Python 行程中的多重直譯器¶

While in most uses, you will only embed a single Python interpreter, there are cases where you need to create several independent interpreters in the same process and perhaps even in the same thread. Sub-interpreters allow you to do that.

The "main" interpreter is the first one created when the runtime initializes. It is usually the only Python interpreter in a process. Unlike sub-interpreters, the main interpreter has unique process-global responsibilities like signal handling. It is also responsible for execution during runtime initialization and is usually the active interpreter during runtime finalization. The PyInterpreterState_Main() function returns a pointer to its state.

You can switch between sub-interpreters using the PyThreadState_Swap() function. You can create and destroy them using the following functions:

type PyInterpreterConfig¶

Structure containing most parameters to configure a sub-interpreter. Its values are used only in Py_NewInterpreterFromConfig() and never modified by the runtime.

在 3.12 版被加入.

結構欄位：

int use_main_obmalloc¶

If this is 0 then the sub-interpreter will use its own "object" allocator state. Otherwise it will use (share) the main interpreter's.

If this is 0 then check_multi_interp_extensions must be 1 (non-zero). If this is 1 then gil must not be PyInterpreterConfig_OWN_GIL.

int allow_fork¶

If this is 0 then the runtime will not support forking the process in any thread where the sub-interpreter is currently active. Otherwise fork is unrestricted.

Note that the subprocess module still works when fork is disallowed.

int allow_exec¶

If this is 0 then the runtime will not support replacing the current process via exec (e.g. os.execv()) in any thread where the sub-interpreter is currently active. Otherwise exec is unrestricted.

Note that the subprocess module still works when exec is disallowed.

int allow_threads¶: If this is 0 then the sub-interpreter's threading module won't create threads. Otherwise threads are allowed.

int allow_daemon_threads¶: If this is 0 then the sub-interpreter's threading module won't create daemon threads. Otherwise daemon threads are allowed (as long as allow_threads is non-zero).

int check_multi_interp_extensions¶

If this is 0 then all extension modules may be imported, including legacy (single-phase init) modules, in any thread where the sub-interpreter is currently active. Otherwise only multi-phase init extension modules (see PEP 489) may be imported. (Also see Py_mod_multiple_interpreters.)

This must be 1 (non-zero) if use_main_obmalloc is 0.

int gil¶

This determines the operation of the GIL for the sub-interpreter. It may be one of the following:

PyInterpreterConfig_DEFAULT_GIL¶: Use the default selection (PyInterpreterConfig_SHARED_GIL).

PyInterpreterConfig_SHARED_GIL¶: Use (share) the main interpreter's GIL.

PyInterpreterConfig_OWN_GIL¶: Use the sub-interpreter's own GIL.

If this is PyInterpreterConfig_OWN_GIL then PyInterpreterConfig.use_main_obmalloc must be 0.

PyStatus Py_NewInterpreterFromConfig(PyThreadState **tstate_p, const PyInterpreterConfig *config)¶

Create a new sub-interpreter. This is an (almost) totally separate environment for the execution of Python code. In particular, the new interpreter has separate, independent versions of all imported modules, including the fundamental modules builtins, __main__ and sys. The table of loaded modules (sys.modules) and the module search path (sys.path) are also separate. The new environment has no sys.argv variable. It has new standard I/O stream file objects sys.stdin, sys.stdout and sys.stderr (however these refer to the same underlying file descriptors).

The given config controls the options with which the interpreter is initialized.

Upon success, tstate_p will be set to the first thread state created in the new sub-interpreter. This thread state is attached. Note that no actual thread is created; see the discussion of thread states below. If creation of the new interpreter is unsuccessful, tstate_p is set to NULL; no exception is set since the exception state is stored in the attached thread state, which might not exist.

Like all other Python/C API functions, an attached thread state must be present before calling this function, but it might be detached upon returning. On success, the returned thread state will be attached. If the sub-interpreter is created with its own GIL then the attached thread state of the calling interpreter will be detached. When the function returns, the new interpreter's thread state will be attached to the current thread and the previous interpreter's attached thread state will remain detached.

在 3.12 版被加入.

Sub-interpreters are most effective when isolated from each other, with certain functionality restricted:

PyInterpreterConfig config = {
    .use_main_obmalloc = 0,
    .allow_fork = 0,
    .allow_exec = 0,
    .allow_threads = 1,
    .allow_daemon_threads = 0,
    .check_multi_interp_extensions = 1,
    .gil = PyInterpreterConfig_OWN_GIL,
};
PyThreadState *tstate = NULL;
PyStatus status = Py_NewInterpreterFromConfig(&tstate, &config);
if (PyStatus_Exception(status)) {
    Py_ExitStatusException(status);
}

Note that the config is used only briefly and does not get modified. During initialization the config's values are converted into various PyInterpreterState values. A read-only copy of the config may be stored internally on the PyInterpreterState.

Extension modules are shared between (sub-)interpreters as follows:

For modules using multi-phase initialization, e.g. PyModule_FromDefAndSpec(), a separate module object is created and initialized for each interpreter. Only C-level static and global variables are shared between these module objects.
For modules using legacy single-phase initialization, e.g. PyModule_Create(), the first time a particular extension is imported, it is initialized normally, and a (shallow) copy of its module's dictionary is squirreled away. When the same extension is imported by another (sub-)interpreter, a new module is initialized and filled with the contents of this copy; the extension's init function is not called. Objects in the module's dictionary thus end up shared across (sub-)interpreters, which might cause unwanted behavior (see Bugs and caveats below).

Note that this is different from what happens when an extension is imported after the interpreter has been completely re-initialized by calling Py_FinalizeEx() and Py_Initialize(); in that case, the extension's initmodule function is called again. As with multi-phase initialization, this means that only C-level static and global variables are shared between these modules.

PyThreadState *Py_NewInterpreter(void)¶: 為穩定 ABI 的一部分.
Create a new sub-interpreter. This is essentially just a wrapper around Py_NewInterpreterFromConfig() with a config that preserves the existing behavior. The result is an unisolated sub-interpreter that shares the main interpreter's GIL, allows fork/exec, allows daemon threads, and allows single-phase init modules.

void Py_EndInterpreter(PyThreadState *tstate)¶

為穩定 ABI 的一部分.

Destroy the (sub-)interpreter represented by the given thread state. The given thread state must be attached. When the call returns, there will be no attached thread state. All thread states associated with this interpreter are destroyed.

Py_FinalizeEx() will destroy all sub-interpreters that haven't been explicitly destroyed at that point.

A per-interpreter GIL¶

在 3.12 版被加入.

Using Py_NewInterpreterFromConfig() you can create a sub-interpreter that is completely isolated from other interpreters, including having its own GIL. The most important benefit of this isolation is that such an interpreter can execute Python code without being blocked by other interpreters or blocking any others. Thus a single Python process can truly take advantage of multiple CPU cores when running Python code. The isolation also encourages a different approach to concurrency than that of just using threads. (See PEP 554 and PEP 684.)

Using an isolated interpreter requires vigilance in preserving that isolation. That especially means not sharing any objects or mutable state without guarantees about thread-safety. Even objects that are otherwise immutable (e.g. None, (1, 5)) can't normally be shared because of the refcount. One simple but less-efficient approach around this is to use a global lock around all use of some state (or object). Alternately, effectively immutable objects (like integers or strings) can be made safe in spite of their refcounts by making them immortal. In fact, this has been done for the builtin singletons, small integers, and a number of other builtin objects.

If you preserve isolation then you will have access to proper multi-core computing without the complications that come with free-threading. Failure to preserve isolation will expose you to the full consequences of free-threading, including races and hard-to-debug crashes.

Aside from that, one of the main challenges of using multiple isolated interpreters is how to communicate between them safely (not break isolation) and efficiently. The runtime and stdlib do not provide any standard approach to this yet. A future stdlib module would help mitigate the effort of preserving isolation and expose effective tools for communicating (and sharing) data between interpreters.