- The `dump_stack()` method could call a `__repr__` method implemented in Python,
causing (infinite) recursion.
I rewrote it to only print out the values for some fundamental types (`int`, `str`, etc.);
for everything else it just prints `<type_name @ 0xdeadbeef>`.
- The lltrace-like feature for uops wrote to `stderr`, while the one in `ceval.c` writes to `stdout`;
I changed the uops to write to stdout as well.
Introducing a new file, stacking.py, that takes over several responsibilities related to symbolic evaluation of push/pop operations, with more generality.
The linked list of objects was a global variable, which broke isolation between interpreters, causing crashes. To solve this, we've moved the linked list to each interpreter.
gh-107184 introduced a refleak in test_import.SubinterpImportTests (specifically test_singlephase_check_with_setting_and_override and test_single_init_extension_compat). We fix it here by making sure _testsinglephase is removed from sys.modules whenever we clear the runtime's internal state for the module.
The underlying problem is strictly contained in the internal function _PyImport_ClearExtension() (AKA _testinternalcapi.clear_extension()), which is only used in tests.
(This also fixes an intermittent segfault introduced in the same place, in test_disallowed_reimport.)
There's no need to use a dummy uop to skip unused cache entries. The macro syntax lets you write `unused/1` instead.
Similarly, move `unused/5` from op `_LOAD_ATTR_INSTANCE_VALUE` to macro `LOAD_ATTR_INSTANCE_VALUE`.
This fixes a crasher due to a race condition, triggered infrequently when two isolated (own GIL) subinterpreters simultaneously initialize their sys or builtins modules. The crash happened due the combination of the "detached" thread state we were using and the "last holder" logic we use for the GIL. It turns out it's tricky to use the same thread state for different threads. Who could have guessed?
We solve the problem by eliminating the one object we were still sharing between interpreters. We replace it with a low-level hashtable, using the "raw" allocator to avoid tying it to the main interpreter.
We also remove the accommodations for "detached" thread states, which were a dubious idea to start with.
The _xxsubinterpreters module should not rely on internal API. Some of the functions it uses were recently moved there however. Here we move them back (and expose them properly).
We tried this before with a dict and for all interned strings. That ran into problems due to interpreter isolation. However, exclusively using a per-interpreter cache caused some inconsistency that can eliminate the benefit of interning. Here we circle back to using a global cache, but only for statically allocated strings. We also use a more-basic _Py_hashtable_t for that global cache instead of a dict.
Ideally we would only have the global cache, but the optional isolation of each interpreter's allocator means that a non-static string object must not outlive its interpreter. Thus we would have to store a copy of each such interned string in the global cache, tied to the main interpreter.
Move private _PyDict functions to the internal C API (pycore_dict.h):
* _PyDict_Contains_KnownHash()
* _PyDict_DebugMallocStats()
* _PyDict_DelItemIf()
* _PyDict_GetItemWithError()
* _PyDict_HasOnlyStringKeys()
* _PyDict_MaybeUntrack()
* _PyDict_MergeEx()
No longer export these functions.
Move private debug _PyObject functions to the internal C API
(pycore_object.h):
* _PyDebugAllocatorStats()
* _PyObject_CheckConsistency()
* _PyObject_DebugTypeStats()
* _PyObject_IsFreed()
No longer export most of these functions, except of
_PyObject_IsFreed().
Move test functions using _PyObject_IsFreed() from _testcapi to
_testinternalcapi. check_pyobject_is_freed() test no longer catch
_testcapi.error: the tested function cannot raise _testcapi.error.
Rename private C API constants:
* Rename PY_MONITORING_UNGROUPED_EVENTS to _PY_MONITORING_UNGROUPED_EVENTS
* Rename PY_MONITORING_EVENTS to _PY_MONITORING_EVENTS
* No longer export most private _PyHash symbols, only export the ones
which are needed by shared extensions.
* Modules/_xxtestfuzz/fuzzer.c now uses the internal C API.
By turning `assert(kwnames == NULL)` into a macro that is not in the "forbidden" list, many instructions that formerly were skipped because they contained such an assert (but no other mention of `kwnames`) are now supported in Tier 2. This covers 10 instructions in total (all specializations of `CALL` that invoke some C code):
- `CALL_NO_KW_TYPE_1`
- `CALL_NO_KW_STR_1`
- `CALL_NO_KW_TUPLE_1`
- `CALL_NO_KW_BUILTIN_O`
- `CALL_NO_KW_BUILTIN_FAST`
- `CALL_NO_KW_LEN`
- `CALL_NO_KW_ISINSTANCE`
- `CALL_NO_KW_METHOD_DESCRIPTOR_O`
- `CALL_NO_KW_METHOD_DESCRIPTOR_NOARGS`
- `CALL_NO_KW_METHOD_DESCRIPTOR_FAST`
These aren't automatically translated because (ironically)
they are macros deferring to POP_JUMP_IF_{TRUE,FALSE},
which are not viable uops (being manually translated).
The hack is that we emit IS_NONE and then set opcode and
jump to the POP_JUMP_IF_{TRUE,FALSE} translation code.
The Tier 2 opcode _IS_ITER_EXHAUSTED_LIST (and _TUPLE)
didn't set it->it_seq to NULL, causing a subtle bug
that resulted in test_exhausted_iterator in list_tests.py
to fail when running all tests with -Xuops.
The bug was introduced in gh-106696.
Added this as an explicit test.
Also fixed the dependencies for ceval.o -- it depends on executor_cases.c.h.
This moves EXIT_TRACE, SAVE_IP, JUMP_TO_TOP, and
_POP_JUMP_IF_{FALSE,TRUE} from ceval.c to bytecodes.c.
They are no less special than before, but this way
they are discoverable o the copy-and-patch tooling.
During superblock generation, a JUMP_BACKWARD instruction is translated to either a JUMP_TO_TOP micro-op (when the target of the jump is exactly the beginning of the superblock, closing the loop), or a SAVE_IP + EXIT_TRACE pair, when the jump goes elsewhere.
The new JUMP_TO_TOP instruction includes a CHECK_EVAL_BREAKER() call, so a closed loop can still be interrupted.
* Convert PyObject_DelAttr() and PyObject_DelAttrString() macros to
functions.
* Add PyObject_DelAttr() and PyObject_DelAttrString() functions to
the stable ABI.
* Replace PyObject_SetAttr(obj, name, NULL) with
PyObject_DelAttr(obj, name).
- Hand-written uops JUMP_IF_{TRUE,FALSE}.
These peek at the top of the stack.
The jump target (in superblock space) is absolute.
- Hand-written translation for POP_JUMP_IF_{TRUE,FALSE},
assuming the jump is unlikely.
Once we implement jump-likelihood profiling,
we can implement the jump-unlikely case (in another PR).
- Tests (including some test cleanup).
- Improvements to len(ex) and ex[i] to expose the whole trace.
This adds several of unspecialized opcodes to superblocks:
TO_BOOL, BINARY_SUBSCR, STORE_SUBSCR,
UNPACK_SEQUENCE, LOAD_GLOBAL, LOAD_ATTR,
COMPARE_OP, BINARY_OP.
While we may not want that eventually, for now this helps finding bugs.
There is a rudimentary test checking for UNPACK_SEQUENCE.
Once we're ready to undo this, that would be simple:
just replace the call to variable_used_unspecialized
with a call to variable_used (as shown in a comment).
Or add individual opcdes to FORBIDDEN_NAMES_IN_UOPS.
Instead of special-casing specific instructions,
we add a few more special values to the 'size' field of expansions,
so in the future we can automatically handle
additional super-instructions in the generator.
The uops test wasn't testing anything by default,
and was failing when run with -Xuops.
Made the two executor-related context managers global,
so TestUops can use them (notably `with temporary_optimizer(opt)`).
Made clear_executor() a little more thorough.
Fixed a crash upon finalizing a uop optimizer,
by adding a `tp_dealloc` handler.
When `_PyOptimizer_BackEdge` returns `NULL`, we should restore `next_instr` (and `stack_pointer`). To accomplish this we should jump to `resume_with_error` instead of just `error`.
The problem this causes is subtle -- the only repro I have is in PR gh-106393, at commit d7df54b139bcc47f5ea094bfaa9824f79bc45adc. But the fix is real (as shown later in that PR).
While we're at it, also improve the debug output: the offsets at which traces are identified are now measured in bytes, and always show the start offset. This makes it easier to correlate executor calls with optimizer calls, and either with `dis` output.
<!-- gh-issue-number: gh-104584 -->
* Issue: gh-104584
<!-- /gh-issue-number -->
Remove private pylifecycle.h functions: move them to the internal C
API ( pycore_atexit.h, pycore_pylifecycle.h and pycore_signal.h). No
longer export most of these functions.
Move _testcapi.test_atexit() to _testinternalcapi.
Remove private _PyUnicode_TransformDecimalAndSpaceToASCII() and other
private _PyUnicode C API functions: move them to the internal C API
(pycore_unicodeobject.h). No longer most of these functions.
Replace _testcapi.unicode_transformdecimalandspacetoascii() with
_testinternal._PyUnicode_TransformDecimalAndSpaceToASCII().
Remove more private _PyUnicode C API functions:
move them to the internal C API (pycore_unicodeobject.h).
No longer export most pycore_unicodeobject.h functions.
- Tweak uops debugging output
- Fix the bug from gh-106290
- Rename `SET_IP` to `SAVE_IP` (per https://github.com/faster-cpython/ideas/issues/558)
- Add a `SAVE_IP` uop at the start of the trace (ditto)
- Allow `unbound_local_error`; this gives us uops for `LOAD_FAST_CHECK`, `LOAD_CLOSURE`, and `DELETE_FAST`
- Longer traces
- Support `STORE_FAST_LOAD_FAST`, `STORE_FAST_STORE_FAST`
- Add deps on pycore_uops.h to Makefile(.pre.in)
Remove the following functions from the C API, move them to the internal C
API: add a new pycore_modsupport.h internal header file:
* PyModule_CreateInitialized()
* _PyArg_NoKwnames()
* _Py_VaBuildStack()
No longer export these functions.
Remove private _PyThreadState and _PyInterpreterState C API
functions: move them to the internal C API (pycore_pystate.h and
pycore_interp.h). Don't export most of these functions anymore, but
still export functions used by tests.
Remove _PyThreadState_Prealloc() and _PyThreadState_Init() from the C
API, but keep it in the stable API.
Remove the "cpython/pytime.h" header file: it only contained private
functions. Move functions to the internal pycore_time.h header file.
Move tests from _testcapi to _testinternalcapi. Rename also test
methods to have the same name than tested C functions.
No longer export these functions:
* _PyTime_Add()
* _PyTime_As100Nanoseconds()
* _PyTime_FromMicrosecondsClamp()
* _PyTime_FromTimespec()
* _PyTime_FromTimeval()
* _PyTime_GetPerfCounterWithInfo()
* _PyTime_MulDiv()
Remove the following private functions of the C API:
* _PyCodecInfo_GetIncrementalDecoder()
* _PyCodecInfo_GetIncrementalEncoder()
* _PyCodec_DecodeText()
* _PyCodec_EncodeText()
* _PyCodec_Forget()
* _PyCodec_Lookup()
* _PyCodec_LookupTextEncoding()
Move these functions to a new pycore_codecs.h internal header file.
These functions are no longer exported.
* EOFError no longer overrides other errors such as MemoryError or OSError at
the start of the object.
* Raise more relevant error when the NULL object occurs as a code object
component.
* Minimize an overhead of calling PyErr_Occurred().
This produces longer traces (superblocks?).
Also improved debug output (uop names are now printed instead of numeric opcodes). This would be simpler if the numeric opcode values were generated by generate_cases.py, but that's another project.
Refactored some code in generate_cases.py so the essential algorithm for cache effects is only run once. (Deciding which effects are used and what the total cache size is, regardless of what's used.)
Remove the following private functions from the public C API:
* _Py_CheckFunctionResult()
* _PyObject_CallMethod()
* _PyObject_CallMethodId()
* _PyObject_CallMethodIdNoArgs()
* _PyObject_CallMethodIdObjArgs()
* _PyObject_CallMethodIdOneArg()
* _PyObject_MakeTpCall()
* _PyObject_VectorcallMethodId()
* _PyStack_AsDict()
Move these functions to the internal C API (pycore_call.h).
No longer export the following functions:
* _PyObject_Call()
* _PyObject_CallMethod()
* _PyObject_CallMethodId()
* _PyObject_CallMethodIdObjArgs()
* _PyObject_Call_Prepend()
* _PyObject_FastCallDictTstate()
* _PyStack_AsDict()
The following functions are still exported for stdlib shared
extensions:
* _Py_CheckFunctionResult()
* _PyObject_MakeTpCall()
Mark the following internal functions as extern:
* _PyStack_UnpackDict()
* _PyStack_UnpackDict_Free()
* _PyStack_UnpackDict_FreeNoDecRef()
This effectively reverts bb578a0, restoring the original DEOPT_IF() macro in ceval_macros.h, and redefining it in the Tier 2 interpreter. We can get rid of the PREDICTED() macros there as well!
Added a new, experimental, tracing optimizer and interpreter (a.k.a. "tier 2"). This currently pessimizes, so don't use yet -- this is infrastructure so we can experiment with optimizing passes. To enable it, pass ``-Xuops`` or set ``PYTHONUOPS=1``. To get debug output, set ``PYTHONUOPSDEBUG=N`` where ``N`` is a debug level (0-4, where 0 is no debug output and 4 is excessively verbose).
All of this code is likely to change dramatically before the 3.13 feature freeze. But this is a first step.
Remove old aliases which were kept backwards compatibility with
Python 3.8:
* _PyObject_CallMethodNoArgs()
* _PyObject_CallMethodOneArg()
* _PyObject_CallOneArg()
* _PyObject_FastCallDict()
* _PyObject_Vectorcall()
* _PyObject_VectorcallMethod()
* _PyVectorcall_Function()
Update code which used these aliases to use new names.
These functions are broken by design because they discard any exceptions raised
inside, including MemoryError and KeyboardInterrupt. They should not be
used in new code.
* PyDict_GetItem() and PyObject_HasAttr() suppress arbitrary errors and
should not be used.
* PyUnicode_CompareWithASCIIString() only works if the second argument
is ASCII string.
* Refleak in get_suggestions_for_name_error.
* Use of borrowed pointer after possible freeing (self).
* Add some missing error checks.
It now raises an exception if sys.modules doesn't hold a strong
reference to the module.
Elaborate the comment explaining why a weak reference is used to
create a borrowed reference.
* Replace PyWeakref_GET_OBJECT() with _PyWeakref_GET_REF().
* _sqlite/blob.c now holds a strong reference to the blob object
while calling close_blob().
* _xidregistry_find_type() now holds a strong reference to registered
while using it.
finalize_modules_clear_weaklist() now holds a strong reference to the
module longer than before: replace PyWeakref_GET_OBJECT() with
_PyWeakref_GET_REF().
* Add tests on PyImport_AddModuleRef(), PyImport_AddModule() and
PyImport_AddModuleObject().
* pythonrun.c: Replace Py_XNewRef(PyImport_AddModule(name)) with
PyImport_AddModuleRef(name).
Refactor PyRun_InteractiveOneObjectEx(), _PyRun_SimpleFileObject()
and PyRun_SimpleStringFlags():
* Keep a strong reference to the __main__ module while using its
dictionary (PyModule_GetDict()). Use PyImport_AddModule() with
Py_XNewRef().
* Declare variables closer to where they are defined.
* Rename variables to use name longer than 1 character.
* Add pyrun_one_parse_ast() sub-function.
* Add table describing possible executable classes for out-of-process debuggers.
* Remove shim code object creation code as it is no longer needed.
* Make lltrace a bit more robust w.r.t. non-standard frames.
This fixes a race during import. The existing _PyRuntimeState.imports.pkgcontext is shared between interpreters, and occasionally this would cause a crash when multiple interpreters were importing extensions modules at the same time. To solve this we add a thread-local variable for the value. We also leave the existing state (and infrequent race) in place for platforms that do not support thread-local variables.
For a while now, pending calls only run in the main thread (in the main interpreter). This PR changes things to allow any thread run a pending call, unless the pending call was explicitly added for the main thread to run.
The risk of a race with this state is relatively low, but we play it safe anyway. We do avoid using the lock in performance-sensitive cases where the risk of a race is very, very low.
This avoids the problematic race in drop_gil() by skipping the FORCE_SWITCHING code there for finalizing threads.
(The idea for this approach came out of discussions with @markshannon.)
Remove functions in the C API:
* PyEval_AcquireLock()
* PyEval_ReleaseLock()
* PyEval_InitThreads()
* PyEval_ThreadsInitialized()
But keep these functions in the stable ABI.
Mention "make regen-limited-abi" in "make regen-all".
Remove the following old functions to configure the Python
initialization, deprecated in Python 3.11:
* PySys_AddWarnOptionUnicode()
* PySys_AddWarnOption()
* PySys_AddXOption()
* PySys_HasWarnOptions()
* PySys_SetArgvEx()
* PySys_SetArgv()
* PySys_SetPath()
* Py_SetPath()
* Py_SetProgramName()
* Py_SetPythonHome()
* Py_SetStandardStreamEncoding()
* _Py_SetProgramFullPath()
Most of these functions are kept in the stable ABI, except:
* Py_SetStandardStreamEncoding()
* _Py_SetProgramFullPath()
Update Doc/extending/embedding.rst and Doc/extending/extending.rst to
use the new PyConfig API.
_testembed.c:
* check_stdio_details() now sets stdio_encoding and stdio_errors
of PyConfig.
* Add definitions of functions removed from the API but kept in the
stable ABI.
* test_init_from_config() and test_init_read_set() now use
PyConfig_SetString() instead of PyConfig_SetBytesString().
Remove _Py_ClearStandardStreamEncoding() internal function.
Deprecate the old Py_UNICODE and PY_UNICODE_TYPE types in the C API:
use wchar_t instead.
Replace Py_UNICODE with wchar_t in multiple C files.
Co-authored-by: Inada Naoki <songofacandy@gmail.com>
Steve Dower
Irit Katriel
Finn Womack
Dong-hee Na
Mark Shannon
Brandt Bucher
Guido van Rossum
Eric Snow
Ivin Lee
Georg Brandl
Victor Stinner
Pablo Galindo Salgado
Carl Meyer
Kirill Podoprigora
Serhiy Storchaka
Inada Naoki
Kevin Diem
Kumar Aditya
Benjamin Peterson
Jeremy Paige
Md Sadman Chowdhury
Nikita Sobolev
Hood Chatham
hms
Erlend E. Aasland
chgnrdv
YAMAMOTO Takashi
Crowthebird
Erlend E. Aasland
Jelle Zijlstra