Sharing mutable (or non-immortal) objects between interpreters is generally not safe. We can work around that but not easily.
There are two restrictions that are critical for objects that break interpreter isolation.
The first is that the object's state be guarded by a global lock. For now the GIL meets this requirement, but a granular global lock is needed once we have a per-interpreter GIL.
The second restriction is that the object (and, for a container, its items) be deallocated/resized only when the interpreter in which it was allocated is the current one. This is because every interpreter has (or will have, see gh-101660) its own object allocator. Deallocating an object with a different allocator can cause crashes.
The dict for the cache of module defs is completely internal, which simplifies what we have to do to meet those requirements. To do so, we do the following:
* add a mechanism for re-using a temporary thread state tied to the main interpreter in an arbitrary thread
* add _PyRuntime.imports.extensions.main_tstate`
* add _PyThreadState_InitDetached() and _PyThreadState_ClearDetached() (pystate.c)
* add _PyThreadState_BindDetached() and _PyThreadState_UnbindDetached() (pystate.c)
* make sure the cache dict (_PyRuntime.imports.extensions.dict) and its items are all owned by the main interpreter)
* add a placeholder using for a granular global lock
Note that the cache is only used for legacy extension modules and not for multi-phase init modules.
https://github.com/python/cpython/issues/100227
This reverts commit 87be8d9.
This approach to keeping the interned strings safe is turning out to be too complex for my taste (due to obmalloc isolation). For now I'm going with the simpler solution, making the dict per-interpreter. We can revisit that later if we want a sharing solution.
This is effectively two changes. The first (the bulk of the change) is where we add _Py_AddToGlobalDict() (and _PyRuntime.cached_objects.main_tstate, etc.). The second (much smaller) change is where we update PyUnicode_InternInPlace() to use _Py_AddToGlobalDict() instead of calling PyDict_SetDefault() directly.
Basically, _Py_AddToGlobalDict() is a wrapper around PyDict_SetDefault() that should be used whenever we need to add a value to a runtime-global dict object (in the few cases where we are leaving the container global rather than moving it to PyInterpreterState, e.g. the interned strings dict). _Py_AddToGlobalDict() does all the necessary work to make sure the target global dict is shared safely between isolated interpreters. This is especially important as we move the obmalloc state to each interpreter (gh-101660), as well as, potentially, the GIL (PEP 684).
https://github.com/python/cpython/issues/100227
Moving it valuable with a per-interpreter GIL. However, it is also useful without one, since it allows us to identify refleaks within a single interpreter or where references are escaping an interpreter. This becomes more important as we move the obmalloc state to PyInterpreterState.
https://github.com/python/cpython/issues/102304
The essentially eliminates the global variable, with the associated benefits. This is also a precursor to isolating this bit of state to PyInterpreterState.
Folks that currently read _Py_RefTotal directly would have to start using _Py_GetGlobalRefTotal() instead.
https://github.com/python/cpython/issues/102304
It doesn't make sense to use multi-phase init for these modules. Using a per-interpreter "m_copy" (instead of PyModuleDef.m_base.m_copy) makes this work okay. (This came up while working on gh-101660.)
Note that we might instead end up disallowing re-load for sys/builtins since they are so special.
https://github.com/python/cpython/issues/102660
This change is almost entirely moving code around and hiding import state behind internal API. We introduce no changes to behavior, nor to non-internal API. (Since there was already going to be a lot of churn, I took this as an opportunity to re-organize import.c into topically-grouped sections of code.) The motivation is to simplify a number of upcoming changes.
Specific changes:
* move existing import-related code to import.c, wherever possible
* add internal API for interacting with import state (both global and per-interpreter)
* use only API outside of import.c (to limit churn there when changing the location, etc.)
* consolidate the import-related state of PyInterpreterState into a single struct field (this changes layout slightly)
* add macros for import state in import.c (to simplify changing the location)
* group code in import.c into sections
*remove _PyState_AddModule()
https://github.com/python/cpython/issues/101758
* Make sure that the current exception is always normalized.
* Remove redundant type and traceback fields for the current exception.
* Add new API functions: PyErr_GetRaisedException, PyErr_SetRaisedException
* Add new API functions: PyException_GetArgs, PyException_SetArgs
The GILState API (PEP 311) implementation from 2003 made the assumption that only one thread state would ever be used for any given OS thread, explicitly disregarding the case of subinterpreters. However, PyThreadState_Swap() still facilitated switching between subinterpreters, meaning the "current" thread state (holding the GIL), and the GILState thread state could end up out of sync, causing problems (including crashes).
This change addresses the issue by keeping the two in sync in PyThreadState_Swap(). I verified the fix against gh-99040.
Note that the other GILState-subinterpreter incompatibility (with autoInterpreterState) is not resolved here.
https://github.com/python/cpython/issues/59956
A PyThreadState can be in one of many states in its lifecycle, represented by some status value. Those statuses haven't been particularly clear, so we're addressing that here. Specifically:
* made the distinct lifecycle statuses clear on PyThreadState
* identified expectations of how various lifecycle-related functions relate to status
* noted the various places where those expectations don't match the actual behavior
At some point we'll need to address the mismatches.
(This change also includes some cleanup.)
https://github.com/python/cpython/issues/59956
We've factored out a struct from the two PyThreadState fields. This accomplishes two things:
* make it clear that the trashcan-related code doesn't need any other parts of PyThreadState
* allows us to use the trashcan mechanism even when there isn't a "current" thread state
We still expect the caller to hold the GIL.
https://github.com/python/cpython/issues/59956
This is a follow-up to gh-101161. The objective is to make it easier to read Python/pystate.c by grouping the functions there in a consistent way. This exclusively involves moving code around and adding various kinds of comments.
https://github.com/python/cpython/issues/59956
The objective of this change is to help make the GILState-related code easier to understand. This mostly involves moving code around and some semantically equivalent refactors. However, there are a also a small number of slight changes in structure and behavior:
* tstate_current is moved out of _PyRuntimeState.gilstate
* autoTSSkey is moved out of _PyRuntimeState.gilstate
* autoTSSkey is initialized earlier
* autoTSSkey is re-initialized (after fork) earlier
https://github.com/python/cpython/issues/59956
* Add API to allow extensions to set callback function on creation and destruction of PyCodeObject
Co-authored-by: Ye11ow-Flash <janshah@cs.stonybrook.edu>
The global allocators were stored in 3 static global variables: _PyMem_Raw, _PyMem, and _PyObject. State for the "small block" allocator was stored in another 13. That makes a total of 16 global variables. We are moving all 16 to the _PyRuntimeState struct as part of the work for gh-81057. (If PEP 684 is accepted then we will follow up by moving them all to PyInterpreterState.)
https://github.com/python/cpython/issues/81057
* Adds EXIT_INTERPRETER instruction to exit PyEval_EvalDefault()
* Simplifies RETURN_VALUE, YIELD_VALUE and RETURN_GENERATOR instructions as they no longer need to check for entry frames.
(see https://github.com/python/cpython/issues/98608)
This change does the following:
1. change the argument to a new `_PyInterpreterConfig` struct
2. rename the function to `_Py_NewInterpreterFromConfig()`, inspired by `Py_InitializeFromConfig()` (takes a `_PyInterpreterConfig` instead of `isolated_subinterpreter`)
3. split up the boolean `isolated_subinterpreter` into the corresponding multiple granular settings
* allow_fork
* allow_subprocess
* allow_threads
4. add `PyInterpreterState.feature_flags` to store those settings
5. add a function for checking if a feature is enabled on an opaque `PyInterpreterState *`
6. drop `PyConfig._isolated_interpreter`
The existing default (see `Py_NewInterpeter()` and `Py_Initialize*()`) allows fork, subprocess, and threads and the optional "isolated" interpreter (see the `_xxsubinterpreters` module) disables all three. None of that changes here; the defaults are preserved.
Note that the given `_PyInterpreterConfig` will not be used outside `_Py_NewInterpreterFromConfig()`, nor preserved. This contrasts with how `PyConfig` is currently preserved, used, and even modified outside `Py_InitializeFromConfig()`. I'd rather just avoid that mess from the start for `_PyInterpreterConfig`. We can preserve it later if we find an actual need.
This change allows us to follow up with a number of improvements (e.g. stop disallowing subprocess and support disallowing exec instead).
(Note that this PR adds "private" symbols. We'll probably make them public, and add docs, in a separate change.)
This was added for bpo-40514 (gh-84694) to test out a per-interpreter GIL. However, it has since proven unnecessary to keep the experiment in the repo. (It can be done as a branch in a fork like normal.) So here we are removing:
* the configure option
* the macro
* the code enabled by the macro
Previously, the main interpreter was allocated on the heap during runtime initialization. Here we instead embed it into _PyRuntimeState, which means it is statically allocated as part of the _PyRuntime global. The same goes for the initial thread state (of each interpreter, including the main one). Consequently there are fewer allocations during runtime/interpreter init, fewer possible failures, and better memory locality.
FYI, this also helps efforts to consolidate globals, which in turns helps work on subinterpreter isolation.
https://bugs.python.org/issue45953
* Do not PUSH/POP traceback or type to the stack as part of exc_info
* Remove exc_traceback and exc_type from _PyErr_StackItem
* Add to what's new, because this change breaks things like Cython
Previously, basic initialization of PyInterprterState happened in PyInterpreterState_New() (along with allocation and adding the new interpreter to the runtime state). This prevented us from initializing interpreter states that were allocated separately (e.g. statically or in a free list). We've addressed that here by factoring out a separate function just for initialization. We've done the same for PyThreadState. _PyRuntimeState was sorted out when we added it since _PyRuntime is statically allocated. However, here we update the existing init code to line up with the functions for PyInterpreterState and PyThreadState.
https://bugs.python.org/issue46008
PyInterpreterState_Main() is a plain function exposed in the public C-API. For internal usage we can take the more efficient approach in this PR.
https://bugs.python.org/issue46008
This simplifies new_threadstate(). We also rename _PyThreadState_Init() to _PyThreadState_SetCurrent() to reflect what it actually does.
https://bugs.python.org/issue46008
Doing so allows us to stop assigning various fields to `NULL` and 0. It also more closely matches the behavior of a static initializer.
Automerge-Triggered-By: GH:ericsnowcurrently
This parallels _PyRuntimeState.interpreters. Doing this helps make it more clear what part of PyInterpreterState relates to its threads.
https://bugs.python.org/issue46008
This falls into the category of keep-allocation-and-initialization separate. It also allows us to use _PyEval_InitState() safely in functions that return void.
https://bugs.python.org/issue46008
* Make internal APIs that take PyFrameConstructor take a PyFunctionObject instead.
* Add reference to function to frame, borrow references to builtins and globals.
* Add COPY_FREE_VARS instruction to allow specialization of calls to inner functions.
Unlike the other locks reinitialized by _PyRuntimeState_ReInitThreads,
the "interpreters.main->id_mutex" is not freed by _PyRuntimeState_Fini
and should not force the default raw allocator.
Add PyThreadState_EnterTracing() and PyThreadState_LeaveTracing()
functions to the limited C API to suspend and resume tracing and
profiling.
Add an unit test on the PyThreadState C API to _testcapi.
Add also internal _PyThreadState_DisableTracing() and
_PyThreadState_ResetTracing().
Redefining the PyThreadState_GET() macro in pycore_pystate.h is
useless since it doesn't affect files not including it. Either use
_PyThreadState_GET() directly, or don't use pycore_pystate.h internal
C API. For example, the _testcapi extension don't use the internal C
API, but use the public PyThreadState_Get() function instead.
Replace PyThreadState_Get() with _PyThreadState_GET(). The
_PyThreadState_GET() macro is more efficient than PyThreadState_Get()
and PyThreadState_GET() function calls which call fail with a fatal
Python error.
posixmodule.c and _ctypes extension now include <windows.h> before
pycore header files (like pycore_call.h).
_PyTraceback_Add() now uses _PyErr_Fetch()/_PyErr_Restore() instead
of PyErr_Fetch()/PyErr_Restore().
The _decimal and _xxsubinterpreters extensions are now built with the
Py_BUILD_CORE_MODULE macro defined to get access to the internal C
API.
Places the locals between the specials and stack. This is the more "natural" layout for a C struct, makes the code simpler and gives a slight speedup (~1%)
* Convert "specials" array to InterpreterFrame struct, adding f_lasti, f_state and other non-debug FrameObject fields to it.
* Refactor, calls pushing the call to the interpreter upward toward _PyEval_Vector.
* Compute f_back when on thread stack, only filling in value when frame object outlives stack invocation.
* Move ownership of InterpreterFrame in generator from frame object to generator object.
* Do not create frame objects for Python calls.
* Do not create frame objects for generators.
* Remove 'zombie' frames. We won't need them once we are allocating fixed-size frames.
* Add co_nlocalplus field to code object to avoid recomputing size of locals + frees + cells.
* Move locals, cells and freevars out of frame object into separate memory buffer.
* Use per-threadstate allocated memory chunks for local variables.
* Move globals and builtins from frame object to per-thread stack.
* Move (slow) locals frame object to per-thread stack.
* Move internal frame functions to internal header.
* Add Py_TPFLAGS_SEQUENCE and Py_TPFLAGS_MAPPING, add to all relevant standard builtin classes.
* Set relevant flags on collections.abc.Sequence and Mapping.
* Use flags in MATCH_SEQUENCE and MATCH_MAPPING opcodes.
* Inherit Py_TPFLAGS_SEQUENCE and Py_TPFLAGS_MAPPING.
* Add NEWS
* Remove interpreter-state map_abc and seq_abc fields.
* Remove redundant tracing_possible field from interpreter state.
* Move 'use_tracing' from tstate onto C stack, for fastest possible checking in dispatch logic.
* Add comments stressing the importance stack discipline when dealing with CFrames.
* Add NEWS
At Python startup, call _PyGILState_Init() before
PyInterpreterState_New() which calls _PyThreadState_GET(). When
Python is built using --with-experimental-isolated-subinterpreters,
_PyThreadState_GET() uses autoTSSkey.
Pass the current interpreter (interp) rather than the current Python
thread state (tstate) to internal functions which only use the
interpreter.
Modified functions:
* _PyXXX_Fini() and _PyXXX_ClearFreeList() functions
* _PyEval_SignalAsyncExc(), make_pending_calls()
* _PySys_GetObject(), sys_set_object(), sys_set_object_id(), sys_set_object_str()
* should_audit(), set_flags_from_config(), make_flags()
* _PyAtExit_Call()
* init_stdio_encoding()
* etc.
Fix the _PyUnicode_FromId() function (_Py_IDENTIFIER(var) API) when
Py_Initialize() / Py_Finalize() is called multiple times:
preserve _PyRuntime.unicode_ids.next_index value.
Use _PyRuntimeState_INIT macro instead memset(0) to reset
_PyRuntimeState members to zero.
Make the type attribute lookup cache per-interpreter.
Add private _PyType_InitCache() function, called by PyInterpreterState_New().
Continue to share next_version_tag between interpreters, since static
types are still shared by interpreters.
Remove MCACHE macro: the cache is no longer disabled if the
EXPERIMENTAL_ISOLATED_SUBINTERPRETERS macro is defined.
Make _PyUnicode_FromId() function compatible with subinterpreters.
Each interpreter now has an array of identifier objects (interned
strings decoded from UTF-8).
* Add PyInterpreterState.unicode.identifiers: array of identifiers
objects.
* Add _PyRuntimeState.unicode_ids used to allocate unique indexes
to _Py_Identifier.
* Rewrite the _Py_Identifier structure.
Microbenchmark on _PyUnicode_FromId(&PyId_a) with _Py_IDENTIFIER(a):
[ref] 2.42 ns +- 0.00 ns -> [atomic] 3.39 ns +- 0.00 ns: 1.40x slower
This change adds 1 ns per _PyUnicode_FromId() call in average.
* Add _PyAtExit_Call() function and remove pyexitfunc and
pyexitmodule members of PyInterpreterState. The function
logs atexit callback errors using _PyErr_WriteUnraisableMsg().
* Add _PyAtExit_Init() and _PyAtExit_Fini() functions.
* Remove traverse, clear and free functions of the atexit module.
Co-authored-by: Dong-hee Na <donghee.na@python.org>
* Inline _PyInterpreterState_SetConfig(): replace it with
_PyConfig_Copy().
* Add _PyErr_SetFromPyStatus()
* Add _PyInterpreterState_GetConfigCopy()
* Add a new _PyInterpreterState_SetConfig() function.
* Add an unit which gets, modifies, and sets the config.
Call _PyAST_Fini() on all interpreters, not only on the main
interpreter. Also, call it ealier to fix a reference leak.
Python types contain a reference to themselves in in their
PyTypeObject.tp_mro member. _PyAST_Fini() must called before the last
GC collection to destroy AST types.
_PyInterpreterState_Clear() now calls _PyAST_Fini(). It now also
calls _PyWarnings_Fini() on subinterpeters, not only on the main
interpreter.
Add an assertion in AST init_types() to ensure that the _ast module
is no longer used after _PyAST_Fini() has been called.
This adds a new function named sys._current_exceptions() which is equivalent ot
sys._current_frames() except that it returns the exceptions currently handled
by other threads. It is equivalent to calling sys.exc_info() for each running
thread.
Eric Snow