cpython/Python/specialize.c

2027 lines
68 KiB
C
Raw Normal View History

#include "Python.h"
#include "pycore_code.h"
#include "pycore_dict.h"
bpo-46541: Replace core use of _Py_IDENTIFIER() with statically initialized global objects. (gh-30928) We're no longer using _Py_IDENTIFIER() (or _Py_static_string()) in any core CPython code. It is still used in a number of non-builtin stdlib modules. The replacement is: PyUnicodeObject (not pointer) fields under _PyRuntimeState, statically initialized as part of _PyRuntime. A new _Py_GET_GLOBAL_IDENTIFIER() macro facilitates lookup of the fields (along with _Py_GET_GLOBAL_STRING() for non-identifier strings). https://bugs.python.org/issue46541#msg411799 explains the rationale for this change. The core of the change is in: * (new) Include/internal/pycore_global_strings.h - the declarations for the global strings, along with the macros * Include/internal/pycore_runtime_init.h - added the static initializers for the global strings * Include/internal/pycore_global_objects.h - where the struct in pycore_global_strings.h is hooked into _PyRuntimeState * Tools/scripts/generate_global_objects.py - added generation of the global string declarations and static initializers I've also added a --check flag to generate_global_objects.py (along with make check-global-objects) to check for unused global strings. That check is added to the PR CI config. The remainder of this change updates the core code to use _Py_GET_GLOBAL_IDENTIFIER() instead of _Py_IDENTIFIER() and the related _Py*Id functions (likewise for _Py_GET_GLOBAL_STRING() instead of _Py_static_string()). This includes adding a few functions where there wasn't already an alternative to _Py*Id(), replacing the _Py_Identifier * parameter with PyObject *. The following are not changed (yet): * stop using _Py_IDENTIFIER() in the stdlib modules * (maybe) get rid of _Py_IDENTIFIER(), etc. entirely -- this may not be doable as at least one package on PyPI using this (private) API * (maybe) intern the strings during runtime init https://bugs.python.org/issue46541
2022-02-08 16:39:07 -04:00
#include "pycore_global_strings.h" // _Py_ID()
#include "pycore_long.h"
#include "pycore_moduleobject.h"
#include "pycore_object.h"
#include "opcode.h"
#include "structmember.h" // struct PyMemberDef, T_OFFSET_EX
#include <stdlib.h> // rand()
/* For guidance on adding or extending families of instructions see
* ./adaptive.md
*/
/* We layout the quickened data as a bi-directional array:
* Instructions upwards, cache entries downwards.
* first_instr is aligned to a SpecializedCacheEntry.
* The nth instruction is located at first_instr[n]
* The nth cache is located at ((SpecializedCacheEntry *)first_instr)[-1-n]
* The first (index 0) cache entry is reserved for the count, to enable finding
* the first instruction from the base pointer.
* The cache_count argument must include space for the count.
* We use the SpecializedCacheOrInstruction union to refer to the data
* to avoid type punning.
Layout of quickened data, each line 8 bytes for M cache entries and N instructions:
<cache_count> <---- co->co_quickened
<cache M-1>
<cache M-2>
...
<cache 0>
<instr 0> <instr 1> <instr 2> <instr 3> <--- co->co_first_instr
<instr 4> <instr 5> <instr 6> <instr 7>
...
<instr N-1>
*/
/* Map from opcode to adaptive opcode.
Values of zero are ignored. */
static uint8_t adaptive_opcodes[256] = {
[LOAD_ATTR] = LOAD_ATTR_ADAPTIVE,
[LOAD_GLOBAL] = LOAD_GLOBAL_ADAPTIVE,
[LOAD_METHOD] = LOAD_METHOD_ADAPTIVE,
[BINARY_SUBSCR] = BINARY_SUBSCR_ADAPTIVE,
[STORE_SUBSCR] = STORE_SUBSCR_ADAPTIVE,
[CALL] = CALL_ADAPTIVE,
[STORE_ATTR] = STORE_ATTR_ADAPTIVE,
[BINARY_OP] = BINARY_OP_ADAPTIVE,
[COMPARE_OP] = COMPARE_OP_ADAPTIVE,
};
/* The number of cache entries required for a "family" of instructions. */
static uint8_t cache_requirements[256] = {
[LOAD_ATTR] = 2, /* _PyAdaptiveEntry and _PyAttrCache */
[LOAD_GLOBAL] = 2, /* _PyAdaptiveEntry and _PyLoadGlobalCache */
[LOAD_METHOD] = 3, /* _PyAdaptiveEntry, _PyAttrCache and _PyObjectCache */
[BINARY_SUBSCR] = 2, /* _PyAdaptiveEntry, _PyObjectCache */
[STORE_SUBSCR] = 0,
[CALL] = 2, /* _PyAdaptiveEntry and _PyObjectCache/_PyCallCache */
[STORE_ATTR] = 2, /* _PyAdaptiveEntry and _PyAttrCache */
[BINARY_OP] = 1, // _PyAdaptiveEntry
[COMPARE_OP] = 1, /* _PyAdaptiveEntry */
};
Py_ssize_t _Py_QuickenedCount = 0;
#ifdef Py_STATS
PyStats _py_stats = { 0 };
#define ADD_STAT_TO_DICT(res, field) \
do { \
PyObject *val = PyLong_FromUnsignedLongLong(stats->field); \
if (val == NULL) { \
Py_DECREF(res); \
return NULL; \
} \
if (PyDict_SetItemString(res, #field, val) == -1) { \
Py_DECREF(res); \
Py_DECREF(val); \
return NULL; \
} \
Py_DECREF(val); \
} while(0);
static PyObject*
stats_to_dict(SpecializationStats *stats)
{
PyObject *res = PyDict_New();
if (res == NULL) {
return NULL;
}
ADD_STAT_TO_DICT(res, success);
ADD_STAT_TO_DICT(res, failure);
ADD_STAT_TO_DICT(res, hit);
ADD_STAT_TO_DICT(res, deferred);
ADD_STAT_TO_DICT(res, miss);
ADD_STAT_TO_DICT(res, deopt);
PyObject *failure_kinds = PyTuple_New(SPECIALIZATION_FAILURE_KINDS);
if (failure_kinds == NULL) {
Py_DECREF(res);
return NULL;
}
for (int i = 0; i < SPECIALIZATION_FAILURE_KINDS; i++) {
PyObject *stat = PyLong_FromUnsignedLongLong(stats->failure_kinds[i]);
if (stat == NULL) {
Py_DECREF(res);
Py_DECREF(failure_kinds);
return NULL;
}
PyTuple_SET_ITEM(failure_kinds, i, stat);
}
if (PyDict_SetItemString(res, "failure_kinds", failure_kinds)) {
Py_DECREF(res);
Py_DECREF(failure_kinds);
return NULL;
}
Py_DECREF(failure_kinds);
return res;
}
#undef ADD_STAT_TO_DICT
static int
add_stat_dict(
PyObject *res,
int opcode,
const char *name) {
SpecializationStats *stats = &_py_stats.opcode_stats[opcode].specialization;
PyObject *d = stats_to_dict(stats);
if (d == NULL) {
return -1;
}
int err = PyDict_SetItemString(res, name, d);
Py_DECREF(d);
return err;
}
#ifdef Py_STATS
PyObject*
_Py_GetSpecializationStats(void) {
PyObject *stats = PyDict_New();
if (stats == NULL) {
return NULL;
}
int err = 0;
err += add_stat_dict(stats, LOAD_ATTR, "load_attr");
err += add_stat_dict(stats, LOAD_GLOBAL, "load_global");
err += add_stat_dict(stats, LOAD_METHOD, "load_method");
err += add_stat_dict(stats, BINARY_SUBSCR, "binary_subscr");
err += add_stat_dict(stats, STORE_SUBSCR, "store_subscr");
err += add_stat_dict(stats, STORE_ATTR, "store_attr");
err += add_stat_dict(stats, CALL, "call");
err += add_stat_dict(stats, BINARY_OP, "binary_op");
err += add_stat_dict(stats, COMPARE_OP, "compare_op");
if (err < 0) {
Py_DECREF(stats);
return NULL;
}
return stats;
}
#endif
#define PRINT_STAT(i, field) \
if (stats[i].field) { \
fprintf(out, " opcode[%d]." #field " : %" PRIu64 "\n", i, stats[i].field); \
}
static void
print_spec_stats(FILE *out, OpcodeStats *stats)
{
/* Mark some opcodes as specializable for stats,
* even though we don't specialize them yet. */
fprintf(out, "opcode[%d].specializable : 1\n", FOR_ITER);
fprintf(out, "opcode[%d].specializable : 1\n", PRECALL_FUNCTION);
fprintf(out, "opcode[%d].specializable : 1\n", PRECALL_METHOD);
fprintf(out, "opcode[%d].specializable : 1\n", UNPACK_SEQUENCE);
for (int i = 0; i < 256; i++) {
if (adaptive_opcodes[i]) {
fprintf(out, "opcode[%d].specializable : 1\n", i);
}
PRINT_STAT(i, specialization.success);
PRINT_STAT(i, specialization.failure);
PRINT_STAT(i, specialization.hit);
PRINT_STAT(i, specialization.deferred);
PRINT_STAT(i, specialization.miss);
PRINT_STAT(i, specialization.deopt);
PRINT_STAT(i, execution_count);
for (int j = 0; j < SPECIALIZATION_FAILURE_KINDS; j++) {
uint64_t val = stats[i].specialization.failure_kinds[j];
if (val) {
fprintf(out, " opcode[%d].specialization.failure_kinds[%d] : %"
PRIu64 "\n", i, j, val);
}
}
for(int j = 0; j < 256; j++) {
if (stats[i].pair_count[j]) {
fprintf(out, "opcode[%d].pair_count[%d] : %" PRIu64 "\n",
i, j, stats[i].pair_count[j]);
}
}
}
}
#undef PRINT_STAT
static void
print_call_stats(FILE *out, CallStats *stats)
{
fprintf(out, "Calls to PyEval_EvalDefault: %" PRIu64 "\n", stats->pyeval_calls);
fprintf(out, "Calls to Python functions inlined: %" PRIu64 "\n", stats->inlined_py_calls);
fprintf(out, "Frames pushed: %" PRIu64 "\n", stats->frames_pushed);
fprintf(out, "Frame objects created: %" PRIu64 "\n", stats->frame_objects_created);
}
static void
print_object_stats(FILE *out, ObjectStats *stats)
{
fprintf(out, "Object allocations: %" PRIu64 "\n", stats->allocations);
fprintf(out, "Object frees: %" PRIu64 "\n", stats->frees);
fprintf(out, "Object new values: %" PRIu64 "\n", stats->new_values);
fprintf(out, "Object materialize dict (on request): %" PRIu64 "\n", stats->dict_materialized_on_request);
fprintf(out, "Object materialize dict (new key): %" PRIu64 "\n", stats->dict_materialized_new_key);
fprintf(out, "Object materialize dict (too big): %" PRIu64 "\n", stats->dict_materialized_too_big);
}
static void
print_stats(FILE *out, PyStats *stats) {
print_spec_stats(out, stats->opcode_stats);
print_call_stats(out, &stats->call_stats);
print_object_stats(out, &stats->object_stats);
}
void
_Py_PrintSpecializationStats(int to_file)
{
FILE *out = stderr;
if (to_file) {
/* Write to a file instead of stderr. */
# ifdef MS_WINDOWS
const char *dirname = "c:\\temp\\py_stats\\";
# else
const char *dirname = "/tmp/py_stats/";
# endif
/* Use random 160 bit number as file name,
* to avoid both accidental collisions and
* symlink attacks. */
unsigned char rand[20];
char hex_name[41];
_PyOS_URandomNonblock(rand, 20);
for (int i = 0; i < 20; i++) {
hex_name[2*i] = "0123456789abcdef"[rand[i]&15];
hex_name[2*i+1] = "0123456789abcdef"[(rand[i]>>4)&15];
}
hex_name[40] = '\0';
char buf[64];
assert(strlen(dirname) + 40 + strlen(".txt") < 64);
sprintf(buf, "%s%s.txt", dirname, hex_name);
FILE *fout = fopen(buf, "w");
if (fout) {
out = fout;
}
}
else {
fprintf(out, "Specialization stats:\n");
}
print_stats(out, &_py_stats);
if (out != stderr) {
fclose(out);
}
}
#ifdef Py_STATS
#define SPECIALIZATION_FAIL(opcode, kind) _py_stats.opcode_stats[opcode].specialization.failure_kinds[kind]++
#endif
#endif
#ifndef SPECIALIZATION_FAIL
#define SPECIALIZATION_FAIL(opcode, kind) ((void)0)
#endif
static SpecializedCacheOrInstruction *
allocate(int cache_count, int instruction_count)
{
assert(sizeof(SpecializedCacheOrInstruction) == 2*sizeof(int32_t));
assert(sizeof(SpecializedCacheEntry) == 2*sizeof(int32_t));
assert(cache_count > 0);
assert(instruction_count > 0);
int count = cache_count + (instruction_count + INSTRUCTIONS_PER_ENTRY -1)/INSTRUCTIONS_PER_ENTRY;
SpecializedCacheOrInstruction *array = (SpecializedCacheOrInstruction *)
PyMem_Malloc(sizeof(SpecializedCacheOrInstruction) * count);
if (array == NULL) {
PyErr_NoMemory();
return NULL;
}
_Py_QuickenedCount++;
array[0].entry.zero.cache_count = cache_count;
return array;
}
static int
get_cache_count(SpecializedCacheOrInstruction *quickened) {
return quickened[0].entry.zero.cache_count;
}
/* Return the oparg for the cache_offset and instruction index.
*
* If no cache is needed then return the original oparg.
* If a cache is needed, but cannot be accessed because
* oparg would be too large, then return -1.
*
* Also updates the cache_offset, as it may need to be incremented by
* more than the cache requirements, if many instructions do not need caches.
*
* See pycore_code.h for details of how the cache offset,
* instruction index and oparg are related */
static int
oparg_from_instruction_and_update_offset(int index, int opcode, int original_oparg, int *cache_offset) {
/* The instruction pointer in the interpreter points to the next
* instruction, so we compute the offset using nexti (index + 1) */
int nexti = index + 1;
uint8_t need = cache_requirements[opcode];
if (need == 0) {
return original_oparg;
}
assert(adaptive_opcodes[opcode] != 0);
int oparg = oparg_from_offset_and_nexti(*cache_offset, nexti);
assert(*cache_offset == offset_from_oparg_and_nexti(oparg, nexti));
/* Some cache space is wasted here as the minimum possible offset is (nexti>>1) */
if (oparg < 0) {
oparg = 0;
*cache_offset = offset_from_oparg_and_nexti(oparg, nexti);
}
else if (oparg > 255) {
return -1;
}
*cache_offset += need;
return oparg;
}
static int
2021-06-12 10:11:59 -03:00
entries_needed(const _Py_CODEUNIT *code, int len)
{
int cache_offset = 0;
int previous_opcode = -1;
for (int i = 0; i < len; i++) {
uint8_t opcode = _Py_OPCODE(code[i]);
if (previous_opcode != EXTENDED_ARG) {
oparg_from_instruction_and_update_offset(i, opcode, 0, &cache_offset);
}
previous_opcode = opcode;
}
return cache_offset + 1; // One extra for the count entry
}
static inline _Py_CODEUNIT *
first_instruction(SpecializedCacheOrInstruction *quickened)
{
return &quickened[get_cache_count(quickened)].code[0];
}
/** Insert adaptive instructions and superinstructions.
*
* Skip instruction preceded by EXTENDED_ARG for adaptive
* instructions as those are both very rare and tricky
* to handle.
*/
static void
optimize(SpecializedCacheOrInstruction *quickened, int len)
{
_Py_CODEUNIT *instructions = first_instruction(quickened);
int cache_offset = 0;
int previous_opcode = -1;
int previous_oparg = 0;
for(int i = 0; i < len; i++) {
int opcode = _Py_OPCODE(instructions[i]);
int oparg = _Py_OPARG(instructions[i]);
uint8_t adaptive_opcode = adaptive_opcodes[opcode];
if (adaptive_opcode && previous_opcode != EXTENDED_ARG) {
int new_oparg = oparg_from_instruction_and_update_offset(
i, opcode, oparg, &cache_offset
);
if (new_oparg < 0) {
/* Not possible to allocate a cache for this instruction */
previous_opcode = opcode;
continue;
}
previous_opcode = adaptive_opcode;
int entries_needed = cache_requirements[opcode];
if (entries_needed) {
/* Initialize the adpative cache entry */
int cache0_offset = cache_offset-entries_needed;
SpecializedCacheEntry *cache =
_GetSpecializedCacheEntry(instructions, cache0_offset);
cache->adaptive.original_oparg = oparg;
cache->adaptive.counter = 0;
} else {
// oparg is the adaptive cache counter
new_oparg = 0;
}
instructions[i] = _Py_MAKECODEUNIT(adaptive_opcode, new_oparg);
}
else {
/* Super instructions don't use the cache,
* so no need to update the offset. */
switch (opcode) {
case JUMP_ABSOLUTE:
instructions[i] = _Py_MAKECODEUNIT(JUMP_ABSOLUTE_QUICK, oparg);
break;
case RESUME:
instructions[i] = _Py_MAKECODEUNIT(RESUME_QUICK, oparg);
break;
case LOAD_FAST:
switch(previous_opcode) {
case LOAD_FAST:
instructions[i-1] = _Py_MAKECODEUNIT(LOAD_FAST__LOAD_FAST, previous_oparg);
break;
case STORE_FAST:
instructions[i-1] = _Py_MAKECODEUNIT(STORE_FAST__LOAD_FAST, previous_oparg);
break;
case LOAD_CONST:
instructions[i-1] = _Py_MAKECODEUNIT(LOAD_CONST__LOAD_FAST, previous_oparg);
break;
}
break;
case STORE_FAST:
if (previous_opcode == STORE_FAST) {
instructions[i-1] = _Py_MAKECODEUNIT(STORE_FAST__STORE_FAST, previous_oparg);
}
break;
case LOAD_CONST:
if (previous_opcode == LOAD_FAST) {
instructions[i-1] = _Py_MAKECODEUNIT(LOAD_FAST__LOAD_CONST, previous_oparg);
}
break;
}
previous_opcode = opcode;
previous_oparg = oparg;
}
}
assert(cache_offset+1 == get_cache_count(quickened));
}
int
_Py_Quicken(PyCodeObject *code) {
if (code->co_quickened) {
return 0;
}
Py_ssize_t size = PyBytes_GET_SIZE(code->co_code);
int instr_count = (int)(size/sizeof(_Py_CODEUNIT));
if (instr_count > MAX_SIZE_TO_QUICKEN) {
code->co_warmup = QUICKENING_WARMUP_COLDEST;
return 0;
}
int entry_count = entries_needed(code->co_firstinstr, instr_count);
SpecializedCacheOrInstruction *quickened = allocate(entry_count, instr_count);
if (quickened == NULL) {
return -1;
}
_Py_CODEUNIT *new_instructions = first_instruction(quickened);
memcpy(new_instructions, code->co_firstinstr, size);
optimize(quickened, instr_count);
code->co_quickened = quickened;
code->co_firstinstr = new_instructions;
return 0;
}
static inline int
initial_counter_value(void) {
/* Starting value for the counter.
* This value needs to be not too low, otherwise
* it would cause excessive de-optimization.
* Neither should it be too high, or that would delay
* de-optimization excessively when it is needed.
* A value around 50 seems to work, and we choose a
* prime number to avoid artifacts.
*/
return 53;
}
/* Common */
2021-08-27 08:01:22 -03:00
#define SPEC_FAIL_OTHER 0
#define SPEC_FAIL_NO_DICT 1
#define SPEC_FAIL_OVERRIDDEN 2
#define SPEC_FAIL_OUT_OF_VERSIONS 3
#define SPEC_FAIL_OUT_OF_RANGE 4
#define SPEC_FAIL_EXPECTED_ERROR 5
#define SPEC_FAIL_WRONG_NUMBER_ARGUMENTS 6
#define SPEC_FAIL_LOAD_GLOBAL_NON_STRING_OR_SPLIT 18
/* Attributes */
#define SPEC_FAIL_ATTR_OVERRIDING_DESCRIPTOR 8
#define SPEC_FAIL_ATTR_NON_OVERRIDING_DESCRIPTOR 9
#define SPEC_FAIL_ATTR_NOT_DESCRIPTOR 10
#define SPEC_FAIL_ATTR_METHOD 11
#define SPEC_FAIL_ATTR_MUTABLE_CLASS 12
#define SPEC_FAIL_ATTR_PROPERTY 13
#define SPEC_FAIL_ATTR_NON_OBJECT_SLOT 14
#define SPEC_FAIL_ATTR_READ_ONLY 15
#define SPEC_FAIL_ATTR_AUDITED_SLOT 16
#define SPEC_FAIL_ATTR_NOT_MANAGED_DICT 17
#define SPEC_FAIL_ATTR_NON_STRING_OR_SPLIT 18
#define SPEC_FAIL_ATTR_MODULE_ATTR_NOT_FOUND 19
/* Methods */
#define SPEC_FAIL_LOAD_METHOD_OVERRIDING_DESCRIPTOR 8
#define SPEC_FAIL_LOAD_METHOD_NON_OVERRIDING_DESCRIPTOR 9
#define SPEC_FAIL_LOAD_METHOD_NOT_DESCRIPTOR 10
#define SPEC_FAIL_LOAD_METHOD_METHOD 11
#define SPEC_FAIL_LOAD_METHOD_MUTABLE_CLASS 12
#define SPEC_FAIL_LOAD_METHOD_PROPERTY 13
#define SPEC_FAIL_LOAD_METHOD_NON_OBJECT_SLOT 14
#define SPEC_FAIL_LOAD_METHOD_IS_ATTR 15
#define SPEC_FAIL_LOAD_METHOD_DICT_SUBCLASS 16
#define SPEC_FAIL_LOAD_METHOD_BUILTIN_CLASS_METHOD 17
#define SPEC_FAIL_LOAD_METHOD_CLASS_METHOD_OBJ 18
#define SPEC_FAIL_LOAD_METHOD_OBJECT_SLOT 19
#define SPEC_FAIL_LOAD_METHOD_HAS_DICT 20
#define SPEC_FAIL_LOAD_METHOD_HAS_MANAGED_DICT 21
#define SPEC_FAIL_LOAD_METHOD_INSTANCE_ATTRIBUTE 22
#define SPEC_FAIL_LOAD_METHOD_METACLASS_ATTRIBUTE 23
/* Binary subscr and store subscr */
#define SPEC_FAIL_SUBSCR_ARRAY_INT 8
#define SPEC_FAIL_SUBSCR_ARRAY_SLICE 9
#define SPEC_FAIL_SUBSCR_LIST_SLICE 10
#define SPEC_FAIL_SUBSCR_TUPLE_SLICE 11
#define SPEC_FAIL_SUBSCR_STRING_INT 12
#define SPEC_FAIL_SUBSCR_STRING_SLICE 13
#define SPEC_FAIL_SUBSCR_BUFFER_INT 15
#define SPEC_FAIL_SUBSCR_BUFFER_SLICE 16
#define SPEC_FAIL_SUBSCR_SEQUENCE_INT 17
/* Store subscr */
#define SPEC_FAIL_SUBSCR_BYTEARRAY_INT 18
#define SPEC_FAIL_SUBSCR_BYTEARRAY_SLICE 19
#define SPEC_FAIL_SUBSCR_PY_SIMPLE 20
#define SPEC_FAIL_SUBSCR_PY_OTHER 21
#define SPEC_FAIL_SUBSCR_DICT_SUBCLASS_NO_OVERRIDE 22
/* Binary add */
#define SPEC_FAIL_BINARY_OP_DIFFERENT_TYPES 12
/* Calls */
#define SPEC_FAIL_CALL_COMPLEX_PARAMETERS 9
#define SPEC_FAIL_CALL_CO_NOT_OPTIMIZED 10
/* SPEC_FAIL_METHOD defined as 11 above */
#define SPEC_FAIL_CALL_INSTANCE_METHOD 11
#define SPEC_FAIL_CALL_CMETHOD 12
#define SPEC_FAIL_CALL_PYCFUNCTION 13
#define SPEC_FAIL_CALL_PYCFUNCTION_WITH_KEYWORDS 14
#define SPEC_FAIL_CALL_PYCFUNCTION_FAST_WITH_KEYWORDS 15
#define SPEC_FAIL_CALL_PYCFUNCTION_NOARGS 16
#define SPEC_FAIL_CALL_BAD_CALL_FLAGS 17
#define SPEC_FAIL_CALL_CLASS 18
#define SPEC_FAIL_CALL_PYTHON_CLASS 19
#define SPEC_FAIL_CALL_METHOD_DESCRIPTOR 20
#define SPEC_FAIL_CALL_BOUND_METHOD 21
#define SPEC_FAIL_CALL_STR 22
#define SPEC_FAIL_CALL_CLASS_NO_VECTORCALL 23
#define SPEC_FAIL_CALL_CLASS_MUTABLE 24
#define SPEC_FAIL_CALL_KWNAMES 25
#define SPEC_FAIL_CALL_METHOD_WRAPPER 26
#define SPEC_FAIL_CALL_OPERATOR_WRAPPER 27
#define SPEC_FAIL_CALL_PYFUNCTION 28
/* COMPARE_OP */
#define SPEC_FAIL_COMPARE_OP_DIFFERENT_TYPES 12
#define SPEC_FAIL_COMPARE_OP_STRING 13
#define SPEC_FAIL_COMPARE_OP_NOT_FOLLOWED_BY_COND_JUMP 14
#define SPEC_FAIL_COMPARE_OP_BIG_INT 15
#define SPEC_FAIL_COMPARE_OP_BYTES 16
#define SPEC_FAIL_COMPARE_OP_TUPLE 17
#define SPEC_FAIL_COMPARE_OP_LIST 18
#define SPEC_FAIL_COMPARE_OP_SET 19
#define SPEC_FAIL_COMPARE_OP_BOOL 20
#define SPEC_FAIL_COMPARE_OP_BASEOBJECT 21
#define SPEC_FAIL_COMPARE_OP_FLOAT_LONG 22
#define SPEC_FAIL_COMPARE_OP_LONG_FLOAT 23
/* FOR_ITER */
#define SPEC_FAIL_FOR_ITER_GENERATOR 10
#define SPEC_FAIL_FOR_ITER_COROUTINE 11
#define SPEC_FAIL_FOR_ITER_ASYNC_GENERATOR 12
#define SPEC_FAIL_FOR_ITER_LIST 13
#define SPEC_FAIL_FOR_ITER_TUPLE 14
#define SPEC_FAIL_FOR_ITER_SET 15
#define SPEC_FAIL_FOR_ITER_STRING 16
#define SPEC_FAIL_FOR_ITER_BYTES 17
#define SPEC_FAIL_FOR_ITER_RANGE 18
#define SPEC_FAIL_FOR_ITER_ITERTOOLS 19
#define SPEC_FAIL_FOR_ITER_DICT_KEYS 20
#define SPEC_FAIL_FOR_ITER_DICT_ITEMS 21
#define SPEC_FAIL_FOR_ITER_DICT_VALUES 22
#define SPEC_FAIL_FOR_ITER_ENUMERATE 23
/* UNPACK_SEQUENCE */
#define SPEC_FAIL_UNPACK_SEQUENCE_TUPLE_0 9
#define SPEC_FAIL_UNPACK_SEQUENCE_TUPLE_1 10
#define SPEC_FAIL_UNPACK_SEQUENCE_TUPLE_2 11
#define SPEC_FAIL_UNPACK_SEQUENCE_TUPLE_3 12
#define SPEC_FAIL_UNPACK_SEQUENCE_TUPLE_4 13
#define SPEC_FAIL_UNPACK_SEQUENCE_TUPLE_N 14
#define SPEC_FAIL_UNPACK_SEQUENCE_LIST_0 15
#define SPEC_FAIL_UNPACK_SEQUENCE_LIST_1 16
#define SPEC_FAIL_UNPACK_SEQUENCE_LIST_2 17
#define SPEC_FAIL_UNPACK_SEQUENCE_LIST_3 18
#define SPEC_FAIL_UNPACK_SEQUENCE_LIST_4 19
#define SPEC_FAIL_UNPACK_SEQUENCE_LIST_N 20
#define SPEC_FAIL_UNPACK_SEQUENCE_OTHER_0 21
#define SPEC_FAIL_UNPACK_SEQUENCE_OTHER_1 22
#define SPEC_FAIL_UNPACK_SEQUENCE_OTHER_2 23
#define SPEC_FAIL_UNPACK_SEQUENCE_OTHER_3 24
#define SPEC_FAIL_UNPACK_SEQUENCE_OTHER_4 25
#define SPEC_FAIL_UNPACK_SEQUENCE_OTHER_N 26
static int
specialize_module_load_attr(
PyObject *owner, _Py_CODEUNIT *instr, PyObject *name,
_PyAdaptiveEntry *cache0, _PyAttrCache *cache1, int opcode,
int opcode_module)
{
PyModuleObject *m = (PyModuleObject *)owner;
PyObject *value = NULL;
assert((owner->ob_type->tp_flags & Py_TPFLAGS_MANAGED_DICT) == 0);
PyDictObject *dict = (PyDictObject *)m->md_dict;
if (dict == NULL) {
SPECIALIZATION_FAIL(opcode, SPEC_FAIL_NO_DICT);
return -1;
}
if (dict->ma_keys->dk_kind != DICT_KEYS_UNICODE) {
SPECIALIZATION_FAIL(opcode, SPEC_FAIL_ATTR_NON_STRING_OR_SPLIT);
return -1;
}
bpo-46541: Replace core use of _Py_IDENTIFIER() with statically initialized global objects. (gh-30928) We're no longer using _Py_IDENTIFIER() (or _Py_static_string()) in any core CPython code. It is still used in a number of non-builtin stdlib modules. The replacement is: PyUnicodeObject (not pointer) fields under _PyRuntimeState, statically initialized as part of _PyRuntime. A new _Py_GET_GLOBAL_IDENTIFIER() macro facilitates lookup of the fields (along with _Py_GET_GLOBAL_STRING() for non-identifier strings). https://bugs.python.org/issue46541#msg411799 explains the rationale for this change. The core of the change is in: * (new) Include/internal/pycore_global_strings.h - the declarations for the global strings, along with the macros * Include/internal/pycore_runtime_init.h - added the static initializers for the global strings * Include/internal/pycore_global_objects.h - where the struct in pycore_global_strings.h is hooked into _PyRuntimeState * Tools/scripts/generate_global_objects.py - added generation of the global string declarations and static initializers I've also added a --check flag to generate_global_objects.py (along with make check-global-objects) to check for unused global strings. That check is added to the PR CI config. The remainder of this change updates the core code to use _Py_GET_GLOBAL_IDENTIFIER() instead of _Py_IDENTIFIER() and the related _Py*Id functions (likewise for _Py_GET_GLOBAL_STRING() instead of _Py_static_string()). This includes adding a few functions where there wasn't already an alternative to _Py*Id(), replacing the _Py_Identifier * parameter with PyObject *. The following are not changed (yet): * stop using _Py_IDENTIFIER() in the stdlib modules * (maybe) get rid of _Py_IDENTIFIER(), etc. entirely -- this may not be doable as at least one package on PyPI using this (private) API * (maybe) intern the strings during runtime init https://bugs.python.org/issue46541
2022-02-08 16:39:07 -04:00
Py_ssize_t index = _PyDict_GetItemHint(dict, &_Py_ID(__getattr__), -1,
&value);
assert(index != DKIX_ERROR);
if (index != DKIX_EMPTY) {
SPECIALIZATION_FAIL(opcode, SPEC_FAIL_ATTR_MODULE_ATTR_NOT_FOUND);
return -1;
}
index = _PyDict_GetItemHint(dict, name, -1, &value);
assert (index != DKIX_ERROR);
if (index != (uint16_t)index) {
SPECIALIZATION_FAIL(opcode, SPEC_FAIL_OUT_OF_RANGE);
return -1;
}
uint32_t keys_version = _PyDictKeys_GetVersionForCurrentState(dict->ma_keys);
if (keys_version == 0) {
SPECIALIZATION_FAIL(opcode, SPEC_FAIL_OUT_OF_VERSIONS);
return -1;
}
cache1->dk_version_or_hint = keys_version;
cache0->index = (uint16_t)index;
*instr = _Py_MAKECODEUNIT(opcode_module, _Py_OPARG(*instr));
return 0;
}
/* Attribute specialization */
typedef enum {
OVERRIDING, /* Is an overriding descriptor, and will remain so. */
METHOD, /* Attribute has Py_TPFLAGS_METHOD_DESCRIPTOR set */
PROPERTY, /* Is a property */
OBJECT_SLOT, /* Is an object slot descriptor */
OTHER_SLOT, /* Is a slot descriptor of another type */
NON_OVERRIDING, /* Is another non-overriding descriptor, and is an instance of an immutable class*/
BUILTIN_CLASSMETHOD, /* Builtin methods with METH_CLASS */
PYTHON_CLASSMETHOD, /* Python classmethod(func) object */
NON_DESCRIPTOR, /* Is not a descriptor, and is an instance of an immutable class */
MUTABLE, /* Instance of a mutable class; might, or might not, be a descriptor */
ABSENT, /* Attribute is not present on the class */
DUNDER_CLASS, /* __class__ attribute */
GETSET_OVERRIDDEN /* __getattribute__ or __setattr__ has been overridden */
} DescriptorClassification;
static DescriptorClassification
analyze_descriptor(PyTypeObject *type, PyObject *name, PyObject **descr, int store)
{
if (store) {
if (type->tp_setattro != PyObject_GenericSetAttr) {
*descr = NULL;
return GETSET_OVERRIDDEN;
}
}
else {
if (type->tp_getattro != PyObject_GenericGetAttr) {
*descr = NULL;
return GETSET_OVERRIDDEN;
}
}
PyObject *descriptor = _PyType_Lookup(type, name);
*descr = descriptor;
if (descriptor == NULL) {
return ABSENT;
}
PyTypeObject *desc_cls = Py_TYPE(descriptor);
if (!(desc_cls->tp_flags & Py_TPFLAGS_IMMUTABLETYPE)) {
return MUTABLE;
}
if (desc_cls->tp_descr_set) {
if (desc_cls == &PyMemberDescr_Type) {
PyMemberDescrObject *member = (PyMemberDescrObject *)descriptor;
struct PyMemberDef *dmem = member->d_member;
if (dmem->type == T_OBJECT_EX) {
return OBJECT_SLOT;
}
return OTHER_SLOT;
}
if (desc_cls == &PyProperty_Type) {
return PROPERTY;
}
if (PyUnicode_CompareWithASCIIString(name, "__class__") == 0) {
if (descriptor == _PyType_Lookup(&PyBaseObject_Type, name)) {
return DUNDER_CLASS;
}
}
return OVERRIDING;
}
if (desc_cls->tp_descr_get) {
if (desc_cls->tp_flags & Py_TPFLAGS_METHOD_DESCRIPTOR) {
return METHOD;
}
if (Py_IS_TYPE(descriptor, &PyClassMethodDescr_Type)) {
return BUILTIN_CLASSMETHOD;
}
if (Py_IS_TYPE(descriptor, &PyClassMethod_Type)) {
return PYTHON_CLASSMETHOD;
}
return NON_OVERRIDING;
}
return NON_DESCRIPTOR;
}
static int
specialize_dict_access(
PyObject *owner, _Py_CODEUNIT *instr, PyTypeObject *type,
DescriptorClassification kind, PyObject *name,
_PyAdaptiveEntry *cache0, _PyAttrCache *cache1,
int base_op, int values_op, int hint_op)
{
assert(kind == NON_OVERRIDING || kind == NON_DESCRIPTOR || kind == ABSENT ||
kind == BUILTIN_CLASSMETHOD || kind == PYTHON_CLASSMETHOD);
// No descriptor, or non overriding.
if ((type->tp_flags & Py_TPFLAGS_MANAGED_DICT) == 0) {
SPECIALIZATION_FAIL(base_op, SPEC_FAIL_ATTR_NOT_MANAGED_DICT);
return 0;
}
PyObject **dictptr = _PyObject_ManagedDictPointer(owner);
PyDictObject *dict = (PyDictObject *)*dictptr;
if (dict == NULL) {
// Virtual dictionary
PyDictKeysObject *keys = ((PyHeapTypeObject *)type)->ht_cached_keys;
assert(PyUnicode_CheckExact(name));
Py_ssize_t index = _PyDictKeys_StringLookup(keys, name);
assert (index != DKIX_ERROR);
if (index != (uint16_t)index) {
SPECIALIZATION_FAIL(base_op, SPEC_FAIL_OUT_OF_RANGE);
return 0;
}
cache1->tp_version = type->tp_version_tag;
cache0->index = (uint16_t)index;
*instr = _Py_MAKECODEUNIT(values_op, _Py_OPARG(*instr));
}
else {
if (!PyDict_CheckExact(dict)) {
SPECIALIZATION_FAIL(base_op, SPEC_FAIL_NO_DICT);
return 0;
}
// We found an instance with a __dict__.
PyObject *value = NULL;
Py_ssize_t hint =
_PyDict_GetItemHint(dict, name, -1, &value);
if (hint != (uint32_t)hint) {
SPECIALIZATION_FAIL(base_op, SPEC_FAIL_OUT_OF_RANGE);
return 0;
}
cache1->dk_version_or_hint = (uint32_t)hint;
cache1->tp_version = type->tp_version_tag;
*instr = _Py_MAKECODEUNIT(hint_op, _Py_OPARG(*instr));
}
return 1;
}
int
_Py_Specialize_LoadAttr(PyObject *owner, _Py_CODEUNIT *instr, PyObject *name, SpecializedCacheEntry *cache)
{
_PyAdaptiveEntry *cache0 = &cache->adaptive;
_PyAttrCache *cache1 = &cache[-1].attr;
if (PyModule_CheckExact(owner)) {
int err = specialize_module_load_attr(owner, instr, name, cache0, cache1,
LOAD_ATTR, LOAD_ATTR_MODULE);
if (err) {
goto fail;
}
goto success;
}
PyTypeObject *type = Py_TYPE(owner);
if (type->tp_dict == NULL) {
if (PyType_Ready(type) < 0) {
return -1;
}
}
PyObject *descr;
DescriptorClassification kind = analyze_descriptor(type, name, &descr, 0);
switch(kind) {
case OVERRIDING:
SPECIALIZATION_FAIL(LOAD_ATTR, SPEC_FAIL_ATTR_OVERRIDING_DESCRIPTOR);
goto fail;
case METHOD:
SPECIALIZATION_FAIL(LOAD_ATTR, SPEC_FAIL_ATTR_METHOD);
goto fail;
case PROPERTY:
SPECIALIZATION_FAIL(LOAD_ATTR, SPEC_FAIL_ATTR_PROPERTY);
goto fail;
case OBJECT_SLOT:
{
PyMemberDescrObject *member = (PyMemberDescrObject *)descr;
struct PyMemberDef *dmem = member->d_member;
Py_ssize_t offset = dmem->offset;
if (dmem->flags & PY_AUDIT_READ) {
SPECIALIZATION_FAIL(LOAD_ATTR, SPEC_FAIL_ATTR_AUDITED_SLOT);
goto fail;
}
if (offset != (uint16_t)offset) {
SPECIALIZATION_FAIL(LOAD_ATTR, SPEC_FAIL_OUT_OF_RANGE);
goto fail;
}
assert(dmem->type == T_OBJECT_EX);
assert(offset > 0);
cache0->index = (uint16_t)offset;
cache1->tp_version = type->tp_version_tag;
*instr = _Py_MAKECODEUNIT(LOAD_ATTR_SLOT, _Py_OPARG(*instr));
goto success;
}
case DUNDER_CLASS:
{
Py_ssize_t offset = offsetof(PyObject, ob_type);
assert(offset == (uint16_t)offset);
cache0->index = (uint16_t)offset;
cache1->tp_version = type->tp_version_tag;
*instr = _Py_MAKECODEUNIT(LOAD_ATTR_SLOT, _Py_OPARG(*instr));
goto success;
}
case OTHER_SLOT:
SPECIALIZATION_FAIL(LOAD_ATTR, SPEC_FAIL_ATTR_NON_OBJECT_SLOT);
goto fail;
case MUTABLE:
SPECIALIZATION_FAIL(LOAD_ATTR, SPEC_FAIL_ATTR_MUTABLE_CLASS);
goto fail;
case GETSET_OVERRIDDEN:
SPECIALIZATION_FAIL(LOAD_ATTR, SPEC_FAIL_OVERRIDDEN);
goto fail;
case BUILTIN_CLASSMETHOD:
case PYTHON_CLASSMETHOD:
case NON_OVERRIDING:
case NON_DESCRIPTOR:
case ABSENT:
break;
}
int err = specialize_dict_access(
owner, instr, type, kind, name, cache0, cache1,
LOAD_ATTR, LOAD_ATTR_INSTANCE_VALUE, LOAD_ATTR_WITH_HINT
);
if (err < 0) {
return -1;
}
if (err) {
goto success;
}
fail:
STAT_INC(LOAD_ATTR, failure);
assert(!PyErr_Occurred());
cache_backoff(cache0);
return 0;
success:
STAT_INC(LOAD_ATTR, success);
assert(!PyErr_Occurred());
cache0->counter = initial_counter_value();
return 0;
}
int
_Py_Specialize_StoreAttr(PyObject *owner, _Py_CODEUNIT *instr, PyObject *name, SpecializedCacheEntry *cache)
{
_PyAdaptiveEntry *cache0 = &cache->adaptive;
_PyAttrCache *cache1 = &cache[-1].attr;
PyTypeObject *type = Py_TYPE(owner);
if (PyModule_CheckExact(owner)) {
SPECIALIZATION_FAIL(STORE_ATTR, SPEC_FAIL_OVERRIDDEN);
goto fail;
}
PyObject *descr;
DescriptorClassification kind = analyze_descriptor(type, name, &descr, 1);
switch(kind) {
case OVERRIDING:
SPECIALIZATION_FAIL(STORE_ATTR, SPEC_FAIL_ATTR_OVERRIDING_DESCRIPTOR);
goto fail;
case METHOD:
SPECIALIZATION_FAIL(STORE_ATTR, SPEC_FAIL_ATTR_METHOD);
goto fail;
case PROPERTY:
SPECIALIZATION_FAIL(STORE_ATTR, SPEC_FAIL_ATTR_PROPERTY);
goto fail;
case OBJECT_SLOT:
{
PyMemberDescrObject *member = (PyMemberDescrObject *)descr;
struct PyMemberDef *dmem = member->d_member;
Py_ssize_t offset = dmem->offset;
if (dmem->flags & READONLY) {
SPECIALIZATION_FAIL(STORE_ATTR, SPEC_FAIL_ATTR_READ_ONLY);
goto fail;
}
if (offset != (uint16_t)offset) {
SPECIALIZATION_FAIL(STORE_ATTR, SPEC_FAIL_OUT_OF_RANGE);
goto fail;
}
assert(dmem->type == T_OBJECT_EX);
assert(offset > 0);
cache0->index = (uint16_t)offset;
cache1->tp_version = type->tp_version_tag;
*instr = _Py_MAKECODEUNIT(STORE_ATTR_SLOT, _Py_OPARG(*instr));
goto success;
}
case DUNDER_CLASS:
case OTHER_SLOT:
SPECIALIZATION_FAIL(STORE_ATTR, SPEC_FAIL_ATTR_NON_OBJECT_SLOT);
goto fail;
case MUTABLE:
SPECIALIZATION_FAIL(STORE_ATTR, SPEC_FAIL_ATTR_MUTABLE_CLASS);
goto fail;
case GETSET_OVERRIDDEN:
SPECIALIZATION_FAIL(STORE_ATTR, SPEC_FAIL_OVERRIDDEN);
goto fail;
case BUILTIN_CLASSMETHOD:
case PYTHON_CLASSMETHOD:
case NON_OVERRIDING:
case NON_DESCRIPTOR:
case ABSENT:
break;
}
int err = specialize_dict_access(
owner, instr, type, kind, name, cache0, cache1,
STORE_ATTR, STORE_ATTR_INSTANCE_VALUE, STORE_ATTR_WITH_HINT
);
if (err < 0) {
return -1;
}
if (err) {
goto success;
}
fail:
STAT_INC(STORE_ATTR, failure);
assert(!PyErr_Occurred());
cache_backoff(cache0);
return 0;
success:
STAT_INC(STORE_ATTR, success);
assert(!PyErr_Occurred());
cache0->counter = initial_counter_value();
return 0;
}
#ifdef Py_STATS
static int
load_method_fail_kind(DescriptorClassification kind)
{
switch (kind) {
case OVERRIDING:
return SPEC_FAIL_LOAD_METHOD_OVERRIDING_DESCRIPTOR;
case METHOD:
return SPEC_FAIL_LOAD_METHOD_METHOD;
case PROPERTY:
return SPEC_FAIL_LOAD_METHOD_PROPERTY;
case OBJECT_SLOT:
return SPEC_FAIL_LOAD_METHOD_OBJECT_SLOT;
case OTHER_SLOT:
return SPEC_FAIL_LOAD_METHOD_NON_OBJECT_SLOT;
case DUNDER_CLASS:
return SPEC_FAIL_OTHER;
case MUTABLE:
return SPEC_FAIL_LOAD_METHOD_MUTABLE_CLASS;
case GETSET_OVERRIDDEN:
return SPEC_FAIL_OVERRIDDEN;
case BUILTIN_CLASSMETHOD:
return SPEC_FAIL_LOAD_METHOD_BUILTIN_CLASS_METHOD;
case PYTHON_CLASSMETHOD:
return SPEC_FAIL_LOAD_METHOD_CLASS_METHOD_OBJ;
case NON_OVERRIDING:
return SPEC_FAIL_LOAD_METHOD_NON_OVERRIDING_DESCRIPTOR;
case NON_DESCRIPTOR:
return SPEC_FAIL_LOAD_METHOD_NOT_DESCRIPTOR;
case ABSENT:
return SPEC_FAIL_LOAD_METHOD_INSTANCE_ATTRIBUTE;
}
Py_UNREACHABLE();
}
#endif
static int
specialize_class_load_method(PyObject *owner, _Py_CODEUNIT *instr, PyObject *name,
_PyAttrCache *cache1, _PyObjectCache *cache2)
{
PyObject *descr = NULL;
DescriptorClassification kind = 0;
kind = analyze_descriptor((PyTypeObject *)owner, name, &descr, 0);
switch (kind) {
case METHOD:
case NON_DESCRIPTOR:
cache1->tp_version = ((PyTypeObject *)owner)->tp_version_tag;
cache2->obj = descr;
*instr = _Py_MAKECODEUNIT(LOAD_METHOD_CLASS, _Py_OPARG(*instr));
return 0;
#ifdef Py_STATS
case ABSENT:
if (_PyType_Lookup(Py_TYPE(owner), name) != NULL) {
SPECIALIZATION_FAIL(LOAD_METHOD, SPEC_FAIL_LOAD_METHOD_METACLASS_ATTRIBUTE);
}
else {
SPECIALIZATION_FAIL(LOAD_METHOD, SPEC_FAIL_EXPECTED_ERROR);
}
return -1;
#endif
default:
SPECIALIZATION_FAIL(LOAD_METHOD, load_method_fail_kind(kind));
return -1;
}
}
// Please collect stats carefully before and after modifying. A subtle change
// can cause a significant drop in cache hits. A possible test is
// python.exe -m test_typing test_re test_dis test_zlib.
int
_Py_Specialize_LoadMethod(PyObject *owner, _Py_CODEUNIT *instr, PyObject *name, SpecializedCacheEntry *cache)
{
_PyAdaptiveEntry *cache0 = &cache->adaptive;
_PyAttrCache *cache1 = &cache[-1].attr;
_PyObjectCache *cache2 = &cache[-2].obj;
PyTypeObject *owner_cls = Py_TYPE(owner);
if (PyModule_CheckExact(owner)) {
int err = specialize_module_load_attr(owner, instr, name, cache0, cache1,
LOAD_METHOD, LOAD_METHOD_MODULE);
if (err) {
goto fail;
}
goto success;
}
if (owner_cls->tp_dict == NULL) {
if (PyType_Ready(owner_cls) < 0) {
return -1;
}
}
if (PyType_Check(owner)) {
int err = specialize_class_load_method(owner, instr, name, cache1, cache2);
if (err) {
goto fail;
}
goto success;
}
PyObject *descr = NULL;
DescriptorClassification kind = 0;
kind = analyze_descriptor(owner_cls, name, &descr, 0);
assert(descr != NULL || kind == ABSENT || kind == GETSET_OVERRIDDEN);
if (kind != METHOD) {
SPECIALIZATION_FAIL(LOAD_METHOD, load_method_fail_kind(kind));
goto fail;
}
if (owner_cls->tp_flags & Py_TPFLAGS_MANAGED_DICT) {
PyObject **owner_dictptr = _PyObject_ManagedDictPointer(owner);
if (*owner_dictptr) {
SPECIALIZATION_FAIL(LOAD_METHOD, SPEC_FAIL_LOAD_METHOD_HAS_MANAGED_DICT);
goto fail;
}
PyDictKeysObject *keys = ((PyHeapTypeObject *)owner_cls)->ht_cached_keys;
Py_ssize_t index = _PyDictKeys_StringLookup(keys, name);
if (index != DKIX_EMPTY) {
SPECIALIZATION_FAIL(LOAD_METHOD, SPEC_FAIL_LOAD_METHOD_IS_ATTR);
goto fail;
}
uint32_t keys_version = _PyDictKeys_GetVersionForCurrentState(keys);
if (keys_version == 0) {
SPECIALIZATION_FAIL(LOAD_METHOD, SPEC_FAIL_OUT_OF_VERSIONS);
goto fail;
}
cache1->dk_version_or_hint = keys_version;
*instr = _Py_MAKECODEUNIT(LOAD_METHOD_CACHED, _Py_OPARG(*instr));
}
else {
if (owner_cls->tp_dictoffset == 0) {
*instr = _Py_MAKECODEUNIT(LOAD_METHOD_NO_DICT, _Py_OPARG(*instr));
}
else {
SPECIALIZATION_FAIL(LOAD_METHOD, SPEC_FAIL_LOAD_METHOD_HAS_DICT);
goto fail;
}
}
/* `descr` is borrowed. This is safe for methods (even inherited ones from
* super classes!) as long as tp_version_tag is validated for two main reasons:
*
* 1. The class will always hold a reference to the method so it will
* usually not be GC-ed. Should it be deleted in Python, e.g.
* `del obj.meth`, tp_version_tag will be invalidated, because of reason 2.
*
* 2. The pre-existing type method cache (MCACHE) uses the same principles
* of caching a borrowed descriptor. The MCACHE infrastructure does all the
* heavy lifting for us. E.g. it invalidates tp_version_tag on any MRO
* modification, on any type object change along said MRO, etc. (see
* PyType_Modified usages in typeobject.c). The MCACHE has been
* working since Python 2.6 and it's battle-tested.
*/
cache1->tp_version = owner_cls->tp_version_tag;
cache2->obj = descr;
// Fall through.
success:
STAT_INC(LOAD_METHOD, success);
assert(!PyErr_Occurred());
cache0->counter = initial_counter_value();
return 0;
fail:
STAT_INC(LOAD_METHOD, failure);
assert(!PyErr_Occurred());
cache_backoff(cache0);
return 0;
}
int
_Py_Specialize_LoadGlobal(
PyObject *globals, PyObject *builtins,
_Py_CODEUNIT *instr, PyObject *name,
SpecializedCacheEntry *cache)
{
_PyAdaptiveEntry *cache0 = &cache->adaptive;
_PyLoadGlobalCache *cache1 = &cache[-1].load_global;
assert(PyUnicode_CheckExact(name));
if (!PyDict_CheckExact(globals)) {
goto fail;
}
PyDictKeysObject * globals_keys = ((PyDictObject *)globals)->ma_keys;
Py_ssize_t index = _PyDictKeys_StringLookup(globals_keys, name);
if (index == DKIX_ERROR) {
SPECIALIZATION_FAIL(LOAD_GLOBAL, SPEC_FAIL_LOAD_GLOBAL_NON_STRING_OR_SPLIT);
goto fail;
}
if (index != DKIX_EMPTY) {
if (index != (uint16_t)index) {
goto fail;
}
uint32_t keys_version = _PyDictKeys_GetVersionForCurrentState(globals_keys);
if (keys_version == 0) {
goto fail;
}
cache1->module_keys_version = keys_version;
cache0->index = (uint16_t)index;
*instr = _Py_MAKECODEUNIT(LOAD_GLOBAL_MODULE, _Py_OPARG(*instr));
goto success;
}
if (!PyDict_CheckExact(builtins)) {
goto fail;
}
PyDictKeysObject * builtin_keys = ((PyDictObject *)builtins)->ma_keys;
index = _PyDictKeys_StringLookup(builtin_keys, name);
if (index == DKIX_ERROR) {
SPECIALIZATION_FAIL(LOAD_GLOBAL, SPEC_FAIL_LOAD_GLOBAL_NON_STRING_OR_SPLIT);
goto fail;
}
if (index != (uint16_t)index) {
goto fail;
}
uint32_t globals_version = _PyDictKeys_GetVersionForCurrentState(globals_keys);
if (globals_version == 0) {
SPECIALIZATION_FAIL(LOAD_GLOBAL, SPEC_FAIL_OUT_OF_VERSIONS);
goto fail;
}
uint32_t builtins_version = _PyDictKeys_GetVersionForCurrentState(builtin_keys);
if (builtins_version == 0) {
SPECIALIZATION_FAIL(LOAD_GLOBAL, SPEC_FAIL_OUT_OF_VERSIONS);
goto fail;
}
cache1->module_keys_version = globals_version;
cache1->builtin_keys_version = builtins_version;
cache0->index = (uint16_t)index;
*instr = _Py_MAKECODEUNIT(LOAD_GLOBAL_BUILTIN, _Py_OPARG(*instr));
goto success;
fail:
STAT_INC(LOAD_GLOBAL, failure);
assert(!PyErr_Occurred());
cache_backoff(cache0);
return 0;
success:
STAT_INC(LOAD_GLOBAL, success);
assert(!PyErr_Occurred());
cache0->counter = initial_counter_value();
return 0;
}
#ifdef Py_STATS
2021-08-27 08:01:22 -03:00
static int
binary_subscr_fail_kind(PyTypeObject *container_type, PyObject *sub)
2021-08-27 08:01:22 -03:00
{
if (container_type == &PyUnicode_Type) {
if (PyLong_CheckExact(sub)) {
return SPEC_FAIL_SUBSCR_STRING_INT;
2021-08-27 08:01:22 -03:00
}
if (PySlice_Check(sub)) {
return SPEC_FAIL_SUBSCR_STRING_SLICE;
2021-08-27 08:01:22 -03:00
}
return SPEC_FAIL_OTHER;
}
else if (strcmp(container_type->tp_name, "array.array") == 0) {
if (PyLong_CheckExact(sub)) {
return SPEC_FAIL_SUBSCR_ARRAY_INT;
2021-08-27 08:01:22 -03:00
}
if (PySlice_Check(sub)) {
return SPEC_FAIL_SUBSCR_ARRAY_SLICE;
2021-08-27 08:01:22 -03:00
}
return SPEC_FAIL_OTHER;
}
else if (container_type->tp_as_buffer) {
if (PyLong_CheckExact(sub)) {
return SPEC_FAIL_SUBSCR_BUFFER_INT;
2021-08-27 08:01:22 -03:00
}
if (PySlice_Check(sub)) {
return SPEC_FAIL_SUBSCR_BUFFER_SLICE;
2021-08-27 08:01:22 -03:00
}
return SPEC_FAIL_OTHER;
}
else if (container_type->tp_as_sequence) {
if (PyLong_CheckExact(sub) && container_type->tp_as_sequence->sq_item) {
return SPEC_FAIL_SUBSCR_SEQUENCE_INT;
2021-08-27 08:01:22 -03:00
}
}
return SPEC_FAIL_OTHER;
}
#endif
#define SIMPLE_FUNCTION 0
static int
function_kind(PyCodeObject *code) {
int flags = code->co_flags;
if ((flags & (CO_VARKEYWORDS | CO_VARARGS)) || code->co_kwonlyargcount) {
return SPEC_FAIL_CALL_COMPLEX_PARAMETERS;
}
if ((flags & CO_OPTIMIZED) == 0) {
return SPEC_FAIL_CALL_CO_NOT_OPTIMIZED;
}
return SIMPLE_FUNCTION;
}
int
_Py_Specialize_BinarySubscr(
PyObject *container, PyObject *sub, _Py_CODEUNIT *instr, SpecializedCacheEntry *cache)
{
_PyAdaptiveEntry *cache0 = &cache->adaptive;
PyTypeObject *container_type = Py_TYPE(container);
if (container_type == &PyList_Type) {
if (PyLong_CheckExact(sub)) {
*instr = _Py_MAKECODEUNIT(BINARY_SUBSCR_LIST_INT, _Py_OPARG(*instr));
goto success;
}
2021-08-27 08:01:22 -03:00
SPECIALIZATION_FAIL(BINARY_SUBSCR,
PySlice_Check(sub) ? SPEC_FAIL_SUBSCR_LIST_SLICE : SPEC_FAIL_OTHER);
2021-08-27 08:01:22 -03:00
goto fail;
}
if (container_type == &PyTuple_Type) {
if (PyLong_CheckExact(sub)) {
*instr = _Py_MAKECODEUNIT(BINARY_SUBSCR_TUPLE_INT, _Py_OPARG(*instr));
goto success;
}
2021-08-27 08:01:22 -03:00
SPECIALIZATION_FAIL(BINARY_SUBSCR,
PySlice_Check(sub) ? SPEC_FAIL_SUBSCR_TUPLE_SLICE : SPEC_FAIL_OTHER);
2021-08-27 08:01:22 -03:00
goto fail;
}
if (container_type == &PyDict_Type) {
*instr = _Py_MAKECODEUNIT(BINARY_SUBSCR_DICT, _Py_OPARG(*instr));
goto success;
}
PyTypeObject *cls = Py_TYPE(container);
bpo-46541: Replace core use of _Py_IDENTIFIER() with statically initialized global objects. (gh-30928) We're no longer using _Py_IDENTIFIER() (or _Py_static_string()) in any core CPython code. It is still used in a number of non-builtin stdlib modules. The replacement is: PyUnicodeObject (not pointer) fields under _PyRuntimeState, statically initialized as part of _PyRuntime. A new _Py_GET_GLOBAL_IDENTIFIER() macro facilitates lookup of the fields (along with _Py_GET_GLOBAL_STRING() for non-identifier strings). https://bugs.python.org/issue46541#msg411799 explains the rationale for this change. The core of the change is in: * (new) Include/internal/pycore_global_strings.h - the declarations for the global strings, along with the macros * Include/internal/pycore_runtime_init.h - added the static initializers for the global strings * Include/internal/pycore_global_objects.h - where the struct in pycore_global_strings.h is hooked into _PyRuntimeState * Tools/scripts/generate_global_objects.py - added generation of the global string declarations and static initializers I've also added a --check flag to generate_global_objects.py (along with make check-global-objects) to check for unused global strings. That check is added to the PR CI config. The remainder of this change updates the core code to use _Py_GET_GLOBAL_IDENTIFIER() instead of _Py_IDENTIFIER() and the related _Py*Id functions (likewise for _Py_GET_GLOBAL_STRING() instead of _Py_static_string()). This includes adding a few functions where there wasn't already an alternative to _Py*Id(), replacing the _Py_Identifier * parameter with PyObject *. The following are not changed (yet): * stop using _Py_IDENTIFIER() in the stdlib modules * (maybe) get rid of _Py_IDENTIFIER(), etc. entirely -- this may not be doable as at least one package on PyPI using this (private) API * (maybe) intern the strings during runtime init https://bugs.python.org/issue46541
2022-02-08 16:39:07 -04:00
PyObject *descriptor = _PyType_Lookup(cls, &_Py_ID(__getitem__));
if (descriptor && Py_TYPE(descriptor) == &PyFunction_Type) {
PyFunctionObject *func = (PyFunctionObject *)descriptor;
PyCodeObject *code = (PyCodeObject *)func->func_code;
int kind = function_kind(code);
if (kind != SIMPLE_FUNCTION) {
SPECIALIZATION_FAIL(BINARY_SUBSCR, kind);
goto fail;
}
if (code->co_argcount != 2) {
SPECIALIZATION_FAIL(BINARY_SUBSCR, SPEC_FAIL_WRONG_NUMBER_ARGUMENTS);
goto fail;
}
assert(cls->tp_version_tag != 0);
cache0->version = cls->tp_version_tag;
int version = _PyFunction_GetVersionForCurrentState(func);
if (version == 0 || version != (uint16_t)version) {
SPECIALIZATION_FAIL(BINARY_SUBSCR, SPEC_FAIL_OUT_OF_VERSIONS);
goto fail;
}
cache0->index = version;
cache[-1].obj.obj = descriptor;
*instr = _Py_MAKECODEUNIT(BINARY_SUBSCR_GETITEM, _Py_OPARG(*instr));
goto success;
}
2021-08-27 08:01:22 -03:00
SPECIALIZATION_FAIL(BINARY_SUBSCR,
binary_subscr_fail_kind(container_type, sub));
fail:
STAT_INC(BINARY_SUBSCR, failure);
assert(!PyErr_Occurred());
cache_backoff(cache0);
return 0;
success:
STAT_INC(BINARY_SUBSCR, success);
assert(!PyErr_Occurred());
cache0->counter = initial_counter_value();
return 0;
}
int
_Py_Specialize_StoreSubscr(PyObject *container, PyObject *sub, _Py_CODEUNIT *instr)
{
PyTypeObject *container_type = Py_TYPE(container);
if (container_type == &PyList_Type) {
if (PyLong_CheckExact(sub)) {
if ((Py_SIZE(sub) == 0 || Py_SIZE(sub) == 1)
&& ((PyLongObject *)sub)->ob_digit[0] < (size_t)PyList_GET_SIZE(container))
{
*instr = _Py_MAKECODEUNIT(STORE_SUBSCR_LIST_INT,
initial_counter_value());
goto success;
}
else {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_OUT_OF_RANGE);
goto fail;
}
}
else if (PySlice_Check(sub)) {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_SUBSCR_LIST_SLICE);
goto fail;
}
else {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_OTHER);
goto fail;
}
}
if (container_type == &PyDict_Type) {
*instr = _Py_MAKECODEUNIT(STORE_SUBSCR_DICT,
initial_counter_value());
goto success;
}
#ifdef Py_STATS
PyMappingMethods *as_mapping = container_type->tp_as_mapping;
if (as_mapping && (as_mapping->mp_ass_subscript
== PyDict_Type.tp_as_mapping->mp_ass_subscript)) {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_SUBSCR_DICT_SUBCLASS_NO_OVERRIDE);
goto fail;
}
if (PyObject_CheckBuffer(container)) {
if (PyLong_CheckExact(sub) && (((size_t)Py_SIZE(sub)) > 1)) {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_OUT_OF_RANGE);
}
else if (strcmp(container_type->tp_name, "array.array") == 0) {
if (PyLong_CheckExact(sub)) {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_SUBSCR_ARRAY_INT);
}
else if (PySlice_Check(sub)) {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_SUBSCR_ARRAY_SLICE);
}
else {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_OTHER);
}
}
else if (PyByteArray_CheckExact(container)) {
if (PyLong_CheckExact(sub)) {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_SUBSCR_BYTEARRAY_INT);
}
else if (PySlice_Check(sub)) {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_SUBSCR_BYTEARRAY_SLICE);
}
else {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_OTHER);
}
}
else {
if (PyLong_CheckExact(sub)) {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_SUBSCR_BUFFER_INT);
}
else if (PySlice_Check(sub)) {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_SUBSCR_BUFFER_SLICE);
}
else {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_OTHER);
}
}
goto fail;
}
bpo-46541: Replace core use of _Py_IDENTIFIER() with statically initialized global objects. (gh-30928) We're no longer using _Py_IDENTIFIER() (or _Py_static_string()) in any core CPython code. It is still used in a number of non-builtin stdlib modules. The replacement is: PyUnicodeObject (not pointer) fields under _PyRuntimeState, statically initialized as part of _PyRuntime. A new _Py_GET_GLOBAL_IDENTIFIER() macro facilitates lookup of the fields (along with _Py_GET_GLOBAL_STRING() for non-identifier strings). https://bugs.python.org/issue46541#msg411799 explains the rationale for this change. The core of the change is in: * (new) Include/internal/pycore_global_strings.h - the declarations for the global strings, along with the macros * Include/internal/pycore_runtime_init.h - added the static initializers for the global strings * Include/internal/pycore_global_objects.h - where the struct in pycore_global_strings.h is hooked into _PyRuntimeState * Tools/scripts/generate_global_objects.py - added generation of the global string declarations and static initializers I've also added a --check flag to generate_global_objects.py (along with make check-global-objects) to check for unused global strings. That check is added to the PR CI config. The remainder of this change updates the core code to use _Py_GET_GLOBAL_IDENTIFIER() instead of _Py_IDENTIFIER() and the related _Py*Id functions (likewise for _Py_GET_GLOBAL_STRING() instead of _Py_static_string()). This includes adding a few functions where there wasn't already an alternative to _Py*Id(), replacing the _Py_Identifier * parameter with PyObject *. The following are not changed (yet): * stop using _Py_IDENTIFIER() in the stdlib modules * (maybe) get rid of _Py_IDENTIFIER(), etc. entirely -- this may not be doable as at least one package on PyPI using this (private) API * (maybe) intern the strings during runtime init https://bugs.python.org/issue46541
2022-02-08 16:39:07 -04:00
PyObject *descriptor = _PyType_Lookup(container_type, &_Py_ID(__setitem__));
if (descriptor && Py_TYPE(descriptor) == &PyFunction_Type) {
PyFunctionObject *func = (PyFunctionObject *)descriptor;
PyCodeObject *code = (PyCodeObject *)func->func_code;
int kind = function_kind(code);
if (kind == SIMPLE_FUNCTION) {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_SUBSCR_PY_SIMPLE);
}
else {
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_SUBSCR_PY_OTHER);
}
goto fail;
}
#endif
SPECIALIZATION_FAIL(STORE_SUBSCR, SPEC_FAIL_OTHER);
fail:
STAT_INC(STORE_SUBSCR, failure);
assert(!PyErr_Occurred());
*instr = _Py_MAKECODEUNIT(_Py_OPCODE(*instr), ADAPTIVE_CACHE_BACKOFF);
return 0;
success:
STAT_INC(STORE_SUBSCR, success);
assert(!PyErr_Occurred());
return 0;
}
static int
specialize_class_call(
PyObject *callable, _Py_CODEUNIT *instr,
int nargs, PyObject *kwnames, SpecializedCacheEntry *cache)
{
PyTypeObject *tp = _PyType_CAST(callable);
if (tp->tp_new == PyBaseObject_Type.tp_new) {
SPECIALIZATION_FAIL(CALL, SPEC_FAIL_CALL_PYTHON_CLASS);
return -1;
}
if (tp->tp_flags & Py_TPFLAGS_IMMUTABLETYPE) {
if (nargs == 1 && kwnames == NULL) {
if (tp == &PyUnicode_Type) {
*instr = _Py_MAKECODEUNIT(CALL_NO_KW_STR_1, _Py_OPARG(*instr));
return 0;
}
else if (tp == &PyType_Type) {
*instr = _Py_MAKECODEUNIT(CALL_NO_KW_TYPE_1, _Py_OPARG(*instr));
return 0;
}
else if (tp == &PyTuple_Type) {
*instr = _Py_MAKECODEUNIT(CALL_NO_KW_TUPLE_1, _Py_OPARG(*instr));
return 0;
}
}
if (tp->tp_vectorcall != NULL) {
*instr = _Py_MAKECODEUNIT(CALL_BUILTIN_CLASS, _Py_OPARG(*instr));
return 0;
}
SPECIALIZATION_FAIL(CALL, tp == &PyUnicode_Type ?
SPEC_FAIL_CALL_STR : SPEC_FAIL_CALL_CLASS_NO_VECTORCALL);
return -1;
}
SPECIALIZATION_FAIL(CALL, SPEC_FAIL_CALL_CLASS_MUTABLE);
return -1;
}
#ifdef Py_STATS
static int
builtin_call_fail_kind(int ml_flags)
{
switch (ml_flags & (METH_VARARGS | METH_FASTCALL | METH_NOARGS | METH_O |
METH_KEYWORDS | METH_METHOD)) {
case METH_VARARGS:
return SPEC_FAIL_CALL_PYCFUNCTION;
case METH_VARARGS | METH_KEYWORDS:
return SPEC_FAIL_CALL_PYCFUNCTION_WITH_KEYWORDS;
case METH_FASTCALL | METH_KEYWORDS:
return SPEC_FAIL_CALL_PYCFUNCTION_FAST_WITH_KEYWORDS;
case METH_NOARGS:
return SPEC_FAIL_CALL_PYCFUNCTION_NOARGS;
/* This case should never happen with PyCFunctionObject -- only
PyMethodObject. See zlib.compressobj()'s methods for an example.
*/
case METH_METHOD | METH_FASTCALL | METH_KEYWORDS:
default:
return SPEC_FAIL_CALL_BAD_CALL_FLAGS;
}
}
#endif
static PyMethodDescrObject *_list_append = NULL;
static int
specialize_method_descriptor(
PyMethodDescrObject *descr, _Py_CODEUNIT *instr,
int nargs, PyObject *kwnames, SpecializedCacheEntry *cache)
{
if (kwnames) {
SPECIALIZATION_FAIL(CALL, SPEC_FAIL_CALL_KWNAMES);
return -1;
}
if (_list_append == NULL) {
bpo-46541: Replace core use of _Py_IDENTIFIER() with statically initialized global objects. (gh-30928) We're no longer using _Py_IDENTIFIER() (or _Py_static_string()) in any core CPython code. It is still used in a number of non-builtin stdlib modules. The replacement is: PyUnicodeObject (not pointer) fields under _PyRuntimeState, statically initialized as part of _PyRuntime. A new _Py_GET_GLOBAL_IDENTIFIER() macro facilitates lookup of the fields (along with _Py_GET_GLOBAL_STRING() for non-identifier strings). https://bugs.python.org/issue46541#msg411799 explains the rationale for this change. The core of the change is in: * (new) Include/internal/pycore_global_strings.h - the declarations for the global strings, along with the macros * Include/internal/pycore_runtime_init.h - added the static initializers for the global strings * Include/internal/pycore_global_objects.h - where the struct in pycore_global_strings.h is hooked into _PyRuntimeState * Tools/scripts/generate_global_objects.py - added generation of the global string declarations and static initializers I've also added a --check flag to generate_global_objects.py (along with make check-global-objects) to check for unused global strings. That check is added to the PR CI config. The remainder of this change updates the core code to use _Py_GET_GLOBAL_IDENTIFIER() instead of _Py_IDENTIFIER() and the related _Py*Id functions (likewise for _Py_GET_GLOBAL_STRING() instead of _Py_static_string()). This includes adding a few functions where there wasn't already an alternative to _Py*Id(), replacing the _Py_Identifier * parameter with PyObject *. The following are not changed (yet): * stop using _Py_IDENTIFIER() in the stdlib modules * (maybe) get rid of _Py_IDENTIFIER(), etc. entirely -- this may not be doable as at least one package on PyPI using this (private) API * (maybe) intern the strings during runtime init https://bugs.python.org/issue46541
2022-02-08 16:39:07 -04:00
_list_append = (PyMethodDescrObject *)_PyType_Lookup(&PyList_Type,
&_Py_ID(append));
}
assert(_list_append != NULL);
if (nargs == 2 && descr == _list_append) {
assert(_Py_OPCODE(instr[-1]) == PRECALL_METHOD);
cache[-1].obj.obj = (PyObject *)_list_append;
*instr = _Py_MAKECODEUNIT(CALL_NO_KW_LIST_APPEND, _Py_OPARG(*instr));
return 0;
}
switch (descr->d_method->ml_flags &
(METH_VARARGS | METH_FASTCALL | METH_NOARGS | METH_O |
METH_KEYWORDS | METH_METHOD)) {
case METH_NOARGS: {
if (nargs != 1) {
SPECIALIZATION_FAIL(CALL, SPEC_FAIL_WRONG_NUMBER_ARGUMENTS);
return -1;
}
*instr = _Py_MAKECODEUNIT(CALL_NO_KW_METHOD_DESCRIPTOR_NOARGS,
_Py_OPARG(*instr));
return 0;
}
case METH_O: {
if (nargs != 2) {
SPECIALIZATION_FAIL(CALL, SPEC_FAIL_OUT_OF_RANGE);
return -1;
}
*instr = _Py_MAKECODEUNIT(CALL_NO_KW_METHOD_DESCRIPTOR_O,
_Py_OPARG(*instr));
return 0;
}
case METH_FASTCALL: {
*instr = _Py_MAKECODEUNIT(CALL_NO_KW_METHOD_DESCRIPTOR_FAST,
_Py_OPARG(*instr));
return 0;
}
}
SPECIALIZATION_FAIL(CALL, builtin_call_fail_kind(descr->d_method->ml_flags));
return -1;
}
static int
specialize_py_call(
PyFunctionObject *func, _Py_CODEUNIT *instr,
int nargs, PyObject *kwnames, SpecializedCacheEntry *cache)
{
_PyCallCache *cache1 = &cache[-1].call;
PyCodeObject *code = (PyCodeObject *)func->func_code;
int kind = function_kind(code);
if (kwnames) {
SPECIALIZATION_FAIL(CALL, SPEC_FAIL_CALL_KWNAMES);
return -1;
}
if (kind != SIMPLE_FUNCTION) {
SPECIALIZATION_FAIL(CALL, kind);
return -1;
}
int argcount = code->co_argcount;
if (argcount > 0xffff) {
SPECIALIZATION_FAIL(CALL, SPEC_FAIL_OUT_OF_RANGE);
return -1;
}
int defcount = func->func_defaults == NULL ? 0 : (int)PyTuple_GET_SIZE(func->func_defaults);
assert(defcount <= argcount);
int min_args = argcount-defcount;
if (nargs > argcount || nargs < min_args) {
SPECIALIZATION_FAIL(CALL, SPEC_FAIL_WRONG_NUMBER_ARGUMENTS);
return -1;
}
assert(nargs <= argcount && nargs >= min_args);
assert(min_args >= 0 && defcount >= 0);
assert(defcount == 0 || func->func_defaults != NULL);
if (min_args > 0xffff || defcount > 0xffff) {
SPECIALIZATION_FAIL(CALL, SPEC_FAIL_OUT_OF_RANGE);
return -1;
}
int version = _PyFunction_GetVersionForCurrentState(func);
if (version == 0) {
SPECIALIZATION_FAIL(CALL, SPEC_FAIL_OUT_OF_VERSIONS);
return -1;
}
cache[0].adaptive.index = nargs;
cache1->func_version = version;
cache1->min_args = min_args;
cache1->defaults_len = defcount;
if (argcount == nargs) {
*instr = _Py_MAKECODEUNIT(CALL_PY_EXACT_ARGS, _Py_OPARG(*instr));
}
else {
*instr = _Py_MAKECODEUNIT(CALL_PY_WITH_DEFAULTS, _Py_OPARG(*instr));
}
return 0;
}
static int
specialize_c_call(PyObject *callable, _Py_CODEUNIT *instr, int nargs,
PyObject *kwnames, SpecializedCacheEntry *cache, PyObject *builtins)
{
_PyObjectCache *cache1 = &cache[-1].obj;
if (PyCFunction_GET_FUNCTION(callable) == NULL) {
return 1;
}
switch (PyCFunction_GET_FLAGS(callable) &
(METH_VARARGS | METH_FASTCALL | METH_NOARGS | METH_O |
METH_KEYWORDS | METH_METHOD)) {
case METH_O: {
if (kwnames) {
SPECIALIZATION_FAIL(CALL, SPEC_FAIL_CALL_KWNAMES);
return -1;
}
if (nargs != 1) {
SPECIALIZATION_FAIL(CALL, SPEC_FAIL_WRONG_NUMBER_ARGUMENTS);
return 1;
}
/* len(o) */
PyObject *builtin_len = PyDict_GetItemString(builtins, "len");
if (callable == builtin_len) {
cache1->obj = builtin_len; // borrowed
*instr = _Py_MAKECODEUNIT(CALL_NO_KW_LEN,
_Py_OPARG(*instr));
return 0;
}
*instr = _Py_MAKECODEUNIT(CALL_NO_KW_BUILTIN_O,
_Py_OPARG(*instr));
return 0;
}
case METH_FASTCALL: {
if (kwnames) {
SPECIALIZATION_FAIL(CALL, SPEC_FAIL_CALL_KWNAMES);
return -1;
}
if (nargs == 2) {
/* isinstance(o1, o2) */
PyObject *builtin_isinstance = PyDict_GetItemString(
builtins, "isinstance");
if (callable == builtin_isinstance) {
cache1->obj = builtin_isinstance; // borrowed
*instr = _Py_MAKECODEUNIT(CALL_NO_KW_ISINSTANCE,
_Py_OPARG(*instr));
return 0;
}
}
*instr = _Py_MAKECODEUNIT(CALL_NO_KW_BUILTIN_FAST,
_Py_OPARG(*instr));
return 0;
}
case METH_FASTCALL | METH_KEYWORDS: {
*instr = _Py_MAKECODEUNIT(CALL_BUILTIN_FAST_WITH_KEYWORDS,
_Py_OPARG(*instr));
return 0;
}
default:
SPECIALIZATION_FAIL(CALL,
builtin_call_fail_kind(PyCFunction_GET_FLAGS(callable)));
return 1;
}
}
#ifdef Py_STATS
static int
call_fail_kind(PyObject *callable)
{
if (PyCFunction_CheckExact(callable)) {
return SPEC_FAIL_CALL_PYCFUNCTION;
}
else if (PyFunction_Check(callable)) {
return SPEC_FAIL_CALL_PYFUNCTION;
}
else if (PyInstanceMethod_Check(callable)) {
return SPEC_FAIL_CALL_INSTANCE_METHOD;
}
else if (PyMethod_Check(callable)) {
return SPEC_FAIL_CALL_BOUND_METHOD;
}
// builtin method
else if (PyCMethod_Check(callable)) {
return SPEC_FAIL_CALL_CMETHOD;
}
else if (PyType_Check(callable)) {
if (((PyTypeObject *)callable)->tp_new == PyBaseObject_Type.tp_new) {
return SPEC_FAIL_CALL_PYTHON_CLASS;
}
else {
return SPEC_FAIL_CALL_CLASS;
}
}
else if (Py_IS_TYPE(callable, &PyMethodDescr_Type)) {
return SPEC_FAIL_CALL_METHOD_DESCRIPTOR;
}
else if (Py_TYPE(callable) == &PyWrapperDescr_Type) {
return SPEC_FAIL_CALL_OPERATOR_WRAPPER;
}
else if (Py_TYPE(callable) == &_PyMethodWrapper_Type) {
return SPEC_FAIL_CALL_METHOD_WRAPPER;
}
return SPEC_FAIL_OTHER;
}
#endif
/* TODO:
- Specialize calling classes.
*/
int
_Py_Specialize_CallNoKw(
PyObject *callable, _Py_CODEUNIT *instr,
int nargs, PyObject *kwnames,
SpecializedCacheEntry *cache, PyObject *builtins)
{
_PyAdaptiveEntry *cache0 = &cache->adaptive;
int fail;
if (PyCFunction_CheckExact(callable)) {
fail = specialize_c_call(callable, instr, nargs, kwnames, cache, builtins);
}
else if (PyFunction_Check(callable)) {
fail = specialize_py_call((PyFunctionObject *)callable, instr, nargs, kwnames, cache);
}
else if (PyType_Check(callable)) {
fail = specialize_class_call(callable, instr, nargs, kwnames, cache);
}
else if (Py_IS_TYPE(callable, &PyMethodDescr_Type)) {
fail = specialize_method_descriptor(
(PyMethodDescrObject *)callable, instr, nargs, kwnames, cache);
}
else {
SPECIALIZATION_FAIL(CALL, call_fail_kind(callable));
fail = -1;
}
if (fail) {
STAT_INC(CALL, failure);
assert(!PyErr_Occurred());
cache_backoff(cache0);
}
else {
STAT_INC(CALL, success);
assert(!PyErr_Occurred());
cache0->counter = initial_counter_value();
}
return 0;
}
void
_Py_Specialize_BinaryOp(PyObject *lhs, PyObject *rhs, _Py_CODEUNIT *instr,
SpecializedCacheEntry *cache)
{
_PyAdaptiveEntry *adaptive = &cache->adaptive;
switch (adaptive->original_oparg) {
case NB_ADD:
case NB_INPLACE_ADD:
if (!Py_IS_TYPE(lhs, Py_TYPE(rhs))) {
SPECIALIZATION_FAIL(BINARY_OP, SPEC_FAIL_BINARY_OP_DIFFERENT_TYPES);
goto failure;
}
if (PyUnicode_CheckExact(lhs)) {
if (_Py_OPCODE(instr[1]) == STORE_FAST && Py_REFCNT(lhs) == 2) {
*instr = _Py_MAKECODEUNIT(BINARY_OP_INPLACE_ADD_UNICODE,
_Py_OPARG(*instr));
goto success;
}
*instr = _Py_MAKECODEUNIT(BINARY_OP_ADD_UNICODE,
_Py_OPARG(*instr));
goto success;
}
if (PyLong_CheckExact(lhs)) {
*instr = _Py_MAKECODEUNIT(BINARY_OP_ADD_INT, _Py_OPARG(*instr));
goto success;
}
if (PyFloat_CheckExact(lhs)) {
*instr = _Py_MAKECODEUNIT(BINARY_OP_ADD_FLOAT,
_Py_OPARG(*instr));
goto success;
}
break;
case NB_MULTIPLY:
case NB_INPLACE_MULTIPLY:
if (!Py_IS_TYPE(lhs, Py_TYPE(rhs))) {
SPECIALIZATION_FAIL(BINARY_OP, SPEC_FAIL_BINARY_OP_DIFFERENT_TYPES);
goto failure;
}
if (PyLong_CheckExact(lhs)) {
*instr = _Py_MAKECODEUNIT(BINARY_OP_MULTIPLY_INT,
_Py_OPARG(*instr));
goto success;
}
if (PyFloat_CheckExact(lhs)) {
*instr = _Py_MAKECODEUNIT(BINARY_OP_MULTIPLY_FLOAT,
_Py_OPARG(*instr));
goto success;
}
break;
case NB_SUBTRACT:
case NB_INPLACE_SUBTRACT:
if (!Py_IS_TYPE(lhs, Py_TYPE(rhs))) {
SPECIALIZATION_FAIL(BINARY_OP, SPEC_FAIL_BINARY_OP_DIFFERENT_TYPES);
goto failure;
}
if (PyLong_CheckExact(lhs)) {
*instr = _Py_MAKECODEUNIT(BINARY_OP_SUBTRACT_INT,
_Py_OPARG(*instr));
goto success;
}
if (PyFloat_CheckExact(lhs)) {
*instr = _Py_MAKECODEUNIT(BINARY_OP_SUBTRACT_FLOAT,
_Py_OPARG(*instr));
goto success;
}
break;
default:
// These operators don't have any available specializations. Rather
// than repeatedly attempting to specialize them, just convert them
// back to BINARY_OP (while still recording a failure, of course)!
*instr = _Py_MAKECODEUNIT(BINARY_OP, adaptive->original_oparg);
}
SPECIALIZATION_FAIL(BINARY_OP, SPEC_FAIL_OTHER);
failure:
STAT_INC(BINARY_OP, failure);
cache_backoff(adaptive);
return;
success:
STAT_INC(BINARY_OP, success);
adaptive->counter = initial_counter_value();
}
#ifdef Py_STATS
static int
compare_op_fail_kind(PyObject *lhs, PyObject *rhs)
{
if (Py_TYPE(lhs) != Py_TYPE(rhs)) {
if (PyFloat_CheckExact(lhs) && PyLong_CheckExact(rhs)) {
return SPEC_FAIL_COMPARE_OP_FLOAT_LONG;
}
if (PyLong_CheckExact(lhs) && PyFloat_CheckExact(rhs)) {
return SPEC_FAIL_COMPARE_OP_LONG_FLOAT;
}
return SPEC_FAIL_COMPARE_OP_DIFFERENT_TYPES;
}
if (PyBytes_CheckExact(lhs)) {
return SPEC_FAIL_COMPARE_OP_BYTES;
}
if (PyTuple_CheckExact(lhs)) {
return SPEC_FAIL_COMPARE_OP_TUPLE;
}
if (PyList_CheckExact(lhs)) {
return SPEC_FAIL_COMPARE_OP_LIST;
}
if (PySet_CheckExact(lhs) || PyFrozenSet_CheckExact(lhs)) {
return SPEC_FAIL_COMPARE_OP_SET;
}
if (PyBool_Check(lhs)) {
return SPEC_FAIL_COMPARE_OP_BOOL;
}
if (Py_TYPE(lhs)->tp_richcompare == PyBaseObject_Type.tp_richcompare) {
return SPEC_FAIL_COMPARE_OP_BASEOBJECT;
}
return SPEC_FAIL_OTHER;
}
#endif
static int compare_masks[] = {
// 1-bit: jump if less than
// 2-bit: jump if equal
// 4-bit: jump if greater
[Py_LT] = 1 | 0 | 0,
[Py_LE] = 1 | 2 | 0,
[Py_EQ] = 0 | 2 | 0,
[Py_NE] = 1 | 0 | 4,
[Py_GT] = 0 | 0 | 4,
[Py_GE] = 0 | 2 | 4,
};
void
_Py_Specialize_CompareOp(PyObject *lhs, PyObject *rhs,
_Py_CODEUNIT *instr, SpecializedCacheEntry *cache)
{
_PyAdaptiveEntry *adaptive = &cache->adaptive;
int op = adaptive->original_oparg;
int next_opcode = _Py_OPCODE(instr[1]);
if (next_opcode != POP_JUMP_IF_FALSE && next_opcode != POP_JUMP_IF_TRUE) {
// Can't ever combine, so don't don't bother being adaptive.
SPECIALIZATION_FAIL(COMPARE_OP, SPEC_FAIL_COMPARE_OP_NOT_FOLLOWED_BY_COND_JUMP);
*instr = _Py_MAKECODEUNIT(COMPARE_OP, adaptive->original_oparg);
goto failure;
}
assert(op <= Py_GE);
int when_to_jump_mask = compare_masks[op];
if (next_opcode == POP_JUMP_IF_FALSE) {
when_to_jump_mask = (1 | 2 | 4) & ~when_to_jump_mask;
}
if (Py_TYPE(lhs) != Py_TYPE(rhs)) {
SPECIALIZATION_FAIL(COMPARE_OP, compare_op_fail_kind(lhs, rhs));
goto failure;
}
if (PyFloat_CheckExact(lhs)) {
*instr = _Py_MAKECODEUNIT(COMPARE_OP_FLOAT_JUMP, _Py_OPARG(*instr));
adaptive->index = when_to_jump_mask;
goto success;
}
if (PyLong_CheckExact(lhs)) {
if (Py_ABS(Py_SIZE(lhs)) <= 1 && Py_ABS(Py_SIZE(rhs)) <= 1) {
*instr = _Py_MAKECODEUNIT(COMPARE_OP_INT_JUMP, _Py_OPARG(*instr));
adaptive->index = when_to_jump_mask;
goto success;
}
else {
SPECIALIZATION_FAIL(COMPARE_OP, SPEC_FAIL_COMPARE_OP_BIG_INT);
goto failure;
}
}
if (PyUnicode_CheckExact(lhs)) {
if (op != Py_EQ && op != Py_NE) {
SPECIALIZATION_FAIL(COMPARE_OP, SPEC_FAIL_COMPARE_OP_STRING);
goto failure;
}
else {
*instr = _Py_MAKECODEUNIT(COMPARE_OP_STR_JUMP, _Py_OPARG(*instr));
adaptive->index = (when_to_jump_mask & 2) == 0;
goto success;
}
}
SPECIALIZATION_FAIL(COMPARE_OP, compare_op_fail_kind(lhs, rhs));
failure:
STAT_INC(COMPARE_OP, failure);
cache_backoff(adaptive);
return;
success:
STAT_INC(COMPARE_OP, success);
adaptive->counter = initial_counter_value();
}
#ifdef Py_STATS
int
_PySpecialization_ClassifyIterator(PyObject *iter)
{
if (PyGen_CheckExact(iter)) {
return SPEC_FAIL_FOR_ITER_GENERATOR;
}
if (PyCoro_CheckExact(iter)) {
return SPEC_FAIL_FOR_ITER_COROUTINE;
}
if (PyAsyncGen_CheckExact(iter)) {
return SPEC_FAIL_FOR_ITER_ASYNC_GENERATOR;
}
PyTypeObject *t = Py_TYPE(iter);
if (t == &PyListIter_Type) {
return SPEC_FAIL_FOR_ITER_LIST;
}
if (t == &PyTupleIter_Type) {
return SPEC_FAIL_FOR_ITER_TUPLE;
}
if (t == &PyDictIterKey_Type) {
return SPEC_FAIL_FOR_ITER_DICT_KEYS;
}
if (t == &PyDictIterValue_Type) {
return SPEC_FAIL_FOR_ITER_DICT_VALUES;
}
if (t == &PyDictIterItem_Type) {
return SPEC_FAIL_FOR_ITER_DICT_ITEMS;
}
if (t == &PySetIter_Type) {
return SPEC_FAIL_FOR_ITER_SET;
}
if (t == &PyUnicodeIter_Type) {
return SPEC_FAIL_FOR_ITER_STRING;
}
if (t == &PyBytesIter_Type) {
return SPEC_FAIL_FOR_ITER_BYTES;
}
if (t == &PyRangeIter_Type) {
return SPEC_FAIL_FOR_ITER_RANGE;
}
if (t == &PyEnum_Type) {
return SPEC_FAIL_FOR_ITER_ENUMERATE;
}
if (strncmp(t->tp_name, "itertools", 8) == 0) {
return SPEC_FAIL_FOR_ITER_ITERTOOLS;
}
return SPEC_FAIL_OTHER;
}
int
_PySpecialization_ClassifySequence(PyObject *seq, int n)
{
assert(n >= 0);
if (n > 4) {
n = 5;
}
if (PyTuple_CheckExact(seq)) {
return SPEC_FAIL_UNPACK_SEQUENCE_TUPLE_0 + n;
}
if (PyList_CheckExact(seq)) {
return SPEC_FAIL_UNPACK_SEQUENCE_LIST_0 + n;
}
return SPEC_FAIL_UNPACK_SEQUENCE_OTHER_0 + n;
}
int
_PySpecialization_ClassifyCallable(PyObject *callable)
{
return call_fail_kind(callable);
}
#endif