Python3字节码指令集

查询字节码指令集：

>>> import opcode
>>> for op in range(len(opcode.opname)):
    print('ox%.2X(%.3d): %s' % (op, op, opcode.opname[op]))

    
ox00(000): <0>
ox01(001): POP_TOP
ox02(002): ROT_TWO
ox03(003): ROT_THREE
ox04(004): DUP_TOP
ox05(005): DUP_TOP_TWO
ox06(006): <6>
ox07(007): <7>
ox08(008): <8>
ox09(009): NOP
ox0A(010): UNARY_POSITIVE
ox0B(011): UNARY_NEGATIVE
ox0C(012): UNARY_NOT
ox0D(013): <13>
ox0E(014): <14>
ox0F(015): UNARY_INVERT
ox10(016): <16>
ox11(017): <17>
ox12(018): <18>
ox13(019): BINARY_POWER
ox14(020): BINARY_MULTIPLY
ox15(021): <21>
ox16(022): BINARY_MODULO
ox17(023): BINARY_ADD
ox18(024): BINARY_SUBTRACT
ox19(025): BINARY_SUBSCR
ox1A(026): BINARY_FLOOR_DIVIDE
ox1B(027): BINARY_TRUE_DIVIDE
ox1C(028): INPLACE_FLOOR_DIVIDE
ox1D(029): INPLACE_TRUE_DIVIDE
ox1E(030): <30>
ox1F(031): <31>
ox20(032): <32>
ox21(033): <33>
ox22(034): <34>
ox23(035): <35>
ox24(036): <36>
ox25(037): <37>
ox26(038): <38>
ox27(039): <39>
ox28(040): <40>
ox29(041): <41>
ox2A(042): <42>
ox2B(043): <43>
ox2C(044): <44>
ox2D(045): <45>
ox2E(046): <46>
ox2F(047): <47>
ox30(048): <48>
ox31(049): <49>
ox32(050): <50>
ox33(051): <51>
ox34(052): <52>
ox35(053): <53>
ox36(054): STORE_MAP
ox37(055): INPLACE_ADD
ox38(056): INPLACE_SUBTRACT
ox39(057): INPLACE_MULTIPLY
ox3A(058): <58>
ox3B(059): INPLACE_MODULO
ox3C(060): STORE_SUBSCR
ox3D(061): DELETE_SUBSCR
ox3E(062): BINARY_LSHIFT
ox3F(063): BINARY_RSHIFT
ox40(064): BINARY_AND
ox41(065): BINARY_XOR
ox42(066): BINARY_OR
ox43(067): INPLACE_POWER
ox44(068): GET_ITER
ox45(069): STORE_LOCALS
ox46(070): PRINT_EXPR
ox47(071): LOAD_BUILD_CLASS
ox48(072): YIELD_FROM
ox49(073): <73>
ox4A(074): <74>
ox4B(075): INPLACE_LSHIFT
ox4C(076): INPLACE_RSHIFT
ox4D(077): INPLACE_AND
ox4E(078): INPLACE_XOR
ox4F(079): INPLACE_OR
ox50(080): BREAK_LOOP
ox51(081): WITH_CLEANUP
ox52(082): <82>
ox53(083): RETURN_VALUE
ox54(084): IMPORT_STAR
ox55(085): <85>
ox56(086): YIELD_VALUE
ox57(087): POP_BLOCK
ox58(088): END_FINALLY
ox59(089): POP_EXCEPT
ox5A(090): STORE_NAME
ox5B(091): DELETE_NAME
ox5C(092): UNPACK_SEQUENCE
ox5D(093): FOR_ITER
ox5E(094): UNPACK_EX
ox5F(095): STORE_ATTR
ox60(096): DELETE_ATTR
ox61(097): STORE_GLOBAL
ox62(098): DELETE_GLOBAL
ox63(099): <99>
ox64(100): LOAD_CONST
ox65(101): LOAD_NAME
ox66(102): BUILD_TUPLE
ox67(103): BUILD_LIST
ox68(104): BUILD_SET
ox69(105): BUILD_MAP
ox6A(106): LOAD_ATTR
ox6B(107): COMPARE_OP
ox6C(108): IMPORT_NAME
ox6D(109): IMPORT_FROM
ox6E(110): JUMP_FORWARD
ox6F(111): JUMP_IF_FALSE_OR_POP
ox70(112): JUMP_IF_TRUE_OR_POP
ox71(113): JUMP_ABSOLUTE
ox72(114): POP_JUMP_IF_FALSE
ox73(115): POP_JUMP_IF_TRUE
ox74(116): LOAD_GLOBAL
ox75(117): <117>
ox76(118): <118>
ox77(119): CONTINUE_LOOP
ox78(120): SETUP_LOOP
ox79(121): SETUP_EXCEPT
ox7A(122): SETUP_FINALLY
ox7B(123): <123>
ox7C(124): LOAD_FAST
ox7D(125): STORE_FAST
ox7E(126): DELETE_FAST
ox7F(127): <127>
ox80(128): <128>
ox81(129): <129>
ox82(130): RAISE_VARARGS
ox83(131): CALL_FUNCTION
ox84(132): MAKE_FUNCTION
ox85(133): BUILD_SLICE
ox86(134): MAKE_CLOSURE
ox87(135): LOAD_CLOSURE
ox88(136): LOAD_DEREF
ox89(137): STORE_DEREF
ox8A(138): DELETE_DEREF
ox8B(139): <139>
ox8C(140): CALL_FUNCTION_VAR
ox8D(141): CALL_FUNCTION_KW
ox8E(142): CALL_FUNCTION_VAR_KW
ox8F(143): SETUP_WITH
ox90(144): EXTENDED_ARG
ox91(145): LIST_APPEND
ox92(146): SET_ADD
ox93(147): MAP_ADD
ox94(148): <148>
ox95(149): <149>
ox96(150): <150>
ox97(151): <151>
ox98(152): <152>
ox99(153): <153>
ox9A(154): <154>
ox9B(155): <155>
ox9C(156): <156>
ox9D(157): <157>
ox9E(158): <158>
ox9F(159): <159>
oxA0(160): <160>
oxA1(161): <161>
oxA2(162): <162>
oxA3(163): <163>
oxA4(164): <164>
oxA5(165): <165>
oxA6(166): <166>
oxA7(167): <167>
oxA8(168): <168>
oxA9(169): <169>
oxAA(170): <170>
oxAB(171): <171>
oxAC(172): <172>
oxAD(173): <173>
oxAE(174): <174>
oxAF(175): <175>
oxB0(176): <176>
oxB1(177): <177>
oxB2(178): <178>
oxB3(179): <179>
oxB4(180): <180>
oxB5(181): <181>
oxB6(182): <182>
oxB7(183): <183>
oxB8(184): <184>
oxB9(185): <185>
oxBA(186): <186>
oxBB(187): <187>
oxBC(188): <188>
oxBD(189): <189>
oxBE(190): <190>
oxBF(191): <191>
oxC0(192): <192>
oxC1(193): <193>
oxC2(194): <194>
oxC3(195): <195>
oxC4(196): <196>
oxC5(197): <197>
oxC6(198): <198>
oxC7(199): <199>
oxC8(200): <200>
oxC9(201): <201>
oxCA(202): <202>
oxCB(203): <203>
oxCC(204): <204>
oxCD(205): <205>
oxCE(206): <206>
oxCF(207): <207>
oxD0(208): <208>
oxD1(209): <209>
oxD2(210): <210>
oxD3(211): <211>
oxD4(212): <212>
oxD5(213): <213>
oxD6(214): <214>
oxD7(215): <215>
oxD8(216): <216>
oxD9(217): <217>
oxDA(218): <218>
oxDB(219): <219>
oxDC(220): <220>
oxDD(221): <221>
oxDE(222): <222>
oxDF(223): <223>
oxE0(224): <224>
oxE1(225): <225>
oxE2(226): <226>
oxE3(227): <227>
oxE4(228): <228>
oxE5(229): <229>
oxE6(230): <230>
oxE7(231): <231>
oxE8(232): <232>
oxE9(233): <233>
oxEA(234): <234>
oxEB(235): <235>
oxEC(236): <236>
oxED(237): <237>
oxEE(238): <238>
oxEF(239): <239>
oxF0(240): <240>
oxF1(241): <241>
oxF2(242): <242>
oxF3(243): <243>
oxF4(244): <244>
oxF5(245): <245>
oxF6(246): <246>
oxF7(247): <247>
oxF8(248): <248>
oxF9(249): <249>
oxFA(250): <250>
oxFB(251): <251>
oxFC(252): <252>
oxFD(253): <253>
oxFE(254): <254>
oxFF(255): <255>

“<n>”形式：是未使用的操作码。 

“字节码”，是指令以字节为单位，最多只能表示256个不同的字节码指令。实际上Python只用了101条字节码指令：

>>> import opcode
>>> import re
>>> [op for op in opcode.opname if re.search(r'^<\d*>$',op)]
['<0>', '<6>', '<7>', '<8>', '<13>', '<14>', '<16>', '<17>', '<18>', '<21>', '<30>', '<31>', '<32>', '<33>', '<34>', '<35>', '<36>', '<37>', '<38>', '<39>', '<40>', '<41>', '<42>', '<43>', '<44>', '<45>', '<46>', '<47>', '<48>', '<49>', '<50>', '<51>', '<52>', '<53>', '<58>', '<73>', '<74>', '<82>', '<85>', '<99>', '<117>', '<118>', '<123>', '<127>', '<128>', '<129>', '<139>', '<148>', '<149>', '<150>', '<151>', '<152>', '<153>', '<154>', '<155>', '<156>', '<157>', '<158>', '<159>', '<160>', '<161>', '<162>', '<163>', '<164>', '<165>', '<166>', '<167>', '<168>', '<169>', '<170>', '<171>', '<172>', '<173>', '<174>', '<175>', '<176>', '<177>', '<178>', '<179>', '<180>', '<181>', '<182>', '<183>', '<184>', '<185>', '<186>', '<187>', '<188>', '<189>', '<190>', '<191>', '<192>', '<193>', '<194>', '<195>', '<196>', '<197>', '<198>', '<199>', '<200>', '<201>', '<202>', '<203>', '<204>', '<205>', '<206>', '<207>', '<208>', '<209>', '<210>', '<211>', '<212>', '<213>', '<214>', '<215>', '<216>', '<217>', '<218>', '<219>', '<220>', '<221>', '<222>', '<223>', '<224>', '<225>', '<226>', '<227>', '<228>', '<229>', '<230>', '<231>', '<232>', '<233>', '<234>', '<235>', '<236>', '<237>', '<238>', '<239>', '<240>', '<241>', '<242>', '<243>', '<244>', '<245>', '<246>', '<247>', '<248>', '<249>', '<250>', '<251>', '<252>', '<253>', '<254>', '<255>']
>>> [op for op in opcode.opname if not re.search(r'^<\d*>$',op)]
['POP_TOP', 'ROT_TWO', 'ROT_THREE', 'DUP_TOP', 'DUP_TOP_TWO', 'NOP', 'UNARY_POSITIVE', 'UNARY_NEGATIVE', 'UNARY_NOT', 'UNARY_INVERT', 'BINARY_POWER', 'BINARY_MULTIPLY', 'BINARY_MODULO', 'BINARY_ADD', 'BINARY_SUBTRACT', 'BINARY_SUBSCR', 'BINARY_FLOOR_DIVIDE', 'BINARY_TRUE_DIVIDE', 'INPLACE_FLOOR_DIVIDE', 'INPLACE_TRUE_DIVIDE', 'STORE_MAP', 'INPLACE_ADD', 'INPLACE_SUBTRACT', 'INPLACE_MULTIPLY', 'INPLACE_MODULO', 'STORE_SUBSCR', 'DELETE_SUBSCR', 'BINARY_LSHIFT', 'BINARY_RSHIFT', 'BINARY_AND', 'BINARY_XOR', 'BINARY_OR', 'INPLACE_POWER', 'GET_ITER', 'STORE_LOCALS', 'PRINT_EXPR', 'LOAD_BUILD_CLASS', 'YIELD_FROM', 'INPLACE_LSHIFT', 'INPLACE_RSHIFT', 'INPLACE_AND', 'INPLACE_XOR', 'INPLACE_OR', 'BREAK_LOOP', 'WITH_CLEANUP', 'RETURN_VALUE', 'IMPORT_STAR', 'YIELD_VALUE', 'POP_BLOCK', 'END_FINALLY', 'POP_EXCEPT', 'STORE_NAME', 'DELETE_NAME', 'UNPACK_SEQUENCE', 'FOR_ITER', 'UNPACK_EX', 'STORE_ATTR', 'DELETE_ATTR', 'STORE_GLOBAL', 'DELETE_GLOBAL', 'LOAD_CONST', 'LOAD_NAME', 'BUILD_TUPLE', 'BUILD_LIST', 'BUILD_SET', 'BUILD_MAP', 'LOAD_ATTR', 'COMPARE_OP', 'IMPORT_NAME', 'IMPORT_FROM', 'JUMP_FORWARD', 'JUMP_IF_FALSE_OR_POP', 'JUMP_IF_TRUE_OR_POP', 'JUMP_ABSOLUTE', 'POP_JUMP_IF_FALSE', 'POP_JUMP_IF_TRUE', 'LOAD_GLOBAL', 'CONTINUE_LOOP', 'SETUP_LOOP', 'SETUP_EXCEPT', 'SETUP_FINALLY', 'LOAD_FAST', 'STORE_FAST', 'DELETE_FAST', 'RAISE_VARARGS', 'CALL_FUNCTION', 'MAKE_FUNCTION', 'BUILD_SLICE', 'MAKE_CLOSURE', 'LOAD_CLOSURE', 'LOAD_DEREF', 'STORE_DEREF', 'DELETE_DEREF', 'CALL_FUNCTION_VAR', 'CALL_FUNCTION_KW', 'CALL_FUNCTION_VAR_KW', 'SETUP_WITH', 'EXTENDED_ARG', 'LIST_APPEND', 'SET_ADD', 'MAP_ADD']
>>> len([op for op in opcode.opname if not re.search(r'^<\d*>$',op)])
101
>>> len(opcode.opmap)
101

字节码指令的编码小于90的为无参数的，指令仅包含操作码自身，共1字节；大于等于90的，则每条指令在码之后还带有1个参数，参数长度为2字节，共3字节。 指令后的参数大于2个字节的情况：见后面的指令含义介绍部分的EXTENDED_ARG(ext)。

Python程序的字节码在运行时以PyStringObject的形式保存在PyCodeObject的co_code域里（前面的PyCodeObject一节已介绍过）。co_code域只含有指令而不包含别的程序数据；变量名、常量等数据均放在别的域里。

 获取编码为88的指令码和获取指令码为'END_FINALLY'的编码：

>>> import opcode
>>> opcode.opname[88]
'END_FINALLY'
>>> opcode.opmap['END_FINALLY']
88

 

opcode模块所在的文件：

>>> opcode.__file__
'D:\\Python33\\lib\\opcode.py'

"""
opcode module - potentially shared between dis and other modules which
operate on bytecodes (e.g. peephole optimizers).
"""

__all__ = ["cmp_op", "hasconst", "hasname", "hasjrel", "hasjabs",
           "haslocal", "hascompare", "hasfree", "opname", "opmap",
           "HAVE_ARGUMENT", "EXTENDED_ARG", "hasnargs"]

cmp_op = ('<', '<=', '==', '!=', '>', '>=', 'in', 'not in', 'is',
        'is not', 'exception match', 'BAD')

hasconst = []
hasname = []
hasjrel = []
hasjabs = []
haslocal = []
hascompare = []
hasfree = []
hasnargs = []

opmap = {}
opname = [''] * 256
for op in range(256): opname[op] = '<%r>' % (op,) #initial opname ,['<0>', '<1>', '<2>', ..., '<255>']
del op

def def_op(name, op):
    opname[op] = name
    opmap[name] = op

def name_op(name, op):
    def_op(name, op)
    hasname.append(op)

def jrel_op(name, op):
    def_op(name, op)
    hasjrel.append(op)

def jabs_op(name, op):
    def_op(name, op)
    hasjabs.append(op)

# Instruction opcodes for compiled code
# Blank lines correspond to available opcodes

def_op('POP_TOP', 1)
def_op('ROT_TWO', 2)
def_op('ROT_THREE', 3)
def_op('DUP_TOP', 4)
def_op('DUP_TOP_TWO', 5)

def_op('NOP', 9)
def_op('UNARY_POSITIVE', 10)
def_op('UNARY_NEGATIVE', 11)
def_op('UNARY_NOT', 12)

def_op('UNARY_INVERT', 15)

def_op('BINARY_POWER', 19)
def_op('BINARY_MULTIPLY', 20)

def_op('BINARY_MODULO', 22)
def_op('BINARY_ADD', 23)
def_op('BINARY_SUBTRACT', 24)
def_op('BINARY_SUBSCR', 25)
def_op('BINARY_FLOOR_DIVIDE', 26)
def_op('BINARY_TRUE_DIVIDE', 27)
def_op('INPLACE_FLOOR_DIVIDE', 28)
def_op('INPLACE_TRUE_DIVIDE', 29)

def_op('STORE_MAP', 54)
def_op('INPLACE_ADD', 55)
def_op('INPLACE_SUBTRACT', 56)
def_op('INPLACE_MULTIPLY', 57)

def_op('INPLACE_MODULO', 59)
def_op('STORE_SUBSCR', 60)
def_op('DELETE_SUBSCR', 61)
def_op('BINARY_LSHIFT', 62)
def_op('BINARY_RSHIFT', 63)
def_op('BINARY_AND', 64)
def_op('BINARY_XOR', 65)
def_op('BINARY_OR', 66)
def_op('INPLACE_POWER', 67)
def_op('GET_ITER', 68)
def_op('STORE_LOCALS', 69)

def_op('PRINT_EXPR', 70)
def_op('LOAD_BUILD_CLASS', 71)
def_op('YIELD_FROM', 72)

def_op('INPLACE_LSHIFT', 75)
def_op('INPLACE_RSHIFT', 76)
def_op('INPLACE_AND', 77)
def_op('INPLACE_XOR', 78)
def_op('INPLACE_OR', 79)
def_op('BREAK_LOOP', 80)
def_op('WITH_CLEANUP', 81)

def_op('RETURN_VALUE', 83)
def_op('IMPORT_STAR', 84)

def_op('YIELD_VALUE', 86)
def_op('POP_BLOCK', 87)
def_op('END_FINALLY', 88)
def_op('POP_EXCEPT', 89)

HAVE_ARGUMENT = 90              # Opcodes from here have an argument:

name_op('STORE_NAME', 90)       # Index in name list
name_op('DELETE_NAME', 91)      # ""
def_op('UNPACK_SEQUENCE', 92)   # Number of tuple items
jrel_op('FOR_ITER', 93)
def_op('UNPACK_EX', 94)
name_op('STORE_ATTR', 95)       # Index in name list
name_op('DELETE_ATTR', 96)      # ""
name_op('STORE_GLOBAL', 97)     # ""
name_op('DELETE_GLOBAL', 98)    # ""
def_op('LOAD_CONST', 100)       # Index in const list
hasconst.append(100)
name_op('LOAD_NAME', 101)       # Index in name list
def_op('BUILD_TUPLE', 102)      # Number of tuple items
def_op('BUILD_LIST', 103)       # Number of list items
def_op('BUILD_SET', 104)        # Number of set items
def_op('BUILD_MAP', 105)        # Number of dict entries (upto 255)
name_op('LOAD_ATTR', 106)       # Index in name list
def_op('COMPARE_OP', 107)       # Comparison operator
hascompare.append(107)
name_op('IMPORT_NAME', 108)     # Index in name list
name_op('IMPORT_FROM', 109)     # Index in name list

jrel_op('JUMP_FORWARD', 110)    # Number of bytes to skip
jabs_op('JUMP_IF_FALSE_OR_POP', 111) # Target byte offset from beginning of code
jabs_op('JUMP_IF_TRUE_OR_POP', 112)  # ""
jabs_op('JUMP_ABSOLUTE', 113)        # ""
jabs_op('POP_JUMP_IF_FALSE', 114)    # ""
jabs_op('POP_JUMP_IF_TRUE', 115)     # ""

name_op('LOAD_GLOBAL', 116)     # Index in name list

jabs_op('CONTINUE_LOOP', 119)   # Target address
jrel_op('SETUP_LOOP', 120)      # Distance to target address
jrel_op('SETUP_EXCEPT', 121)    # ""
jrel_op('SETUP_FINALLY', 122)   # ""

def_op('LOAD_FAST', 124)        # Local variable number
haslocal.append(124)
def_op('STORE_FAST', 125)       # Local variable number
haslocal.append(125)
def_op('DELETE_FAST', 126)      # Local variable number
haslocal.append(126)

def_op('RAISE_VARARGS', 130)    # Number of raise arguments (1, 2, or 3)
def_op('CALL_FUNCTION', 131)    # #args + (#kwargs << 8)
hasnargs.append(131)
def_op('MAKE_FUNCTION', 132)    # Number of args with default values
def_op('BUILD_SLICE', 133)      # Number of items
def_op('MAKE_CLOSURE', 134)
def_op('LOAD_CLOSURE', 135)
hasfree.append(135)
def_op('LOAD_DEREF', 136)
hasfree.append(136)
def_op('STORE_DEREF', 137)
hasfree.append(137)
def_op('DELETE_DEREF', 138)
hasfree.append(138)

def_op('CALL_FUNCTION_VAR', 140)     # #args + (#kwargs << 8)
hasnargs.append(140)
def_op('CALL_FUNCTION_KW', 141)      # #args + (#kwargs << 8)
hasnargs.append(141)
def_op('CALL_FUNCTION_VAR_KW', 142)  # #args + (#kwargs << 8)
hasnargs.append(142)

jrel_op('SETUP_WITH', 143)

def_op('LIST_APPEND', 145)
def_op('SET_ADD', 146)
def_op('MAP_ADD', 147)

def_op('EXTENDED_ARG', 144)
EXTENDED_ARG = 144

del def_op, name_op, jrel_op, jabs_op

 

 

各字节码指令的含义：http://docs.python.org/3/library/dis.html

31.12.1. Python Bytecode Instructions

The Python compiler currently generates the following bytecode instructions.

General instructions

NOP

Do nothing code. Used as a placeholder by the bytecode optimizer.

POP_TOP

Removes the top-of-stack (TOS) item.

ROT_TWO

Swaps the two top-most stack items.

ROT_THREE

Lifts second and third stack item one position up, moves top down to position three.

DUP_TOP

Duplicates the reference on top of the stack.

DUP_TOP_TWO

Duplicates the two references on top of the stack, leaving them in the same order.

Unary operations

Unary operations take the top of the stack, apply the operation, and push the result back on the stack.

UNARY_POSITIVE

Implements TOS = +TOS.

UNARY_NEGATIVE

Implements TOS = -TOS.

UNARY_NOT

Implements TOS = not TOS.

UNARY_INVERT

Implements TOS = ~TOS.

GET_ITER

Implements TOS = iter(TOS).

Binary operations

Binary operations remove the top of the stack (TOS) and the second top-most stack item (TOS1) from the stack. They perform the operation, and put the result back on the stack.

BINARY_POWER

Implements TOS = TOS1 ** TOS.

BINARY_MULTIPLY

Implements TOS = TOS1 * TOS.

BINARY_FLOOR_DIVIDE

Implements TOS = TOS1 // TOS.

BINARY_TRUE_DIVIDE

Implements TOS = TOS1 / TOS.

BINARY_MODULO

Implements TOS = TOS1 % TOS.

BINARY_ADD

Implements TOS = TOS1 + TOS.

BINARY_SUBTRACT

Implements TOS = TOS1 - TOS.

BINARY_SUBSCR

Implements TOS = TOS1[TOS].

BINARY_LSHIFT

Implements TOS = TOS1 << TOS.

BINARY_RSHIFT

Implements TOS = TOS1 >> TOS.

BINARY_AND

Implements TOS = TOS1 & TOS.

BINARY_XOR

Implements TOS = TOS1 ^ TOS.

BINARY_OR

Implements TOS = TOS1 | TOS.

In-place operations

In-place operations are like binary operations, in that they remove TOS and TOS1, and push the result back on the stack, but the operation is done in-place when TOS1 supports it, and the resulting TOS may be (but does not have to be) the original TOS1.

INPLACE_POWER

Implements in-place TOS = TOS1 ** TOS.

INPLACE_MULTIPLY

Implements in-place TOS = TOS1 * TOS.

INPLACE_FLOOR_DIVIDE

Implements in-place TOS = TOS1 // TOS.

INPLACE_TRUE_DIVIDE

Implements in-place TOS = TOS1 / TOS.

INPLACE_MODULO

Implements in-place TOS = TOS1 % TOS.

INPLACE_ADD

Implements in-place TOS = TOS1 + TOS.

INPLACE_SUBTRACT

Implements in-place TOS = TOS1 - TOS.

INPLACE_LSHIFT

Implements in-place TOS = TOS1 << TOS.

INPLACE_RSHIFT

Implements in-place TOS = TOS1 >> TOS.

INPLACE_AND

Implements in-place TOS = TOS1 & TOS.

INPLACE_XOR

Implements in-place TOS = TOS1 ^ TOS.

INPLACE_OR

Implements in-place TOS = TOS1 | TOS.

STORE_SUBSCR

Implements TOS1[TOS] = TOS2.

DELETE_SUBSCR

Implements del TOS1[TOS].

Miscellaneous opcodes

PRINT_EXPR

Implements the expression statement for the interactive mode. TOS is removed from the stack and printed. In non-interactive mode, an expression statement is terminated with POP_STACK.

BREAK_LOOP

Terminates a loop due to a break statement.

CONTINUE_LOOP(target)

Continues a loop due to a continue statement. target is the address to jump to (which should be a FOR_ITER instruction).

SET_ADD(i)

Calls set.add(TOS1[-i], TOS). Used to implement set comprehensions.

LIST_APPEND(i)

Calls list.append(TOS[-i], TOS). Used to implement list comprehensions.

MAP_ADD(i)

Calls dict.setitem(TOS1[-i], TOS, TOS1). Used to implement dict comprehensions.

For all of the SET_ADD, LIST_APPEND and MAP_ADD instructions, while the added value or key/value pair is popped off, the container object remains on the stack so that it is available for further iterations of the loop.

RETURN_VALUE

Returns with TOS to the caller of the function.

YIELD_VALUE

Pops TOS and yields it from a generator.

YIELD_FROM

Pops TOS and delegates to it as a subiterator from a generator.

New in version 3.3.

IMPORT_STAR

Loads all symbols not starting with '_' directly from the module TOS to the local namespace. The module is popped after loading all names. This opcode implements from module import *.

POP_BLOCK

Removes one block from the block stack. Per frame, there is a stack of blocks, denoting nested loops, try statements, and such.

POP_EXCEPT

Removes one block from the block stack. The popped block must be an exception handler block, as implicitly created when entering an except handler. In addition to popping extraneous values from the frame stack, the last three popped values are used to restore the exception state.

END_FINALLY

Terminates a finally clause. The interpreter recalls whether the exception has to be re-raised, or whether the function returns, and continues with the outer-next block.

LOAD_BUILD_CLASS

Pushes builtins.__build_class__() onto the stack. It is later called by CALL_FUNCTION to construct a class.

SETUP_WITH(delta)

This opcode performs several operations before a with block starts. First, it loads __exit__() from the context manager and pushes it onto the stack for later use by WITH_CLEANUP. Then, __enter__() is called, and a finally block pointing to delta is pushed. Finally, the result of calling the enter method is pushed onto the stack. The next opcode will either ignore it (POP_TOP), or store it in (a) variable(s) (STORE_FAST, STORE_NAME, or UNPACK_SEQUENCE).

WITH_CLEANUP

Cleans up the stack when a with statement block exits. TOS is the context manager’s __exit__() bound method. Below TOS are 1–3 values indicating how/why the finally clause was entered:

• SECOND = None
• (SECOND, THIRD) = (WHY_{RETURN,CONTINUE}), retval
• SECOND = WHY_*; no retval below it
• (SECOND, THIRD, FOURTH) = exc_info()

In the last case, TOS(SECOND, THIRD, FOURTH) is called, otherwise TOS(None, None, None). In addition, TOS is removed from the stack.

If the stack represents an exception, and the function call returns a ‘true’ value, this information is “zapped” and replaced with a single WHY_SILENCED to prevent END_FINALLY from re-raising the exception. (But non-local gotos will still be resumed.)

STORE_LOCALS

Pops TOS from the stack and stores it as the current frame’s f_locals. This is used in class construction.

All of the following opcodes expect arguments. An argument is two bytes, with the more significant byte last.

STORE_NAME(namei)

Implements name = TOS. namei is the index of name in the attribute co_names of the code object. The compiler tries to use STORE_FAST orSTORE_GLOBAL if possible.

DELETE_NAME(namei)

Implements del name, where namei is the index into co_names attribute of the code object.

UNPACK_SEQUENCE(count)

Unpacks TOS into count individual values, which are put onto the stack right-to-left.

UNPACK_EX(counts)

Implements assignment with a starred target: Unpacks an iterable in TOS into individual values, where the total number of values can be smaller than the number of items in the iterable: one the new values will be a list of all leftover items.

The low byte of counts is the number of values before the list value, the high byte of counts the number of values after it. The resulting values are put onto the stack right-to-left.

STORE_ATTR(namei)

Implements TOS.name = TOS1, where namei is the index of name in co_names.

DELETE_ATTR(namei)

Implements del TOS.name, using namei as index into co_names.

STORE_GLOBAL(namei)

Works as STORE_NAME, but stores the name as a global.

DELETE_GLOBAL(namei)

Works as DELETE_NAME, but deletes a global name.

LOAD_CONST(consti)

Pushes co_consts[consti] onto the stack.

LOAD_NAME(namei)

Pushes the value associated with co_names[namei] onto the stack.

BUILD_TUPLE(count)

Creates a tuple consuming count items from the stack, and pushes the resulting tuple onto the stack.

BUILD_LIST(count)

Works as BUILD_TUPLE, but creates a list.

BUILD_SET(count)

Works as BUILD_TUPLE, but creates a set.

BUILD_MAP(count)

Pushes a new dictionary object onto the stack. The dictionary is pre-sized to hold count entries.

LOAD_ATTR(namei)

Replaces TOS with getattr(TOS, co_names[namei]).

COMPARE_OP(opname)

Performs a Boolean operation. The operation name can be found in cmp_op[opname].

IMPORT_NAME(namei)

Imports the module co_names[namei]. TOS and TOS1 are popped and provide the fromlist and level arguments of __import__(). The module object is pushed onto the stack. The current namespace is not affected: for a proper import statement, a subsequent STORE_FAST instruction modifies the namespace.

IMPORT_FROM(namei)

Loads the attribute co_names[namei] from the module found in TOS. The resulting object is pushed onto the stack, to be subsequently stored by aSTORE_FAST instruction.

JUMP_FORWARD(delta)

Increments bytecode counter by delta.

POP_JUMP_IF_TRUE(target)

If TOS is true, sets the bytecode counter to target. TOS is popped.

POP_JUMP_IF_FALSE(target)

If TOS is false, sets the bytecode counter to target. TOS is popped.

JUMP_IF_TRUE_OR_POP(target)

If TOS is true, sets the bytecode counter to target and leaves TOS on the stack. Otherwise (TOS is false), TOS is popped.

JUMP_IF_FALSE_OR_POP(target)

If TOS is false, sets the bytecode counter to target and leaves TOS on the stack. Otherwise (TOS is true), TOS is popped.

JUMP_ABSOLUTE(target)

Set bytecode counter to target.

FOR_ITER(delta)

TOS is an iterator. Call its __next__() method. If this yields a new value, push it on the stack (leaving the iterator below it). If the iterator indicates it is exhausted TOS is popped, and the byte code counter is incremented by delta.

LOAD_GLOBAL(namei)

Loads the global named co_names[namei] onto the stack.

SETUP_LOOP(delta)

Pushes a block for a loop onto the block stack. The block spans from the current instruction with a size of delta bytes.

SETUP_EXCEPT(delta)

Pushes a try block from a try-except clause onto the block stack. delta points to the first except block.

SETUP_FINALLY(delta)

Pushes a try block from a try-except clause onto the block stack. delta points to the finally block.

STORE_MAP

Store a key and value pair in a dictionary. Pops the key and value while leaving the dictionary on the stack.

LOAD_FAST(var_num)

Pushes a reference to the local co_varnames[var_num] onto the stack.

STORE_FAST(var_num)

Stores TOS into the local co_varnames[var_num].

DELETE_FAST(var_num)

Deletes local co_varnames[var_num].

LOAD_CLOSURE(i)

Pushes a reference to the cell contained in slot i of the cell and free variable storage. The name of the variable is co_cellvars[i] if i is less than the length of co_cellvars. Otherwise it is co_freevars[i - len(co_cellvars)].

LOAD_DEREF(i)

Loads the cell contained in slot i of the cell and free variable storage. Pushes a reference to the object the cell contains on the stack.

STORE_DEREF(i)

Stores TOS into the cell contained in slot i of the cell and free variable storage.

DELETE_DEREF(i)

Empties the cell contained in slot i of the cell and free variable storage. Used by the del statement.

RAISE_VARARGS(argc)

Raises an exception. argc indicates the number of parameters to the raise statement, ranging from 0 to 3. The handler will find the traceback as TOS2, the parameter as TOS1, and the exception as TOS.

CALL_FUNCTION(argc)

Calls a function. The low byte of argc indicates the number of positional parameters, the high byte the number of keyword parameters. On the stack, the opcode finds the keyword parameters first. For each keyword argument, the value is on top of the key. Below the keyword parameters, the positional parameters are on the stack, with the right-most parameter on top. Below the parameters, the function object to call is on the stack. Pops all function arguments, and the function itself off the stack, and pushes the return value.

MAKE_FUNCTION(argc)

Pushes a new function object on the stack. TOS is the qualified name of the function; TOS1 is the code associated with the function. The function object is defined to have argc default parameters, which are found below TOS1.

MAKE_CLOSURE(argc)

Creates a new function object, sets its __closure__ slot, and pushes it on the stack. TOS is the qualified name of the function, TOS1 is the code associated with the function, and TOS2 is the tuple containing cells for the closure’s free variables. The function also has argc default parameters, which are found below the cells.

BUILD_SLICE(argc)

Pushes a slice object on the stack. argc must be 2 or 3. If it is 2, slice(TOS1, TOS) is pushed; if it is 3, slice(TOS2, TOS1, TOS) is pushed. See theslice() built-in function for more information.

EXTENDED_ARG(ext)

Prefixes any opcode which has an argument too big to fit into the default two bytes. ext holds two additional bytes which, taken together with the subsequent opcode’s argument, comprise a four-byte argument, ext being the two most-significant bytes.

CALL_FUNCTION_VAR(argc)

Calls a function. argc is interpreted as in CALL_FUNCTION. The top element on the stack contains the variable argument list, followed by keyword and positional arguments.

CALL_FUNCTION_KW(argc)

Calls a function. argc is interpreted as in CALL_FUNCTION. The top element on the stack contains the keyword arguments dictionary, followed by explicit keyword and positional arguments.

CALL_FUNCTION_VAR_KW(argc)

Calls a function. argc is interpreted as in CALL_FUNCTION. The top element on the stack contains the keyword arguments dictionary, followed by the variable-arguments tuple, followed by explicit keyword and positional arguments.

HAVE_ARGUMENT

This is not really an opcode. It identifies the dividing line between opcodes which don’t take arguments < HAVE_ARGUMENT and those which do >=HAVE_ARGUMENT.

 

简单分析一个函数的字节码：

>>> import dis
>>> def f(x, y):
    return x + y

>>> dis.dis(f)
  2           0 LOAD_FAST                0 (x) 
              3 LOAD_FAST                1 (y) 
              6 BINARY_ADD           
              7 RETURN_VALUE         
>>> f.__code__.co_varnames
('x', 'y')
>>> f.__code__.co_varnames[0]
'x'
>>> f.__code__.co_varnames[1]
'y'

dis.dis(f)的显示结果说明：

例如：2     0 LOAD_FAST      0 (x)

最左边的一列显示的是字节码指令对应的源码中行数，左起第2列显示的是当前的字节码指令在co_code中的偏移位置，第3列显示的是当前的字节码指令，最后一列显示的是当期字节码指令的参数。

其中三条指令的含义：

（1）LOAD_FAST(var_num)：Pushes a reference to the local co_varnames[var_num] onto the stack.

（2）BINARY_ADD：Implements TOS = TOS1 + TOS.

（3）RETURN_VALUE：Returns with TOS to the caller of the function.