PMOV_AVX_INSTRUCTION(instruction, prefix, escape1,      \
                                          escape2, opcode)                   \
  void v##instruction(XMMRegister dst, XMMRegister src) {                    \
    vinstr(0x##opcode, dst, xmm0, src, k##prefix, k##escape1##escape2, kW0); \
  }                                                                          \
  void v##instruction(XMMRegister dst, Operand src) {                        \
    vinstr(0x##opcode, dst, xmm0, src, k##prefix, k##escape1##escape2, kW0); \
  }
  SSE4_UNOP_INSTRUCTION_LIST(DECLARE_SSE4_PMOV_AVX_INSTRUCTION)
#undef DECLARE_SSE4_PMOV_AVX_INSTRUCTION
 
#define DECLARE_SSE4_PMOV_AVX2_INSTRUCTION(instruction, prefix, escape1,     \
                                           escape2, opcode)                  \
  void v##instruction(YMMRegister dst, XMMRegister src) {                    \
    vinstr(0x##opcode, dst, xmm0, src, k##prefix, k##escape1##escape2, kW0); \
  }                                                                          \
  void v##instruction(YMMRegister dst, Operand src) {                        \
    vinstr(0x##opcode, dst, xmm0, src, k##prefix, k##escape1##escape2, kW0); \
  }
  SSE4_UNOP_INSTRUCTION_LIST_PMOV(DECLARE_SSE4_PMOV_AVX2_INSTRUCTION)
#undef DECLARE_SSE4_PMOV_AVX2_INSTRUCTION
 
  void vptest(YMMRegister dst, YMMRegister src) {
    vinstr(0x17, dst, ymm0, src, k66, k0F38, kW0, AVX);
  }
  void vptest(YMMRegister dst, Operand src) {
    vinstr(0x17, dst, ymm0, src, k66, k0F38, kW0, AVX);
  }
 
#define DECLARE_AVX_INSTRUCTION(instruction, prefix, escape1, escape2, opcode) \
  void v##instruction(Register dst, XMMRegister src, uint8_t imm8) {           \
    XMMRegister idst = XMMRegister::from_code(dst.code());                     \
    vinstr(0x##opcode, src, xmm0, idst, k##prefix, k##escape1##escape2, kW0);  \
    emit(imm8);                                                                \
  }                                                                            \
  void v##instruction(Operand dst, XMMRegister src, uint8_t imm8) {            \
    vinstr(0x##opcode, src, xmm0, dst, k##prefix, k##escape1##escape2, kW0);   \
    emit(imm8);                                                                \
  }
 
  SSE4_EXTRACT_INSTRUCTION_LIST(DECLARE_AVX_INSTRUCTION)
#undef DECLARE_AVX_INSTRUCTION
 
  void movd(XMMRegister dst, Register src);
  void movd(XMMRegister dst, Operand src);
  void movd(Register dst, XMMRegister src);
  void movq(XMMRegister dst, Register src);
  void movq(XMMRegister dst, Operand src);
  void movq(Register dst, XMMRegister src);
  void movq(XMMRegister dst, XMMRegister src);
 
  // Don't use this unless it's important to keep the
  // top half of the destination register unchanged.
  // Use movapd when moving double values and movq for integer
  // values in xmm registers.
  void movsd(XMMRegister dst, XMMRegister src);
 
  void movsd(Operand dst, XMMRegister src);
  void movsd(XMMRegister dst, Operand src);
 
  void movdqa(Operand dst, XMMRegister src);
  void movdqa(XMMRegister dst, Operand src);
  void movdqa(XMMRegister dst, XMMRegister src);
 
  void movdqu(Operand dst, XMMRegister src);
  void movdqu(XMMRegister dst, Operand src);
  void movdqu(XMMRegister dst, XMMRegister src);
 
  void movapd(XMMRegister dst, XMMRegister src);
  void movupd(XMMRegister dst, Operand src);
  void movupd(Operand dst, XMMRegister src);
 
  void cvtdq2pd(XMMRegister dst, XMMRegister src);
 
  void cvttsd2si(Register dst, Operand src);
  void cvttsd2si(Register dst, XMMRegister src);
  void cvttss2siq(Register dst, XMMRegister src);
  void cvttss2siq(Register dst, Operand src);
  void cvttsd2siq(Register dst, XMMRegister src);
  void cvttsd2siq(Register dst, Operand src);
  void cvttps2dq(XMMRegister dst, Operand src);
  void cvttps2dq(XMMRegister dst, XMMRegister src);
 
  void cvtlsi2sd(XMMRegister dst, Operand src);
  void cvtlsi2sd(XMMRegister dst, Register src);
 
  void cvtqsi2ss(XMMRegister dst, Operand src);
  void cvtqsi2ss(XMMRegister dst, Register src);
 
  void cvtqsi2sd(XMMRegister dst, Operand src);
  void cvtqsi2sd(XMMRegister dst, Register src);
 
  void cvtsd2si(Register dst, XMMRegister src);
  void cvtsd2siq(Register dst, XMMRegister src);
 
  void haddps(XMMRegister dst, XMMRegister src);
  void haddps(XMMRegister dst, Operand src);
 
  void cmpeqsd(XMMRegister dst, XMMRegister src);
  void cmpeqss(XMMRegister dst, XMMRegister src);
  void cmpltsd(XMMRegister dst, XMMRegister src);
 
  void movmskpd(Register dst, XMMRegister src);
 
  void pmovmskb(Register dst, XMMRegister src);
 
  void pinsrw(XMMRegister dst, Register src, uint8_t imm8);
  void pinsrw(XMMRegister dst, Operand src, uint8_t imm8);
 
  // SSE 4.1 instruction
  void insertps(XMMRegister dst, XMMRegister src, uint8_t imm8);
  void insertps(XMMRegister dst, Operand src, uint8_t imm8);
  void pextrq(Register dst, XMMRegister src, int8_t imm8);
  void pinsrb(XMMRegister dst, Register src, uint8_t imm8);
  void pinsrb(XMMRegister dst, Operand src, uint8_t imm8);
  void pinsrd(XMMRegister dst, Register src, uint8_t imm8);
  void pinsrd(XMMRegister dst, Operand src, uint8_t imm8);
  void pinsrq(XMMRegister dst, Register src, uint8_t imm8);
  void pinsrq(XMMRegister dst, Operand src, uint8_t imm8);
 
  void roundss(XMMRegister dst, XMMRegister src, RoundingMode mode);
  void roundss(XMMRegister dst, Operand src, RoundingMode mode);
  void roundsd(XMMRegister dst, XMMRegister src, RoundingMode mode);
  void roundsd(XMMRegister dst, Operand src, RoundingMode mode);
  void roundps(XMMRegister dst, XMMRegister src, RoundingMode mode);
  void roundpd(XMMRegister dst, XMMRegister src, RoundingMode mode);
 
  void cmpps(XMMRegister dst, XMMRegister src, int8_t cmp);
  void cmpps(XMMRegister dst, Operand src, int8_t cmp);
  void cmppd(XMMRegister dst, XMMRegister src, int8_t cmp);
  void cmppd(XMMRegister dst, Operand src, int8_t cmp);
 
#define SSE_CMP_P(instr, imm8)                                                \
  void instr##ps(XMMRegister dst, XMMRegister src) { cmpps(dst, src, imm8); } \
  void instr##ps(XMMRegister dst, Operand src) { cmpps(dst, src, imm8); }     \
  void instr##pd(XMMRegister dst, XMMRegister src) { cmppd(dst, src, imm8); } \
  void instr##pd(XMMRegister dst, Operand src) { cmppd(dst, src, imm8); }
 
  SSE_CMP_P(cmpeq, 0x0)
  SSE_CMP_P(cmplt, 0x1)
  SSE_CMP_P(cmple, 0x2)
  SSE_CMP_P(cmpunord, 0x3)
  SSE_CMP_P(cmpneq, 0x4)
  SSE_CMP_P(cmpnlt, 0x5)
  SSE_CMP_P(cmpnle, 0x6)
 
#undef SSE_CMP_P
 
  void movups(XMMRegister dst, XMMRegister src);
  void movups(XMMRegister dst, Operand src);
  void movups(Operand dst, XMMRegister src);
  void psrldq(XMMRegister dst, uint8_t shift);
  void pshufd(XMMRegister dst, XMMRegister src, uint8_t shuffle);
  void pshufd(XMMRegister dst, Operand src, uint8_t shuffle);
  void pshufhw(XMMRegister dst, XMMRegister src, uint8_t shuffle);
  void pshufhw(XMMRegister dst, Operand src, uint8_t shuffle);
  void pshuflw(XMMRegister dst, XMMRegister src, uint8_t shuffle);
  void pshuflw(XMMRegister dst, Operand src, uint8_t shuffle);
 
  void movhlps(XMMRegister dst, XMMRegister src) {
    sse_instr(dst, src, 0x0F, 0x12);
  }
  void movlhps(XMMRegister dst, XMMRegister src) {
    sse_instr(dst, src, 0x0F, 0x16);
  }
 
  // AVX instruction
  void vmovddup(XMMRegister dst, XMMRegister src);
  void vmovddup(XMMRegister dst, Operand src);
  void vmovddup(YMMRegister dst, YMMRegister src);
  void vmovddup(YMMRegister dst, Operand src);
  void vmovshdup(XMMRegister dst, XMMRegister src);
  void vmovshdup(YMMRegister dst, YMMRegister src);
  void vbroadcastss(XMMRegister dst, Operand src);
  void vbroadcastss(XMMRegister dst, XMMRegister src);
  void vbroadcastss(YMMRegister dst, Operand src);
  void vbroadcastss(YMMRegister dst, XMMRegister src);
  void vbroadcastsd(YMMRegister dst, XMMRegister src);
  void vbroadcastsd(YMMRegister dst, Operand src);
 
  void vinserti128(YMMRegister dst, YMMRegister src1, XMMRegister src2,
                   uint8_t lane);
  void vperm2f128(YMMRegister dst, YMMRegister src1, YMMRegister src2,
                  uint8_t lane);
  void vextractf128(XMMRegister dst, YMMRegister src, uint8_t lane);
 
  template <typename Reg1, typename Reg2, typename Op>
  void fma_instr(uint8_t op, Reg1 dst, Reg2 src1, Op src2, VectorLength l,
                 SIMDPrefix pp, LeadingOpcode m, VexW w);
 
#define FMA(instr, prefix, escape1, escape2, extension, opcode)     \
  void instr(XMMRegister dst, XMMRegister src1, XMMRegister src2) { \
    fma_instr(0x##opcode, dst, src1, src2, kL128, k##prefix,        \
              k##escape1##escape2, k##extension);                   \
  }                                                                 \
  void instr(XMMRegister dst, XMMRegister src1, Operand src2) {     \
    fma_instr(0x##opcode, dst, src1, src2, kL128, k##prefix,        \
              k##escape1##escape2, k##extension);                   \
  }
  FMA_INSTRUCTION_LIST(FMA)
#undef FMA
 
#define DECLARE_FMA_YMM_INSTRUCTION(instr, prefix, escape1, escape2, \
                                    extension, opcode)               \
  void instr(YMMRegister dst, YMMRegister src1, YMMRegister src2) {  \
    fma_instr(0x##opcode, dst, src1, src2, kL256, k##prefix,         \
              k##escape1##escape2, k##extension);                    \
  }                                                                  \
  void instr(YMMRegister dst, YMMRegister src1, Operand src2) {      \
    fma_instr(0x##opcode, dst, src1, src2, kL256, k##prefix,         \
              k##escape1##escape2, k##extension);                    \
  }
  FMA_PS_INSTRUCTION_LIST(DECLARE_FMA_YMM_INSTRUCTION)
  FMA_PD_INSTRUCTION_LIST(DECLARE_FMA_YMM_INSTRUCTION)
#undef DECLARE_FMA_YMM_INSTRUCTION
 
  void vmovd(XMMRegister dst, Register src);
  void vmovd(XMMRegister dst, Operand src);
  void vmovd(Register dst, XMMRegister src);
  void vmovq(XMMRegister dst, Register src);
  void vmovq(XMMRegister dst, Operand src);
  void vmovq(Register dst, XMMRegister src);
 
  void vmovsd(XMMRegister dst, XMMRegister src1, XMMRegister src2) {
    vsd(0x10, dst, src1, src2);
  }
  void vmovsd(XMMRegister dst, Operand src) { vsd(0x10, dst, xmm0, src); }
  void vmovsd(Operand dst, XMMRegister src) { vsd(0x11, src, xmm0, dst); }
  void vmovdqa(XMMRegister dst, Operand src);
  void vmovdqa(XMMRegister dst, XMMRegister src);
  void vmovdqa(YMMRegister dst, Operand src);
  void vmovdqa(YMMRegister dst, YMMRegister src);
  void vmovdqu(XMMRegister dst, Operand src);
  void vmovdqu(Operand dst, XMMRegister src);
  void vmovdqu(XMMRegister dst, XMMRegister src);
  void vmovdqu(YMMRegister dst, Operand src);
  void vmovdqu(Operand dst, YMMRegister src);
  void vmovdqu(YMMRegister dst, YMMRegister src);
 
  void vmovlps(XMMRegister dst, XMMRegister src1, Operand src2);
  void vmovlps(Operand dst, XMMRegister src);
 
  void vmovhps(XMMRegister dst, XMMRegister src1, Operand src2);
  void vmovhps(Operand dst, XMMRegister src);
 
#define AVX_SSE_UNOP(instr, escape, opcode)          \
  void v##instr(XMMRegister dst, XMMRegister src2) { \
    vps(0x##opcode, dst, xmm0, src2);                \
  }                                                  \
  void v##instr(XMMRegister dst, Operand src2) {     \
    vps(0x##opcode, dst, xmm0, src2);                \
  }                                                  \
  void v##instr(YMMRegister dst, YMMRegister src2) { \
    vps(0x##opcode, dst, ymm0, src2);                \
  }                                                  \
  void v##instr(YMMRegister dst, Operand src2) {     \
    vps(0x##opcode, dst, ymm0, src2);                \
  }
  SSE_UNOP_INSTRUCTION_LIST(AVX_SSE_UNOP)
#undef AVX_SSE_UNOP
 
#define AVX_SSE_BINOP(instr, escape, opcode)                           \
  void v##instr(XMMRegister dst, XMMRegister src1, XMMRegister src2) { \
    vps(0x##opcode, dst, src1, src2);                                  \
  }                                                                    \
  void v##instr(XMMRegister dst, XMMRegister src1, Operand src2) {     \
    vps(0x##opcode, dst, src1, src2);                                  \
  }                                                                    \
  void v##instr(YMMRegister dst, YMMRegister src1, YMMRegister src2) { \
    vps(0x##opcode, dst, src1, src2);                                  \
  }                                                                    \
  void v##instr(YMMRegister dst, YMMRegister src1, Operand src2) {     \
    vps(0x##opcode, dst, src1, src2);                                  \
  }
  SSE_BINOP_INSTRUCTION_LIST(AVX_SSE_BINOP)
#undef AVX_SSE_BINOP
 
#define AVX_3(instr, opcode, impl, SIMDRegister)                       \
  void instr(SIMDRegister dst, SIMDRegister src1, SIMDRegister src2) { \
    impl(opcode, dst, src1, src2);                                     \
  }                                                                    \
  void instr(SIMDRegister dst, SIMDRegister src1, Operand src2) {      \
    impl(opcode, dst, src1, src2);                                     \
  }
 
  AVX_3(vhaddps, 0x7c, vsd, XMMRegister)
  AVX_3(vhaddps, 0x7c, vsd, YMMRegister)
 
#define AVX_SCALAR(instr, prefix, escape, opcode)                      \
  void v##instr(XMMRegister dst, XMMRegister src1, XMMRegister src2) { \
    vinstr(0x##opcode, dst, src1, src2, k##prefix, k##escape, kWIG);   \
  }                                                                    \
  void v##instr(XMMRegister dst, XMMRegister src1, Operand src2) {     \
    vinstr(0x##opcode, dst, src1, src2, k##prefix, k##escape, kWIG);   \
  }
  SSE_INSTRUCTION_LIST_SS(AVX_SCALAR)
  SSE2_INSTRUCTION_LIST_SD(AVX_SCALAR)
#undef AVX_SCALAR
 
#undef AVX_3
 
#define AVX_SSE2_SHIFT_IMM(instr, prefix, escape, opcode, extension)   \
  void v##instr(XMMRegister dst, XMMRegister src, uint8_t imm8) {      \
    XMMRegister ext_reg = XMMRegister::from_code(extension);           \
    vinstr(0x##opcode, ext_reg, dst, src, k##prefix, k##escape, kWIG); \
    emit(imm8);                                                        \
  }                                                                    \
                                                                       \
  void v##instr(YMMRegister dst, YMMRegister src, uint8_t imm8) {      \
    YMMRegister ext_reg = YMMRegister::from_code(extension);           \
    vinstr(0x##opcode, ext_reg, dst, src, k##prefix, k##escape, kWIG); \
    emit(imm8);                                                        \
  }
  SSE2_INSTRUCTION_LIST_SHIFT_IMM(AVX_SSE2_SHIFT_IMM)
#undef AVX_SSE2_SHIFT_IMM
 
  void vmovlhps(XMMRegister dst, XMMRegister src1, XMMRegister src2) {
    vinstr(0x16, dst, src1, src2, kNoPrefix, k0F, kWIG);
  }
  void vmovhlps(XMMRegister dst, XMMRegister src1, XMMRegister src2) {
    vinstr(0x12, dst, src1, src2, kNoPrefix, k0F, kWIG);
  }
  void vcvtdq2pd(XMMRegister dst, XMMRegister src) {
    vinstr(0xe6, dst, xmm0, src, kF3, k0F, kWIG);
  }
  void vcvtdq2pd(YMMRegister dst, XMMRegister src) {
    vinstr(0xe6, dst, xmm0, src, kF3, k0F, kWIG, AVX);
  }
  void vcvtdq2pd(YMMRegister dst, Operand src) {
    vinstr(0xe6, dst, xmm0, src, kF3, k0F, kWIG, AVX);
  }
  void vcvttps2dq(XMMRegister dst, XMMRegister src) {
    vinstr(0x5b, dst, xmm0, src, kF3, k0F, kWIG);
  }
  void vcvttps2dq(YMMRegister dst, YMMRegister src) {
    vinstr(0x5b, dst, ymm0, src, kF3, k0F, kWIG, AVX);
  }
  void vcvttps2dq(YMMRegister dst, Operand src) {
    vinstr(0x5b, dst, ymm0, src, kF3, k0F, kWIG, AVX);
  }
  void vcvtlsi2sd(XMMRegister dst, XMMRegister src1, Register src2) {
    XMMRegister isrc2 = XMMRegister::from_code(src2.code());
    vinstr(0x2a, dst, src1, isrc2, kF2, k0F, kW0);
  }
  void vcvtlsi2sd(XMMRegister dst, XMMRegister src1, Operand src2) {
    vinstr(0x2a, dst, src1, src2, kF2, k0F, kW0);
  }
  void vcvtlsi2ss(XMMRegister dst, XMMRegister src1, Register src2) {
    XMMRegister isrc2 = XMMRegister::from_code(src2.code());
    vinstr(0x2a, dst, src1, isrc2, kF3, k0F, kW0);
  }
  void vcvtlsi2ss(XMMRegister dst, XMMRegister src1, Operand src2) {
    vinstr(0x2a, dst, src1, src2, kF3, k0F, kW0);
  }
  void vcvtqsi2ss(XMMRegister dst, XMMRegister src1, Register src2) {
    XMMRegister isrc2 = XMMRegister::from_code(src2.code());
    vinstr(0x2a, dst, src1, isrc2, kF3, k0F, kW1);
  }
  void vcvtqsi2ss(XMMRegister dst, XMMRegister src1, Operand src2) {
    vinstr(0x2a, dst, src1, src2, kF3, k0F, kW1);
  }
  void vcvtqsi2sd(XMMRegister dst, XMMRegister src1, Register src2) {
    XMMRegister isrc2 = XMMRegister::from_code(src2.code());
    vinstr(0x2a, dst, src1, isrc2, kF2, k0F, kW1);
  }
  void vcvtqsi2sd(XMMRegister dst, XMMRegister src1, Operand src2) {
    vinstr(0x2a, dst, src1, src2, kF2, k0F, kW1);
  }
  void vcvttss2si(Register dst, XMMRegister src) {
    XMMRegister idst = XMMRegister::from_code(dst.code());
    vinstr(0x2c, idst, xmm0, src, kF3, k0F, kW0);
  }
  void vcvttss2si(Register dst, Operand src) {
    XMMRegister idst = XMMRegister::from_code(dst.code());
    vinstr(0x2c, idst, xmm0, src, kF3, k0F, kW0);
  }
  void vcvttsd2si(Register dst, XMMRegister src) {
    XMMRegister idst = XMMRegister::from_code(dst.code());
    vinstr(0x2c, idst, xmm0, src, kF2, k0F, kW0);
  }
  void vcvttsd2si(Register dst, Operand src) {
    XMMRegister idst = XMMRegister::from_code(dst.code());
    vinstr(0x2c, idst, xmm0, src, kF2, k0F, kW0);
  }
  void vcvttss2siq(Register dst, XMMRegister src) {
    XMMRegister idst = XMMRegister::from_code(dst.code());
    vinstr(0x2c, idst, xmm0, src, kF3, k0F, kW1);
  }
  void vcvttss2siq(Register dst, Operand src) {
    XMMRegister idst = XMMRegister::from_code(dst.code());
    vinstr(0x2c, idst, xmm0, src, kF3, k0F, kW1);
  }
  void vcvttsd2siq(Register dst, XMMRegister src) {
    XMMRegister idst = XMMRegister::from_code(dst.code());
    vinstr(0x2c, idst, xmm0, src, kF2, k0F, kW1);
  }
  void vcvttsd2siq(Register dst, Operand src) {
    XMMRegister idst = XMMRegister::from_code(dst.code());
    vinstr(0x2c, idst, xmm0, src, kF2, k0F, kW1);
  }
  void vcvtsd2si(Register dst, XMMRegister src) {
    XMMRegister idst = XMMRegister::from_code(dst.code());
    vinstr(0x2d, idst, xmm0, src, kF2, k0F, kW0);
  }
  void vroundss(XMMRegister dst, XMMRegister src1, XMMRegister src2,
                RoundingMode mode) {
    vinstr(0x0a, dst, src1, src2, k66, k0F3A, kWIG);
    emit(static_cast<uint8_t>(mode) | 0x8);  // Mask precision exception.
  }
  void vroundss(XMMRegister dst, XMMRegister src1, Operand src2,
                RoundingMode mode) {
    vinstr(0x0a, dst, src1, src2, k66, k0F3A, kWIG);
    emit(static_cast<uint8_t>(mode) | 0x8);  // Mask precision exception.
  }
  void vroundsd(XMMRegister dst, XMMRegister src1, XMMRegister src2,
                RoundingMode mode) {
    vinstr(0x0b, dst, src1, src2, k66, k0F3A, kWIG);
    emit(static_cast<uint8_t>(mode) | 0x8);  // Mask precision exception.
  }
  void vroundsd(XMMRegister dst, XMMRegister src1, Operand src2,
                RoundingMode mode) {
    vinstr(0x0b, dst, src1, src2, k66, k0F3A, kWIG);
    emit(static_cast<uint8_t>(mode) | 0x8);  // Mask precision exception.
  }
  void vroundps(XMMRegister dst, XMMRegister src, RoundingMode mode) {
    vinstr(0x08, dst, xmm0, src, k66, k0F3A, kWIG);
    emit(static_cast<uint8_t>(mode) | 0x8);  // Mask precision exception.
  }
  void vroundps(YMMRegister dst, YMMRegister src, RoundingMode mode) {
    vinstr(0x08, dst, ymm0, src, k66, k0F3A, kWIG, AVX);
    emit(static_cast<uint8_t>(mode) | 0x8);  // Mask precision exception.
  }
  void vroundpd(XMMRegister dst, XMMRegister src, RoundingMode mode) {
    vinstr(0x09, dst, xmm0, src, k66, k0F3A, kWIG);
    emit(static_cast<uint8_t>(mode) | 0x8);  // Mask precision exception.
  }
  void vroundpd(YMMRegister dst, YMMRegister src, RoundingMode mode) {
    vinstr(0x09, dst, ymm0, src, k66, k0F3A, kWIG, AVX);
    emit(static_cast<uint8_t>(mode) | 0x8);  // Mask precision exception.
  }
 
  template <typename Reg, typename Op>
  void vsd(uint8_t op, Reg dst, Reg src1, Op src2) {
    vinstr(op, dst, src1, src2, kF2, k0F, kWIG, AVX);
  }
 
  void vmovss(XMMRegister dst, XMMRegister src1, XMMRegister src2) {
    vss(0x10, dst, src1, src2);
  }
  void vmovss(XMMRegister dst, Operand src) { vss(0x10, dst, xmm0, src); }
  void vmovss(Operand dst, XMMRegister src) { vss(0x11, src, xmm0, dst); }
  void vucomiss(XMMRegister dst, XMMRegister src);
  void vucomiss(XMMRegister dst, Operand src);
  void vss(uint8_t op, XMMRegister dst, XMMRegister src1, XMMRegister src2);
  void vss(uint8_t op, XMMRegister dst, XMMRegister src1, Operand src2);
 
  void vshufps(XMMRegister dst, XMMRegister src1, XMMRegister src2,
               uint8_t imm8) {
    vps(0xC6, dst, src1, src2, imm8);
  }
  void vshufps(YMMRegister dst, YMMRegister src1, YMMRegister src2,
               uint8_t imm8) {
    vps(0xC6, dst, src1, src2, imm8);
  }
 
  void vmovaps(XMMRegister dst, XMMRegister src) { vps(0x28, dst, xmm0, src); }
  void vmovaps(YMMRegister dst, YMMRegister src) { vps(0x28, dst, ymm0, src); }
  void vmovaps(XMMRegister dst, Operand src) { vps(0x28, dst, xmm0, src); }
  void vmovaps(YMMRegister dst, Operand src) { vps(0x28, dst, ymm0, src); }
  void vmovups(XMMRegister dst, XMMRegister src) { vps(0x10, dst, xmm0, src); }
  void vmovups(YMMRegister dst, YMMRegister src) { vps(0x10, dst, ymm0, src); }
  void vmovups(XMMRegister dst, Operand src) { vps(0x10, dst, xmm0, src); }
  void vmovups(YMMRegister dst, Operand src) { vps(0x10, dst, ymm0, src); }
  void vmovups(Operand dst, XMMRegister src) { vps(0x11, src, xmm0, dst); }
  void vmovups(Operand dst, YMMRegister src) { vps(0x11, src, ymm0, dst); }
  void vmovapd(XMMRegister dst, XMMRegister src) { vpd(0x28, dst, xmm0, src); }
  void vmovapd(YMMRegister dst, YMMRegister src) { vpd(0x28, dst, ymm0, src); }
  void vmovupd(XMMRegister dst, Operand src) { vpd(0x10, dst, xmm0, src); }
  void vmovupd(YMMRegister dst, Operand src) { vpd(0x10, dst, ymm0, src); }
  void vmovupd(Operand dst, XMMRegister src) { vpd(0x11, src, xmm0, dst); }
  void vmovupd(Operand dst, YMMRegister src) { vpd(0x11, src, ymm0, dst); }
  void vmovmskps(Register dst, XMMRegister src) {
    XMMRegister idst = XMMRegister::from_code(dst.code());
    vps(0x50, idst, xmm0, src);
  }
  void vmovmskpd(Register dst, XMMRegister src) {
    XMMRegister idst = XMMRegister::from_code(dst.code());
    vpd(0x50, idst, xmm0, src);
  }
  void vpmovmskb(Register dst, XMMRegister src);
  void vcmpeqss(XMMRegister dst, XMMRegister src) {
    vss(0xC2, dst, dst, src);
    emit(0x00);  // EQ == 0
  }
  void vcmpeqsd(XMMRegister dst, XMMRegister src) {
    vsd(0xC2, dst, dst, src);
    emit(0x00);  // EQ == 0
  }
  void vcmpps(XMMRegister dst, XMMRegister src1, XMMRegister src2, int8_t cmp) {
    vps(0xC2, dst, src1, src2);
    emit(cmp);
  }
  void vcmpps(YMMRegister dst, YMMRegister src1, YMMRegister src2, int8_t cmp) {
    vps(0xC2, dst, src1, src2);
    emit(cmp);
  }
  void vcmpps(XMMRegister dst, XMMRegister src1, Operand src2, int8_t cmp) {
    vps(0xC2, dst, src1, src2);
    emit(cmp);
  }
  void vcmpps(YMMRegister dst, YMMRegister src1, Operand src2, int8_t cmp) {
    vps(0xC2, dst, src1, src2);
    emit(cmp);
  }
  void vcmppd(XMMRegister dst, XMMRegister src1, XMMRegister src2, int8_t cmp) {
    vpd(0xC2, dst, src1, src2);
    emit(cmp);
  }
  void vcmppd(YMMRegister dst, YMMRegister src1, YMMRegister src2, int8_t cmp) {
    vpd(0xC2, dst, src1, src2);
    emit(cmp);
  }
  void vcmppd(XMMRegister dst, XMMRegister src1, Operand src2, int8_t cmp) {
    vpd(0xC2, dst, src1, src2);
    emit(cmp);
  }
  void vcmppd(YMMRegister dst, YMMRegister src1, Operand src2, int8_t cmp) {
    vpd(0xC2, dst, src1, src2);
    emit(cmp);
  }
#define AVX_CMP_P(instr, imm8, SIMDRegister)                               \
  void instr##ps(SIMDRegister dst, SIMDRegister src1, SIMDRegister src2) { \
    vcmpps(dst, src1, src2, imm8);                                         \
  }                                                                        \
  void instr##ps(SIMDRegister dst, SIMDRegister src1, Operand src2) {      \
    vcmpps(dst, src1, src2, imm8);                                         \
  }                                                                        \
  void instr##pd(SIMDRegister dst, SIMDRegister src1, SIMDRegister src2) { \
    vcmppd(dst, src1, src2, imm8);                                         \
  }                                                                        \
  void instr##pd(SIMDRegister dst, SIMDRegister src1, Operand src2) {      \
    vcmppd(dst, src1, src2, imm8);                                         \
  }
 
  AVX_CMP_P(vcmpeq, 0x0, XMMRegister)
  AVX_CMP_P(vcmpeq, 0x0, YMMRegister)
  AVX_CMP_P(vcmplt, 0x1, XMMRegister)
  AVX_CMP_P(vcmplt, 0x1, YMMRegister)
  AVX_CMP_P(vcmple, 0x2, XMMRegister)
  AVX_CMP_P(vcmple, 0x2, YMMRegister)
  AVX_CMP_P(vcmpunord, 0x3, XMMRegister)
  AVX_CMP_P(vcmpunord, 0x3, YMMRegister)
  AVX_CMP_P(vcmpneq, 0x4, XMMRegister)
  AVX_CMP_P(vcmpneq, 0x4, YMMRegister)
  AVX_CMP_P(vcmpnlt, 0x5, XMMRegister)
  AVX_CMP_P(vcmpnlt, 0x5, YMMRegister)
  AVX_CMP_P(vcmpnle, 0x6, XMMRegister)
  AVX_CMP_P(vcmpnle, 0x6, YMMRegister)
  AVX_CMP_P(vcmpge, 0xd, XMMRegister)
  AVX_CMP_P(vcmpge, 0xd, YMMRegister)
 
#undef AVX_CMP_P
 
  void vlddqu(XMMRegister dst, Operand src) {
    vinstr(0xF0, dst, xmm0, src, kF2, k0F, kWIG);
  }
  void vinsertps(XMMRegister dst, XMMRegister src1, XMMRegister src2,
                 uint8_t imm8) {
    vinstr(0x21, dst, src1, src2, k66, k0F3A, kWIG);
    emit(imm8);
  }
  void vinsertps(XMMRegister dst, XMMRegister src1, Operand src2,
                 uint8_t imm8) {
    vinstr(0x21, dst, src1, src2, k66, k0F3A, kWIG);
    emit(imm8);
  }
  void vpextrq(Register dst, XMMRegister src, int8_t imm8) {
    XMMRegister idst = XMMRegister::from_code(dst.code());
    vinstr(0x16, src, xmm0, idst, k66, k0F3A, kW1);
    emit(imm8);
  }
  void vpinsrb(XMMRegister dst, XMMRegister src1, Register src2, uint8_t imm8) {
    XMMRegister isrc = XMMRegister::from_code(src2.code());
    vinstr(0x20, dst, src1, isrc, k66, k0F3A, kW0);
    emit(imm8);
  }
  void vpinsrb(XMMRegister dst, XMMRegister src1, Operand src2, uint8_t imm8) {
    vinstr(0x20, dst, src1, src2, k66, k0F3A, kW0);
    emit(imm8);
  }
  void vpinsrw(XMMRegister dst, XMMRegister src1, Register src2, uint8_t imm8) {
    XMMRegister isrc = XMMRegister::from_code(src2.code());
    vinstr(0xc4, dst, src1, isrc, k66, k0F, kW0);
    emit(imm8);
  }
  void vpinsrw(XMMRegister dst, XMMRegister src1, Operand src2, uint8_t imm8) {
    vinstr(0xc4, dst, src1, src2, k66, k0F, kW0);
    emit(imm8);
  }
  void vpinsrd(XMMRegister dst, XMMRegister src1, Register src2, uint8_t imm8) {
    XMMRegister isrc = XMMRegister::from_code(src2.code());
    vinstr(0x22, dst, src1, isrc, k66, k0F3A, kW0);
    emit(imm8);
  }
  void vpinsrd(XMMRegister dst, XMMRegister src1, Operand src2, uint8_t imm8) {
    vinstr(0x22, dst, src1, src2, k66, k0F3A, kW0);
    emit(imm8);
  }
  void vpinsrq(XMMRegister dst, XMMRegister src1, Register src2, uint8_t imm8) {
    XMMRegister isrc = XMMRegister::from_code(src2.code());
    vinstr(0x22, dst, src1, isrc, k66, k0F3A, kW1);
    emit(imm8);
  }
  void vpinsrq(XMMRegister dst, XMMRegister src1, Operand src2, uint8_t imm8) {
    vinstr(0x22, dst, src1, src2, k66, k0F3A, kW1);
    emit(imm8);
  }
 
  void vpshufd(XMMRegister dst, XMMRegister src, uint8_t imm8) {
    vinstr(0x70, dst, xmm0, src, k66, k0F, kWIG);
    emit(imm8);
  }
  void vpshufd(YMMRegister dst, YMMRegister src, uint8_t imm8) {
    vinstr(0x70, dst, ymm0, src, k66, k0F, kWIG);
    emit(imm8);
  }
  void vpshufd(XMMRegister dst, Operand src, uint8_t imm8) {
    vinstr(0x70, dst, xmm0, src, k66, k0F, kWIG);
    emit(imm8);
  }
  void vpshufd(YMMRegister dst, Operand src, uint8_t imm8) {
    vinstr(0x70, dst, ymm0, src, k66, k0F, kWIG);
    emit(imm8);
  }
  void vpshuflw(XMMRegister dst, XMMRegister src, uint8_t imm8) {
    vinstr(0x70, dst, xmm0, src, kF2, k0F, kWIG);
    emit(imm8);
  }
  void vpshuflw(YMMRegister dst, YMMRegister src, uint8_t imm8) {
    vinstr(0x70, dst, ymm0, src, kF2, k0F, kWIG);
    emit(imm8);
  }
  void vpshuflw(XMMRegister dst, Operand src, uint8_t imm8) {
    vinstr(0x70, dst, xmm0, src, kF2, k0F, kWIG);
    emit(imm8);
  }
  void vpshuflw(YMMRegister dst, Operand src, uint8_t imm8) {
    vinstr(0x70, dst, ymm0, src, kF2, k0F, kWIG);
    emit(imm8);
  }
  void vpshufhw(XMMRegister dst, XMMRegister src, uint8_t imm8) {
    vinstr(0x70, dst, xmm0, src, kF3, k0F, kWIG);
    emit(imm8);
  }
  void vpshufhw(YMMRegister dst, YMMRegister src, uint8_t imm8) {
    vinstr(0x70, dst, ymm0, src, kF3, k0F, kWIG);
    emit(imm8);
  }
  void vpshufhw(XMMRegister dst, Operand src, uint8_t imm8) {
    vinstr(0x70, dst, xmm0, src, kF3, k0F, kWIG);
    emit(imm8);
  }
  void vpshufhw(YMMRegister dst, Operand src, uint8_t imm8) {
    vinstr(0x70, dst, ymm0, src, kF3, k0F, kWIG);
    emit(imm8);
  }
 
  void vpblendw(XMMRegister dst, XMMRegister src1, XMMRegister src2,
                uint8_t mask) {
    vinstr(0x0E, dst, src1, src2, k66, k0F3A, kWIG);
    emit(mask);
  }
  void vpblendw(YMMRegister dst, YMMRegister src1, YMMRegister src2,
                uint8_t mask) {
    vinstr(0x0E, dst, src1, src2, k66, k0F3A, kWIG);
    emit(mask);
  }
  void vpblendw(XMMRegister dst, XMMRegister src1, Operand src2, uint8_t mask) {
    vinstr(0x0E, dst, src1, src2, k66, k0F3A, kWIG);
    emit(mask);
  }
  void vpblendw(YMMRegister dst, YMMRegister src1, Operand src2, uint8_t mask) {
    vinstr(0x0E, dst, src1, src2, k66, k0F3A, kWIG);
    emit(mask);
  }
 
  void vpalignr(XMMRegister dst, XMMRegister src1, XMMRegister src2,
                uint8_t imm8) {
    vinstr(0x0F, dst, src1, src2, k66, k0F3A, kWIG);
    emit(imm8);
  }
  void vpalignr(YMMRegister dst, YMMRegister src1, YMMRegister src2,
                uint8_t imm8) {
    vinstr(0x0F, dst, src1, src2, k66, k0F3A, kWIG);
    emit(imm8);
  }
  void vpalignr(XMMRegister dst, XMMRegister src1, Operand src2, uint8_t imm8) {
    vinstr(0x0F, dst, src1, src2, k66, k0F3A, kWIG);
    emit(imm8);
  }
  void vpalignr(YMMRegister dst, YMMRegister src1, Operand src2, uint8_t imm8) {
    vinstr(0x0F, dst, src1, src2, k66, k0F3A, kWIG);
    emit(imm8);
  }
 
  void vps(uint8_t op, XMMRegister dst, XMMRegister src1, XMMRegister src2);
  void vps(uint8_t op, YMMRegister dst, YMMRegister src1, YMMRegister src2);
  void vps(uint8_t op, XMMRegister dst, XMMRegister src1, Operand src2);
  void vps(uint8_t op, YMMRegister dst, YMMRegister src1, Operand src2);
  void vps(uint8_t op, XMMRegister dst, XMMRegister src1, XMMRegister src2,
           uint8_t imm8);
  void vps(uint8_t op, YMMRegister dst, YMMRegister src1, YMMRegister src2,
           uint8_t imm8);
  void vpd(uint8_t op, XMMRegister dst, XMMRegister src1, XMMRegister src2);
  void vpd(uint8_t op, YMMRegister dst, YMMRegister src1, YMMRegister src2);
  void vpd(uint8_t op, XMMRegister dst, YMMRegister src1, YMMRegister src2);
  void vpd(uint8_t op, XMMRegister dst, XMMRegister src1, Operand src2);
  void vpd(uint8_t op, YMMRegister dst, YMMRegister src1, Operand src2);
  void vpd(uint8_t op, XMMRegister dst, YMMRegister src1, Operand src2);
 
  // AVX2 instructions
#define AVX2_INSTRUCTION(instr, prefix, escape1, escape2, opcode)           \
  template <typename Reg, typename Op>                                      \
  void instr(Reg dst, Op src) {                                             \
    vinstr(0x##opcode, dst, xmm0, src, k##prefix, k##escape1##escape2, kW0, \
           AVX2);                                                           \
  }
  AVX2_BROADCAST_LIST(AVX2_INSTRUCTION)
#undef AVX2_INSTRUCTION
 
  // F16C Instructions.
  void vcvtph2ps(XMMRegister dst, XMMRegister src);
  void vcvtph2ps(YMMRegister dst, XMMRegister src);
  void vcvtps2ph(XMMRegister dst, XMMRegister src, uint8_t imm8);
  void vcvtps2ph(XMMRegister dst, YMMRegister src, uint8_t imm8);
 
  // AVX-VNNI instruction
  void vpdpbusd(XMMRegister dst, XMMRegister src1, XMMRegister src2) {
    vinstr(0x50, dst, src1, src2, k66, k0F38, kW0, AVX_VNNI);
  }
  void vpdpbusd(YMMRegister dst, YMMRegister src1, YMMRegister src2) {
    vinstr(0x50, dst, src1, src2, k66, k0F38, kW0, AVX_VNNI);
  }
 
  // AVX-VNNI-INT8 instruction
  void vpdpbssd(XMMRegister dst, XMMRegister src1, XMMRegister src2) {
    vinstr(0x50, dst, src1, src2, kF2, k0F38, kW0, AVX_VNNI_INT8);
  }
  void vpdpbssd(YMMRegister dst, YMMRegister src1, YMMRegister src2) {
    vinstr(0x50, dst, src1, src2, kF2, k0F38, kW0, AVX_VNNI_INT8);
  }
 
  // BMI instruction
  void andnq(Register dst, Register src1, Register src2) {
    bmi1q(0xf2, dst, src1, src2);
  }
  void andnq(Register dst, Register src1, Operand src2) {
    bmi1q(0xf2, dst, src1, src2);
  }
  void andnl(Register dst, Register src1, Register src2) {
    bmi1l(0xf2, dst, src1, src2);
  }
  void andnl(Register dst, Register src1, Operand src2) {
    bmi1l(0xf2, dst, src1, src2);
  }
  void bextrq(Register dst, Register src1, Register src2) {
    bmi1q(0xf7, dst, src2, src1);
  }
  void bextrq(Register dst, Operand src1, Register src2) {
    bmi1q(0xf7, dst, src2, src1);
  }
  void bextrl(Register dst, Register src1, Register src2) {
    bmi1l(0xf7, dst, src2, src1);
  }
  void bextrl(Register dst, Operand src1, Register src2) {
    bmi1l(0xf7, dst, src2, src1);
  }
  void blsiq(Register dst, Register src) { bmi1q(0xf3, rbx, dst, src); }
  void blsiq(Register dst, Operand src) { bmi1q(0xf3, rbx, dst, src); }
  void blsil(Register dst, Register src) { bmi1l(0xf3, rbx, dst, src); }
  void blsil(Register dst, Operand src) { bmi1l(0xf3, rbx, dst, src); }
  void blsmskq(Register dst, Register src) { bmi1q(0xf3, rdx, dst, src); }
  void blsmskq(Register dst, Operand src) { bmi1q(0xf3, rdx, dst, src); }
  void blsmskl(Register dst, Register src) { bmi1l(0xf3, rdx, dst, src); }
  void blsmskl(Register dst, Operand src) { bmi1l(0xf3, rdx, dst, src); }
  void blsrq(Register dst, Register src) { bmi1q(0xf3, rcx, dst, src); }
  void blsrq(Register dst, Operand src) { bmi1q(0xf3, rcx, dst, src); }
  void blsrl(Register dst, Register src) { bmi1l(0xf3, rcx, dst, src); }
  void blsrl(Register dst, Operand src) { bmi1l(0xf3, rcx, dst, src); }
  void tzcntq(Register dst, Register src);
  void tzcntq(Register dst, Operand src);
  void tzcntl(Register dst, Register src);
  void tzcntl(Register dst, Operand src);
 
  void lzcntq(Register dst, Register src);
  void lzcntq(Register dst, Operand src);
  void lzcntl(Register dst, Register src);
  void lzcntl(Register dst, Operand src);
 
  void popcntq(Register dst, Register src);
  void popcntq(Register dst, Operand src);
  void popcntl(Register dst, Register src);
  void popcntl(Register dst, Operand src);
 
  void bzhiq(Register dst, Register src1, Register src2) {
    bmi2q(kNoPrefix, 0xf5, dst, src2, src1);
  }
  void bzhiq(Register dst, Operand src1, Register src2) {
    bmi2q(kNoPrefix, 0xf5, dst, src2, src1);
  }
  void bzhil(Register dst, Register src1, Register src2) {
    bmi2l(kNoPrefix, 0xf5, dst, src2, src1);
  }
  void bzhil(Register dst, Operand src1, Register src2) {
    bmi2l(kNoPrefix, 0xf5, dst, src2, src1);
  }
  void mulxq(Register dst1, Register dst2, Register src) {
    bmi2q(kF2, 0xf6, dst1, dst2, src);
  }
  void mulxq(Register dst1, Register dst2, Operand src) {
    bmi2q(kF2, 0xf6, dst1, dst2, src);
  }
  void mulxl(Register dst1, Register dst2, Register src) {
    bmi2l(kF2, 0xf6, dst1, dst2, src);
  }
  void mulxl(Register dst1, Register dst2, Operand src) {
    bmi2l(kF2, 0xf6, dst1, dst2, src);
  }
  void pdepq(Register dst, Register src1, Register src2) {
    bmi2q(kF2, 0xf5, dst, src1, src2);
  }
  void pdepq(Register dst, Register src1, Operand src2) {
    bmi2q(kF2, 0xf5, dst, src1, src2);
  }
  void pdepl(Register dst, Register src1, Register src2) {
    bmi2l(kF2, 0xf5, dst, src1, src2);
  }
  void pdepl(Register dst, Register src1, Operand src2) {
    bmi2l(kF2, 0xf5, dst, src1, src2);
  }
  void pextq(Register dst, Register src1, Register src2) {
    bmi2q(kF3, 0xf5, dst, src1, src2);
  }
  void pextq(Register dst, Register src1, Operand src2) {
    bmi2q(kF3, 0xf5, dst, src1, src2);
  }
  void pextl(Register dst, Register src1, Register src2) {
    bmi2l(kF3, 0xf5, dst, src1, src2);
  }
  void pextl(Register dst, Register src1, Operand src2) {
    bmi2l(kF3, 0xf5, dst, src1, src2);
  }
  void sarxq(Register dst, Register src1, Register src2) {
    bmi2q(kF3, 0xf7, dst, src2, src1);
  }
  void sarxq(Register dst, Operand src1, Register src2) {
    bmi2q(kF3, 0xf7, dst, src2, src1);
  }
  void sarxl(Register dst, Register src1, Register src2) {
    bmi2l(kF3, 0xf7, dst, src2, src1);
  }
  void sarxl(Register dst, Operand src1, Register src2) {
    bmi2l(kF3, 0xf7, dst, src2, src1);
  }
  void shlxq(Register dst, Register src1, Register src2) {
    bmi2q(k66, 0xf7, dst, src2, src1);
  }
  void shlxq(Register dst, Operand src1, Register src2) {
    bmi2q(k66, 0xf7, dst, src2, src1);
  }
  void shlxl(Register dst, Register src1, Register src2) {
    bmi2l(k66, 0xf7, dst, src2, src1);
  }
  void shlxl(Register dst, Operand src1, Register src2) {
    bmi2l(k66, 0xf7, dst, src2, src1);
  }
  void shrxq(Register dst, Register src1, Register src2) {
    bmi2q(kF2, 0xf7, dst, src2, src1);
  }
  void shrxq(Register dst, Operand src1, Register src2) {
    bmi2q(kF2, 0xf7, dst, src2, src1);
  }
  void shrxl(Register dst, Register src1, Register src2) {
    bmi2l(kF2, 0xf7, dst, src2, src1);
  }
  void shrxl(Register dst, Operand src1, Register src2) {
    bmi2l(kF2, 0xf7, dst, src2, src1);
  }
  void rorxq(Register dst, Register src, uint8_t imm8);
  void rorxq(Register dst, Operand src, uint8_t imm8);
  void rorxl(Register dst, Register src, uint8_t imm8);
  void rorxl(Register dst, Operand src, uint8_t imm8);
 
  void mfence();
  void lfence();
  void pause();
 
  // Check the code size generated from label to here.
  int SizeOfCodeGeneratedSince(Label* label) {
    return pc_offset() - label->pos();
  }
 
  // Record a deoptimization reason that can be used by a log or cpu profiler.
  // Use --trace-deopt to enable.
  void RecordDeoptReason(DeoptimizeReason reason, uint32_t node_id,
                         SourcePosition position, int id);
 
  // Writes a single word of data in the code stream.
  // Used for inline tables, e.g., jump-tables.
  void db(uint8_t data);
  void dd(uint32_t data);
  void dq(uint64_t data);
  void dp(uintptr_t data) { dq(data); }
  void dq(Label* label);
 
  void WriteBuiltinJumpTableEntry(Label* label, int table_pos);
 
  // Patch entries for partial constant pool.
  void PatchConstPool();
 
  // Check if use partial constant pool for this rmode.
  static bool UseConstPoolFor(RelocInfo::Mode rmode);
 
  // Check if there is less than kGap bytes available in the buffer.
  // If this is the case, we need to grow the buffer before emitting
  // an instruction or relocation information.
  bool buffer_overflow() const { return available_space() < kGap; }
 
  // Get the number of bytes available in the buffer.
  int available_space() const {
    DCHECK_GE(reloc_info_writer.pos(), pc_);
    DCHECK_GE(kMaxInt, reloc_info_writer.pos() - pc_);
    return static_cast<int>(reloc_info_writer.pos() - pc_);
  }
 
  static bool IsNop(Address addr);
  static bool IsJmpRel(Address addr);
 
  // Avoid overflows for displacements etc.
  static constexpr int kMaximalBufferSize = 512 * MB;
 
  uint8_t byte_at(int pos) { return buffer_start_[pos]; }
  void set_byte_at(int pos, uint8_t value) { buffer_start_[pos] = value; }
 
#if defined(V8_OS_WIN_X64)
  win64_unwindinfo::BuiltinUnwindInfo GetUnwindInfo() const;
#endif
 
 protected:
  // Call near indirect
  void call(Operand operand);
 
 private:
  Address addr_at(int pos) {
    DCHECK_GE(pos, 0);
    DCHECK_LT(pos, pc_offset());
    return reinterpret_cast<Address>(buffer_start_ + pos);
  }
  uint32_t long_at(int pos) {
    return ReadUnalignedValue<uint32_t>(addr_at(pos));
  }
  void long_at_put(int pos, uint32_t x) {
    WriteUnalignedValue(addr_at(pos), x);
  }
 
  // InstructionStream emission.
  V8_NOINLINE V8_PRESERVE_MOST void GrowBuffer();
 
  template <typename T>
  static uint8_t* emit(uint8_t* __restrict pc, T t) {
    WriteUnalignedValue(reinterpret_cast<Address>(pc), t);
    return pc + sizeof(T);
  }
 
  void emit(uint8_t x) { pc_ = emit(pc_, x); }
  void emitw(uint16_t x) { pc_ = emit(pc_, x); }
  void emitl(uint32_t x) { pc_ = emit(pc_, x); }
  void emitq(uint64_t x) { pc_ = emit(pc_, x); }
 
  void emit(Immediate x) {
    if (!RelocInfo::IsNoInfo(x.rmode_)) RecordRelocInfo(x.rmode_);
    emitl(x.value_);
  }
 
  void emit(Immediate64 x) {
    if (!RelocInfo::IsNoInfo(x.rmode_)) RecordRelocInfo(x.rmode_);
    emitq(static_cast<uint64_t>(x.value_));
  }
 
  // Emits a REX prefix that encodes a 64-bit operand size and
  // the top bit of both register codes.
  // High bit of reg goes to REX.R, high bit of rm_reg goes to REX.B.
  // REX.W is set.
  inline void emit_rex_64(XMMRegister reg, Register rm_reg);
  inline void emit_rex_64(Register reg, XMMRegister rm_reg);
  inline void emit_rex_64(Register reg, Register rm_reg);
  inline void emit_rex_64(XMMRegister reg, XMMRegister rm_reg);
 
  // Emits a REX prefix that encodes a 64-bit operand size and
  // the top bit of the destination, index, and base register codes.
  // The high bit of reg is used for REX.R, the high bit of op's base
  // register is used for REX.B, and the high bit of op's index register
  // is used for REX.X.  REX.W is set.
  inline void emit_rex_64(Register reg, Operand op);
  inline void emit_rex_64(XMMRegister reg, Operand op);
 
  // Emits a REX prefix that encodes a 64-bit operand size and
  // the top bit of the register code.
  // The high bit of register is used for REX.B.
  // REX.W is set and REX.R and REX.X are clear.
  inline void emit_rex_64(Register rm_reg);
 
  // Emits a REX prefix that encodes a 64-bit operand size and
  // the top bit of the index and base register codes.
  // The high bit of op's base register is used for REX.B, and the high
  // bit of op's index register is used for REX.X.
  // REX.W is set and REX.R clear.
  inline void emit_rex_64(Operand op);
 
  // Emit a REX prefix that only sets REX.W to choose a 64-bit operand size.
  void emit_rex_64() { emit(0x48); }
 
  // High bit of reg goes to REX.R, high bit of rm_reg goes to REX.B.
  // REX.W is clear.
  inline void emit_rex_32(Register reg, Register rm_reg);
 
  // The high bit of reg is used for REX.R, the high bit of op's base
  // register is used for REX.B, and the high bit of op's index register
  // is used for REX.X.  REX.W is cleared.
  inline void emit_rex_32(Register reg, Operand op);
 
  // High bit of rm_reg goes to REX.B.
  // REX.W, REX.R and REX.X are clear.
  inline void emit_rex_32(Register rm_reg);
 
  // High bit of base goes to REX.B and high bit of index to REX.X.
  // REX.W and REX.R are clear.
  inline void emit_rex_32(Operand op);
 
  // High bit of reg goes to REX.R, high bit of rm_reg goes to REX.B.
  // REX.W is cleared.  If no REX bits are set, no byte is emitted.
  inline void emit_optional_rex_32(Register reg, Register rm_reg);
 
  // The high bit of reg is used for REX.R, the high bit of op's base
  // register is used for REX.B, and the high bit of op's index register
  // is used for REX.X.  REX.W is cleared.  If no REX bits are set, nothing
  // is emitted.
  inline void emit_optional_rex_32(Register reg, Operand op);
 
  // As for emit_optional_rex_32(Register, Register), except that
  // the registers are XMM registers.
  inline void emit_optional_rex_32(XMMRegister reg, XMMRegister base);
 
  // As for emit_optional_rex_32(Register, Register), except that
  // one of the registers is an XMM registers.
  inline void emit_optional_rex_32(XMMRegister reg, Register base);
 
  // As for emit_optional_rex_32(Register, Register), except that
  // one of the registers is an XMM registers.
  inline void emit_optional_rex_32(Register reg, XMMRegister base);
 
  // As for emit_optional_rex_32(Register, Operand), except that
  // the register is an XMM register.
  inline void emit_optional_rex_32(XMMRegister reg, Operand op);
 
  // Optionally do as emit_rex_32(Register) if the register number has
  // the high bit set.
  inline void emit_optional_rex_32(Register rm_reg);
  inline void emit_optional_rex_32(XMMRegister rm_reg);
 
  // Optionally do as emit_rex_32(Operand) if the operand register
  // numbers have a high bit set.
  inline void emit_optional_rex_32(Operand op);
 
  // Calls emit_rex_32(Register) for all non-byte registers.
  inline void emit_optional_rex_8(Register reg);
 
  // Calls emit_rex_32(Register, Operand) for all non-byte registers, and
  // emit_optional_rex_32(Register, Operand) for byte registers.
  inline void emit_optional_rex_8(Register reg, Operand op);
 
  void emit_rex(int size) {
    if (size == kInt64Size) {
      emit_rex_64();
    } else {
      DCHECK_EQ(size, kInt32Size);
    }
  }
 
  template <class P1>
  void emit_rex(P1 p1, int size) {
    if (size == kInt64Size) {
      emit_rex_64(p1);
    } else {
      DCHECK_EQ(size, kInt32Size);
      emit_optional_rex_32(p1);
    }
  }
 
  template <class P1, class P2>
  void emit_rex(P1 p1, P2 p2, int size) {
    if (size == kInt64Size) {
      emit_rex_64(p1, p2);
    } else {
      DCHECK_EQ(size, kInt32Size);
      emit_optional_rex_32(p1, p2);
    }
  }
 
  // Emit vex prefix
  void emit_vex2_byte0() { emit(0xc5); }
  inline void emit_vex2_byte1(XMMRegister reg, XMMRegister v, VectorLength l,
                              SIMDPrefix pp);
  void emit_vex3_byte0() { emit(0xc4); }
  inline void emit_vex3_byte1(XMMRegister reg, XMMRegister rm, LeadingOpcode m);
  inline void emit_vex3_byte1(XMMRegister reg, Operand rm, LeadingOpcode m);
  inline void emit_vex3_byte2(VexW w, XMMRegister v, VectorLength l,
                              SIMDPrefix pp);
  inline void emit_vex_prefix(XMMRegister reg, XMMRegister v, XMMRegister rm,
                              VectorLength l, SIMDPrefix pp, LeadingOpcode m,
                              VexW w);
  inline void emit_vex_prefix(Register reg, Register v, Register rm,
                              VectorLength l, SIMDPrefix pp, LeadingOpcode m,
                              VexW w);
  inline void emit_vex_prefix(XMMRegister reg, XMMRegister v, Operand rm,
                              VectorLength l, SIMDPrefix pp, LeadingOpcode m,
                              VexW w);
  inline void emit_vex_prefix(Register reg, Register v, Operand rm,
                              VectorLength l, SIMDPrefix pp, LeadingOpcode m,
                              VexW w);
 
  // Emit the ModR/M byte, and optionally the SIB byte and
  // 1- or 4-byte offset for a memory operand.  Also encodes
  // the second operand of the operation, a register or operation
  // subcode, into the reg field of the ModR/M byte.
  void emit_operand(Register reg, Operand adr) {
    emit_operand(reg.low_bits(), adr);
  }
 
  // Emit the ModR/M byte, and optionally the SIB byte and
  // 1- or 4-byte offset for a memory operand.
  // Also used to encode a three-bit opcode extension into the ModR/M byte.
  void emit_operand(int rm, Operand adr);
 
  // Emit a RIP-relative operand.
  // Also used to encode a three-bit opcode extension into the ModR/M byte.
  V8_NOINLINE void emit_label_operand(int rm, Label* label, int addend = 0);
 
  // Emit a ModR/M byte with registers coded in the reg and rm_reg fields.
  void emit_modrm(Register reg, Register rm_reg) {
    emit(0xC0 | reg.low_bits() << 3 | rm_reg.low_bits());
  }
 
  // Emit a ModR/M byte with an operation subcode in the reg field and
  // a register in the rm_reg field.
  void emit_modrm(int code, Register rm_reg) {
    DCHECK(is_uint3(code));
    emit(0xC0 | code << 3 | rm_reg.low_bits());
  }
 
  // Emit the code-object-relative offset of the label's position
  inline void emit_code_relative_offset(Label* label);
 
  // The first argument is the reg field, the second argument is the r/m field.
  void emit_sse_operand(XMMRegister dst, XMMRegister src);
  void emit_sse_operand(XMMRegister reg, Operand adr);
  void emit_sse_operand(Register reg, Operand adr);
  void emit_sse_operand(XMMRegister dst, Register src);
  void emit_sse_operand(Register dst, XMMRegister src);
  void emit_sse_operand(XMMRegister dst);
 
  // Emit machine code for one of the operations ADD, ADC, SUB, SBC,
  // AND, OR, XOR, or CMP.  The encodings of these operations are all
  // similar, differing just in the opcode or in the reg field of the
  // ModR/M byte.
  void arithmetic_op_8(uint8_t opcode, Register reg, Register rm_reg);
  void arithmetic_op_8(uint8_t opcode, Register reg, Operand rm_reg);
  void arithmetic_op_16(uint8_t opcode, Register reg, Register rm_reg);
  void arithmetic_op_16(uint8_t opcode, Register reg, Operand rm_reg);
  // Operate on operands/registers with pointer size, 32-bit or 64-bit size.
  void arithmetic_op(uint8_t opcode, Register reg, Register rm_reg, int size);
  void arithmetic_op(uint8_t opcode, Register reg, Operand rm_reg, int size);
  // Operate on a byte in memory or register.
  void immediate_arithmetic_op_8(uint8_t subcode, Register dst, Immediate src);
  void immediate_arithmetic_op_8(uint8_t subcode, Operand dst, Immediate src);
  // Operate on a word in memory or register.
  void immediate_arithmetic_op_16(uint8_t subcode, Register dst, Immediate src);
  void immediate_arithmetic_op_16(uint8_t subcode, Operand dst, Immediate src);
  // Operate on operands/registers with pointer size, 32-bit or 64-bit size.
  void immediate_arithmetic_op(uint8_t subcode, Register dst, Immediate src,
                               int size);
  void immediate_arithmetic_op(uint8_t subcode, Operand dst, Immediate src,
                               int size);
 
  // Emit machine code for a shift operation.
  void shift(Operand dst, Immediate shift_amount, int subcode, int size);
  void shift(Register dst, Immediate shift_amount, int subcode, int size);
  // Shift dst by cl % 64 bits.
  void shift(Register dst, int subcode, int size);
  void shift(Operand dst, int subcode, int size);
 
  void emit_farith(int b1, int b2, int i);
 
  // labels
  // void print(Label* L);
  void bind_to(Label* L, int pos);
 
  // record reloc info for current pc_
  void RecordRelocInfo(RelocInfo::Mode rmode, intptr_t data = 0);
 
  // Arithmetics
  void emit_add(Register dst, Register src, int size) {
    arithmetic_op(0x03, dst, src, size);
  }
 
  void emit_add(Register dst, Immediate src, int size) {
    immediate_arithmetic_op(0x0, dst, src, size);
  }
 
  void emit_add(Register dst, Operand src, int size) {
    arithmetic_op(0x03, dst, src, size);
  }
 
  void emit_add(Operand dst, Register src, int size) {
    arithmetic_op(0x1, src, dst, size);
  }
 
  void emit_add(Operand dst, Immediate src, int size) {
    immediate_arithmetic_op(0x0, dst, src, size);
  }
 
  void emit_and(Register dst, Register src, int size) {
    arithmetic_op(0x23, dst, src, size);
  }
 
  void emit_and(Register dst, Operand src, int size) {
    arithmetic_op(0x23, dst, src, size);
  }
 
  void emit_and(Operand dst, Register src, int size) {
    arithmetic_op(0x21, src, dst, size);
  }
 
  void emit_and(Register dst, Immediate src, int size) {
    immediate_arithmetic_op(0x4, dst, src, size);
  }
 
  void emit_and(Operand dst, Immediate src, int size) {
    immediate_arithmetic_op(0x4, dst, src, size);
  }
 
  void emit_cmp(Register dst, Register src, int size) {
    arithmetic_op(0x3B, dst, src, size);
  }
 
  // Used for JCC erratum performance mitigation.
  void emit_aligned_cmp(Register dst, Register src, int size) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* cmp */ 3 + /* jcc */ 6
    const int kMaxMacroFusionLength = 9;
    AlignForJCCErratum(kMaxMacroFusionLength);
    emit_cmp(dst, src, size);
  }
 
  void emit_cmp(Register dst, Operand src, int size) {
    arithmetic_op(0x3B, dst, src, size);
  }
 
  // Used for JCC erratum performance mitigation.
  void emit_aligned_cmp(Register dst, Operand src, int size) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* cmp */ 8 + /* jcc */ 6
    const int kMaxMacroFusionLength = 14;
    AlignForJCCErratum(kMaxMacroFusionLength);
    emit_cmp(dst, src, size);
  }
 
  void emit_cmp(Operand dst, Register src, int size) {
    arithmetic_op(0x39, src, dst, size);
  }
 
  // Used for JCC erratum performance mitigation.
  void emit_aligned_cmp(Operand dst, Register src, int size) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* cmp */ 8 + /* jcc */ 6
    const int kMaxMacroFusionLength = 14;
    AlignForJCCErratum(kMaxMacroFusionLength);
    emit_cmp(dst, src, size);
  }
 
  void emit_cmp(Register dst, Immediate src, int size) {
    immediate_arithmetic_op(0x7, dst, src, size);
  }
 
  // Used for JCC erratum performance mitigation.
  void emit_aligned_cmp(Register dst, Immediate src, int size) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* cmpl */ 7 + /* jcc */ 6
    // /* cmpq */ 11 + /* jcc */ 6
    const int kMaxMacroFusionLength = 9 + size;
    AlignForJCCErratum(kMaxMacroFusionLength);
    emit_cmp(dst, src, size);
  }
 
  void emit_cmp(Operand dst, Immediate src, int size) {
    immediate_arithmetic_op(0x7, dst, src, size);
  }
 
  // Used for JCC erratum performance mitigation.
  void emit_aligned_cmp(Operand dst, Immediate src, int size) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // cmp can not be fused when comparing MEM-IMM, so we would not align this
    // instruction.
    emit_cmp(dst, src, size);
  }
 
  // Compare {al,ax,eax,rax} with src.  If equal, set ZF and write dst into
  // src. Otherwise clear ZF and write src into {al,ax,eax,rax}.  This
  // operation is only atomic if prefixed by the lock instruction.
  void emit_cmpxchg(Operand dst, Register src, int size);
 
  void emit_dec(Register dst, int size);
  void emit_dec(Operand dst, int size);
 
  // Divide rdx:rax by src.  Quotient in rax, remainder in rdx when size is 64.
  // Divide edx:eax by lower 32 bits of src.  Quotient in eax, remainder in edx
  // when size is 32.
  void emit_idiv(Register src, int size);
  void emit_div(Register src, int size);
 
  // Signed multiply instructions.
  // rdx:rax = rax * src when size is 64 or edx:eax = eax * src when size is 32.
  void emit_imul(Register src, int size);
  void emit_imul(Operand src, int size);
  void emit_imul(Register dst, Register src, int size);
  void emit_imul(Register dst, Operand src, int size);
  void emit_imul(Register dst, Register src, Immediate imm, int size);
  void emit_imul(Register dst, Operand src, Immediate imm, int size);
 
  void emit_inc(Register dst, int size);
  void emit_inc(Operand dst, int size);
 
  void emit_lea(Register dst, Operand src, int size);
 
  void emit_mov(Register dst, Operand src, int size);
  void emit_mov(Register dst, Register src, int size);
  void emit_mov(Operand dst, Register src, int size);
  void emit_mov(Register dst, Immediate value, int size);
  void emit_mov(Operand dst, Immediate value, int size);
  void emit_mov(Register dst, Immediate64 value, int size);
 
  void emit_movzxb(Register dst, Operand src, int size);
  void emit_movzxb(Register dst, Register src, int size);
  void emit_movzxw(Register dst, Operand src, int size);
  void emit_movzxw(Register dst, Register src, int size);
 
  void emit_neg(Register dst, int size);
  void emit_neg(Operand dst, int size);
 
  void emit_not(Register dst, int size);
  void emit_not(Operand dst, int size);
 
  void emit_or(Register dst, Register src, int size) {
    arithmetic_op(0x0B, dst, src, size);
  }
 
  void emit_or(Register dst, Operand src, int size) {
    arithmetic_op(0x0B, dst, src, size);
  }
 
  void emit_or(Operand dst, Register src, int size) {
    arithmetic_op(0x9, src, dst, size);
  }
 
  void emit_or(Register dst, Immediate src, int size) {
    immediate_arithmetic_op(0x1, dst, src, size);
  }
 
  void emit_or(Operand dst, Immediate src, int size) {
    immediate_arithmetic_op(0x1, dst, src, size);
  }
 
  void emit_repmovs(int size);
 
  void emit_sbb(Register dst, Register src, int size) {
    arithmetic_op(0x1b, dst, src, size);
  }
 
  void emit_sub(Register dst, Register src, int size) {
    arithmetic_op(0x2B, dst, src, size);
  }
 
  void emit_sub(Register dst, Immediate src, int size) {
    immediate_arithmetic_op(0x5, dst, src, size);
  }
 
  void emit_sub(Register dst, Operand src, int size) {
    arithmetic_op(0x2B, dst, src, size);
  }
 
  void emit_sub(Operand dst, Register src, int size) {
    arithmetic_op(0x29, src, dst, size);
  }
 
  void emit_sub(Operand dst, Immediate src, int size) {
    immediate_arithmetic_op(0x5, dst, src, size);
  }
 
  void emit_test(Register dst, Register src, int size);
  // Used for JCC erratum performance mitigation.
  void emit_aligned_test(Register dst, Register src, int size) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* test */ 3 + /* jcc */ 6
    const int kMaxMacroFusionLength = 9;
    AlignForJCCErratum(kMaxMacroFusionLength);
    emit_test(dst, src, size);
  }
 
  void emit_test(Register reg, Immediate mask, int size);
  // Used for JCC erratum performance mitigation.
  void emit_aligned_test(Register reg, Immediate mask, int size) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* testl */ 7 + /* jcc */ 6
    // /* testq */ 11 + /* jcc */ 6
    const int kMaxMacroFusionLength = 9 + size;
    AlignForJCCErratum(kMaxMacroFusionLength);
    emit_test(reg, mask, size);
  }
 
  void emit_test(Operand op, Register reg, int size);
  // Used for JCC erratum performance mitigation.
  void emit_aligned_test(Operand op, Register reg, int size) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // /* test */ 8 + /* jcc */ 6
    const int kMaxMacroFusionLength = 14;
    AlignForJCCErratum(kMaxMacroFusionLength);
    emit_test(op, reg, size);
  }
 
  void emit_test(Operand op, Immediate mask, int size);
  // Used for JCC erratum performance mitigation.
  void emit_aligned_test(Operand op, Immediate mask, int size) {
    DCHECK(CpuFeatures::IsSupported(INTEL_JCC_ERRATUM_MITIGATION));
    // test can not be fused when comparing MEM-IMM, so we would not align this
    // instruction.
    emit_test(op, mask, size);
  }
 
  void emit_test(Register reg, Operand op, int size) {
    return emit_test(op, reg, size);
  }
 
  // Used for JCC erratum performance mitigation.
  void emit_aligned_test(Register reg, Operand op, int size) {
    return emit_aligned_test(op, reg, size);
  }
 
  void emit_xchg(Register dst, Register src, int size);
  void emit_xchg(Register dst, Operand src, int size);
 
  void emit_xor(Register dst, Register src, int size) {
    if (size == kInt64Size && dst.code() == src.code()) {
      // 32 bit operations zero the top 32 bits of 64 bit registers. Therefore
      // there is no need to make this a 64 bit operation.
      arithmetic_op(0x33, dst, src, kInt32Size);
    } else {
      arithmetic_op(0x33, dst, src, size);
    }
  }
 
  void emit_xor(Register dst, Operand src, int size) {
    arithmetic_op(0x33, dst, src, size);
  }
 
  void emit_xor(Register dst, Immediate src, int size) {
    immediate_arithmetic_op(0x6, dst, src, size);
  }
 
  void emit_xor(Operand dst, Immediate src, int size) {
    immediate_arithmetic_op(0x6, dst, src, size);
  }
 
  void emit_xor(Operand dst, Register src, int size) {
    arithmetic_op(0x31, src, dst, size);
  }
 
  // Most BMI instructions are similar.
  void bmi1q(uint8_t op, Register reg, Register vreg, Register rm);
  void bmi1q(uint8_t op, Register reg, Register vreg, Operand rm);
  void bmi1l(uint8_t op, Register reg, Register vreg, Register rm);
  void bmi1l(uint8_t op, Register reg, Register vreg, Operand rm);
  void bmi2q(SIMDPrefix pp, uint8_t op, Register reg, Register vreg,
             Register rm);
  void bmi2q(SIMDPrefix pp, uint8_t op, Register reg, Register vreg,
             Operand rm);
  void bmi2l(SIMDPrefix pp, uint8_t op, Register reg, Register vreg,
             Register rm);
  void bmi2l(SIMDPrefix pp, uint8_t op, Register reg, Register vreg,
             Operand rm);
 
  // record the position of jmp/jcc instruction
  void record_farjmp_position(Label* L, int pos);
 
  bool is_optimizable_farjmp(int idx);
 
  void AllocateAndInstallRequestedHeapNumbers(LocalIsolate* isolate);
 
  int WriteCodeComments();
  int WriteBuiltinJumpTableInfos();
 
  void GetCode(LocalIsolate* isolate, CodeDesc* desc,
               int safepoint_table_offset, int handler_table_offset);
 
  friend class EnsureSpace;
  friend class RegExpMacroAssemblerX64;
 
  // code generation
  RelocInfoWriter reloc_info_writer;
 
  // Internal reference positions, required for (potential) patching in
  // GrowBuffer(); contains only those internal references whose labels
  // are already bound.
  std::deque<int> internal_reference_positions_;
 
  ConstPool constpool_;
 
  friend class ConstPool;
 
  BuiltinJumpTableInfoWriter builtin_jump_table_info_writer_;
 
#if defined(V8_OS_WIN_X64)
  std::unique_ptr<win64_unwindinfo::XdataEncoder> xdata_encoder_;
#endif
};
 
extern template EXPORT_TEMPLATE_DECLARE(
    V8_EXPORT_PRIVATE) void Assembler::fma_instr(uint8_t op, XMMRegister dst,
                                                 XMMRegister src1,
                                                 XMMRegister src2,
                                                 VectorLength l, SIMDPrefix pp,
                                                 LeadingOpcode m, VexW w);
 
extern template EXPORT_TEMPLATE_DECLARE(
    V8_EXPORT_PRIVATE) void Assembler::fma_instr(uint8_t op, YMMRegister dst,
                                                 YMMRegister src1,
                                                 YMMRegister src2,
                                                 VectorLength l, SIMDPrefix pp,
                                                 LeadingOpcode m, VexW w);
 
extern template EXPORT_TEMPLATE_DECLARE(
    V8_EXPORT_PRIVATE) void Assembler::fma_instr(uint8_t op, XMMRegister dst,
                                                 XMMRegister src1, Operand src2,
                                                 VectorLength l, SIMDPrefix pp,
                                                 LeadingOpcode m, VexW w);
 
extern template EXPORT_TEMPLATE_DECLARE(
    V8_EXPORT_PRIVATE) void Assembler::fma_instr(uint8_t op, YMMRegister dst,
                                                 YMMRegister src1, Operand src2,
                                                 VectorLength l, SIMDPrefix pp,
                                                 LeadingOpcode m, VexW w);
 
extern template EXPORT_TEMPLATE_DECLARE(
    V8_EXPORT_PRIVATE) void Assembler::vinstr(uint8_t op, YMMRegister dst,
                                              YMMRegister src1,
                                              YMMRegister src2, SIMDPrefix pp,
                                              LeadingOpcode m, VexW w,
                                              CpuFeature feature);
extern template EXPORT_TEMPLATE_DECLARE(
    V8_EXPORT_PRIVATE) void Assembler::vinstr(uint8_t op, YMMRegister dst,
                                              XMMRegister src1,
                                              XMMRegister src2, SIMDPrefix pp,
                                              LeadingOpcode m, VexW w,
                                              CpuFeature feature);
extern template EXPORT_TEMPLATE_DECLARE(
    V8_EXPORT_PRIVATE) void Assembler::vinstr(uint8_t op, YMMRegister dst,
                                              YMMRegister src1, Operand src2,
                                              SIMDPrefix pp, LeadingOpcode m,
                                              VexW w, CpuFeature feature);
extern template EXPORT_TEMPLATE_DECLARE(
    V8_EXPORT_PRIVATE) void Assembler::vinstr(uint8_t op, YMMRegister dst,
                                              YMMRegister src1,
                                              XMMRegister src2, SIMDPrefix pp,
                                              LeadingOpcode m, VexW w,
                                              CpuFeature feature);
extern template EXPORT_TEMPLATE_DECLARE(
    V8_EXPORT_PRIVATE) void Assembler::vinstr(uint8_t op, YMMRegister dst,
                                              XMMRegister src1, Operand src2,
                                              SIMDPrefix pp, LeadingOpcode m,
                                              VexW w, CpuFeature feature);
extern template EXPORT_TEMPLATE_DECLARE(
    V8_EXPORT_PRIVATE) void Assembler::vinstr(uint8_t op, YMMRegister dst,
                                              XMMRegister src1,
                                              YMMRegister src2, SIMDPrefix pp,
                                              LeadingOpcode m, VexW w,
                                              CpuFeature feature);
 
// Helper class that ensures that there is enough space for generating
// instructions and relocation information.  The constructor makes
// sure that there is enough space and (in debug mode) the destructor
// checks that we did not generate too much.
class EnsureSpace {
 public:
  explicit V8_INLINE EnsureSpace(Assembler* assembler) : assembler_(assembler) {
    if (V8_UNLIKELY(assembler_->buffer_overflow())) assembler_->GrowBuffer();
#ifdef DEBUG
    space_before_ = assembler_->available_space();
#endif
  }
 
#ifdef DEBUG
  ~EnsureSpace() {
    int bytes_generated = space_before_ - assembler_->available_space();
    DCHECK(bytes_generated < assembler_->kGap);
  }
#endif
 
 private:
  Assembler* const assembler_;
#ifdef DEBUG
  int space_before_;
#endif
};
 
}  // namespace internal
}  // namespace v8
 
#endif  // V8_CODEGEN_X64_ASSEMBLER_X64_H_