大师兄的Python源码学习笔记(五十八）: Python的内存

大师兄的Python源码学习笔记(五十七）: Python的内存管理机制(十二）
大师兄的Python源码学习笔记(五十九）: Python的内存管理机制(十四）

五、Python中的垃圾收集

4. 垃圾收集全景

• 回顾实际完成垃圾收集的collect方法：

gcmodule.c

/* This is the main function.  Read this to understand how the
 * collection process works. */
static Py_ssize_t
collect(int generation, Py_ssize_t *n_collected, Py_ssize_t *n_uncollectable,
        int nofail)
{
    int i;
    Py_ssize_t m = 0; /* # objects collected */
    Py_ssize_t n = 0; /* # unreachable objects that couldn't be collected */
    PyGC_Head *young; /* the generation we are examining */
    PyGC_Head *old; /* next older generation */
    PyGC_Head unreachable; /* non-problematic unreachable trash */
    PyGC_Head finalizers;  /* objects with, & reachable from, __del__ */
    PyGC_Head *gc;
    _PyTime_t t1 = 0;   /* initialize to prevent a compiler warning */

    struct gc_generation_stats *stats = &_PyRuntime.gc.generation_stats[generation];

    if (_PyRuntime.gc.debug & DEBUG_STATS) {
        PySys_WriteStderr("gc: collecting generation %d...\n",
                          generation);
        PySys_WriteStderr("gc: objects in each generation:");
        for (i = 0; i < NUM_GENERATIONS; i++)
            PySys_FormatStderr(" %zd",
                              gc_list_size(GEN_HEAD(i)));
        PySys_WriteStderr("\ngc: objects in permanent generation: %zd",
                         gc_list_size(&_PyRuntime.gc.permanent_generation.head));
        t1 = _PyTime_GetMonotonicClock();

        PySys_WriteStderr("\n");
    }

    if (PyDTrace_GC_START_ENABLED())
        PyDTrace_GC_START(generation);

    /* update collection and allocation counters */
    1. 将比当前处理的更年轻的代链表合并到当前代中
    if (generation+1 < NUM_GENERATIONS)
        _PyRuntime.gc.generations[generation+1].count += 1;
    for (i = 0; i <= generation; i++)
        _PyRuntime.gc.generations[i].count = 0;

    /* merge younger generations with one we are currently collecting */
    for (i = 0; i < generation; i++) {
        gc_list_merge(GEN_HEAD(i), GEN_HEAD(generation));
    }

    /* handy references */
    young = GEN_HEAD(generation);
    if (generation < NUM_GENERATIONS-1)
        old = GEN_HEAD(generation+1);
    else
        old = young;

    /* Using ob_refcnt and gc_refs, calculate which objects in the
     * container set are reachable from outside the set (i.e., have a
     * refcount greater than 0 when all the references within the
     * set are taken into account).
     */
    2. 在待处理链表上进行打破循环的模拟，寻找root object
    update_refs(young);
    subtract_refs(young);

    /* Leave everything reachable from outside young in young, and move
     * everything else (in young) to unreachable.
     * NOTE:  This used to move the reachable objects into a reachable
     * set instead.  But most things usually turn out to be reachable,
     * so it's more efficient to move the unreachable things.
     */
    3. 将待处理链表中的unreachable object转移到unreachable链表中，当前代中只剩下reachable object。
    gc_list_init(&unreachable);
    move_unreachable(young, &unreachable);
    
    /* Move reachable objects to next generation. */
    4. 将当前代中的reachable object合并到更老的代中。
    if (young != old) {
        if (generation == NUM_GENERATIONS - 2) {
            _PyRuntime.gc.long_lived_pending += gc_list_size(young);
        }
        gc_list_merge(young, old);
    }
    else {
        /* We only untrack dicts in full collections, to avoid quadratic
           dict build-up. See issue #14775. */
        untrack_dicts(young);
        _PyRuntime.gc.long_lived_pending = 0;
        _PyRuntime.gc.long_lived_total = gc_list_size(young);
    }

    /* All objects in unreachable are trash, but objects reachable from
     * legacy finalizers (e.g. tp_del) can't safely be deleted.
     */
    5. 将 unreachable链表中带有__del__函数的对象与其引用对象收集到finalizers链表中。
    gc_list_init(&finalizers);
    move_legacy_finalizers(&unreachable, &finalizers);
    /* finalizers contains the unreachable objects with a legacy finalizer;
     * unreachable objects reachable *from* those are also uncollectable,
     * and we move those into the finalizers list too.
     */
    move_legacy_finalizer_reachable(&finalizers);

    /* Collect statistics on collectable objects found and print
     * debugging information.
     */
    for (gc = unreachable.gc.gc_next; gc != &unreachable;
                    gc = gc->gc.gc_next) {
        m++;
        if (_PyRuntime.gc.debug & DEBUG_COLLECTABLE) {
            debug_cycle("collectable", FROM_GC(gc));
        }
    }

    /* Clear weakrefs and invoke callbacks as necessary. */
    6. 处理弱引用，尝试调用其callback操作
    m += handle_weakrefs(&unreachable, old);

    /* Call tp_finalize on objects which have one. */
    finalize_garbage(&unreachable);

    if (check_garbage(&unreachable)) {
        revive_garbage(&unreachable);
        gc_list_merge(&unreachable, old);
    }
    else {
        /* Call tp_clear on objects in the unreachable set.  This will cause
         * the reference cycles to be broken.  It may also cause some objects
         * in finalizers to be freed.
         */
    7. 对unreachable链表上的对象进行垃圾回收操作
        delete_garbage(&unreachable, old);
    }

    /* Collect statistics on uncollectable objects found and print
     * debugging information. */
    for (gc = finalizers.gc.gc_next;
         gc != &finalizers;
         gc = gc->gc.gc_next) {
        n++;
        if (_PyRuntime.gc.debug & DEBUG_UNCOLLECTABLE)
            debug_cycle("uncollectable", FROM_GC(gc));
    }
    if (_PyRuntime.gc.debug & DEBUG_STATS) {
        _PyTime_t t2 = _PyTime_GetMonotonicClock();

        if (m == 0 && n == 0)
            PySys_WriteStderr("gc: done");
        else
            PySys_FormatStderr(
                "gc: done, %zd unreachable, %zd uncollectable",
                n+m, n);
        PySys_WriteStderr(", %.4fs elapsed\n",
                          _PyTime_AsSecondsDouble(t2 - t1));
    }

    /* Append instances in the uncollectable set to a Python
     * reachable list of garbage.  The programmer has to deal with
     * this if they insist on creating this type of structure.
     */
    8. 将含有__del__操作的实例收集到garbage链表中，同时将finalizers链表中所有对象加入到old链表中。
    handle_legacy_finalizers(&finalizers, old);

    /* Clear free list only during the collection of the highest
     * generation */
    if (generation == NUM_GENERATIONS-1) {
        clear_freelists();
    }

    if (PyErr_Occurred()) {
        if (nofail) {
            PyErr_Clear();
        }
        else {
            if (gc_str == NULL)
                gc_str = PyUnicode_FromString("garbage collection");
            PyErr_WriteUnraisable(gc_str);
            Py_FatalError("unexpected exception during garbage collection");
        }
    }

    /* Update stats */
    if (n_collected)
        *n_collected = m;
    if (n_uncollectable)
        *n_uncollectable = n;
    stats->collections++;
    stats->collected += m;
    stats->uncollectable += n;

    if (PyDTrace_GC_DONE_ENABLED())
        PyDTrace_GC_DONE(n+m);

    return n+m;
}

• 可以注意到Python对弱引用weakref的处理，因为weakref能够注册callback操作，所以这个行为类似带有__del__的实例对象。
• 区别是weakref能够被正确的清理掉，而带有__del__的实例对象不能被自动清除，而是被放入garbage链表中。

gcmodule.c

static int
handle_weakrefs(PyGC_Head *unreachable, PyGC_Head *old)
{
    PyGC_Head *gc;
    PyObject *op;               /* generally FROM_GC(gc) */
    PyWeakReference *wr;        /* generally a cast of op */
    PyGC_Head wrcb_to_call;     /* weakrefs with callbacks to call */
    PyGC_Head *next;
    int num_freed = 0;

    gc_list_init(&wrcb_to_call);

    /* Clear all weakrefs to the objects in unreachable.  If such a weakref
     * also has a callback, move it into `wrcb_to_call` if the callback
     * needs to be invoked.  Note that we cannot invoke any callbacks until
     * all weakrefs to unreachable objects are cleared, lest the callback
     * resurrect an unreachable object via a still-active weakref.  We
     * make another pass over wrcb_to_call, invoking callbacks, after this
     * pass completes.
     */
    for (gc = unreachable->gc.gc_next; gc != unreachable; gc = next) {
        PyWeakReference **wrlist;

        op = FROM_GC(gc);
        assert(IS_TENTATIVELY_UNREACHABLE(op));
        next = gc->gc.gc_next;

        if (! PyType_SUPPORTS_WEAKREFS(Py_TYPE(op)))
            continue;

        /* It supports weakrefs.  Does it have any? */
        wrlist = (PyWeakReference **)
                                PyObject_GET_WEAKREFS_LISTPTR(op);

        /* `op` may have some weakrefs.  March over the list, clear
         * all the weakrefs, and move the weakrefs with callbacks
         * that must be called into wrcb_to_call.
         */
        for (wr = *wrlist; wr != NULL; wr = *wrlist) {
            PyGC_Head *wrasgc;                  /* AS_GC(wr) */

            /* _PyWeakref_ClearRef clears the weakref but leaves
             * the callback pointer intact.  Obscure:  it also
             * changes *wrlist.
             */
            assert(wr->wr_object == op);
            _PyWeakref_ClearRef(wr);
            assert(wr->wr_object == Py_None);
            if (wr->wr_callback == NULL)
                continue;                       /* no callback */

    /* Headache time.  `op` is going away, and is weakly referenced by
     * `wr`, which has a callback.  Should the callback be invoked?  If wr
     * is also trash, no:
     *
     * 1. There's no need to call it.  The object and the weakref are
     *    both going away, so it's legitimate to pretend the weakref is
     *    going away first.  The user has to ensure a weakref outlives its
     *    referent if they want a guarantee that the wr callback will get
     *    invoked.
     *
     * 2. It may be catastrophic to call it.  If the callback is also in
     *    cyclic trash (CT), then although the CT is unreachable from
     *    outside the current generation, CT may be reachable from the
     *    callback.  Then the callback could resurrect insane objects.
     *
     * Since the callback is never needed and may be unsafe in this case,
     * wr is simply left in the unreachable set.  Note that because we
     * already called _PyWeakref_ClearRef(wr), its callback will never
     * trigger.
     *
     * OTOH, if wr isn't part of CT, we should invoke the callback:  the
     * weakref outlived the trash.  Note that since wr isn't CT in this
     * case, its callback can't be CT either -- wr acted as an external
     * root to this generation, and therefore its callback did too.  So
     * nothing in CT is reachable from the callback either, so it's hard
     * to imagine how calling it later could create a problem for us.  wr
     * is moved to wrcb_to_call in this case.
     */
            if (IS_TENTATIVELY_UNREACHABLE(wr))
                continue;
            assert(IS_REACHABLE(wr));

            /* Create a new reference so that wr can't go away
             * before we can process it again.
             */
            Py_INCREF(wr);

            /* Move wr to wrcb_to_call, for the next pass. */
            wrasgc = AS_GC(wr);
            assert(wrasgc != next); /* wrasgc is reachable, but
                                       next isn't, so they can't
                                       be the same */
            gc_list_move(wrasgc, &wrcb_to_call);
        }
    }

    /* Invoke the callbacks we decided to honor.  It's safe to invoke them
     * because they can't reference unreachable objects.
     */
    while (! gc_list_is_empty(&wrcb_to_call)) {
        PyObject *temp;
        PyObject *callback;

        gc = wrcb_to_call.gc.gc_next;
        op = FROM_GC(gc);
        assert(IS_REACHABLE(op));
        assert(PyWeakref_Check(op));
        wr = (PyWeakReference *)op;
        callback = wr->wr_callback;
        assert(callback != NULL);

        /* copy-paste of weakrefobject.c's handle_callback() */
        temp = PyObject_CallFunctionObjArgs(callback, wr, NULL);
        if (temp == NULL)
            PyErr_WriteUnraisable(callback);
        else
            Py_DECREF(temp);

        /* Give up the reference we created in the first pass.  When
         * op's refcount hits 0 (which it may or may not do right now),
         * op's tp_dealloc will decref op->wr_callback too.  Note
         * that the refcount probably will hit 0 now, and because this
         * weakref was reachable to begin with, gc didn't already
         * add it to its count of freed objects.  Example:  a reachable
         * weak value dict maps some key to this reachable weakref.
         * The callback removes this key->weakref mapping from the
         * dict, leaving no other references to the weakref (excepting
         * ours).
         */
        Py_DECREF(op);
        if (wrcb_to_call.gc.gc_next == gc) {
            /* object is still alive -- move it */
            gc_list_move(gc, old);
        }
        else
            ++num_freed;
    }

    return num_freed;
}

• 到这里可以看出，Python的垃圾收集机制完全是为了处理循环引用而设计的。
• 虽然几乎大多数对象在创建时都会通过PyObject_GC_New，并最终调用_PyObject_GC_New，将创建的对象纳入垃圾收集机制的监控中。
• 但垃圾收集监控的对象并非只有垃圾收集机制才能回收，正常的引用计数就能销毁掉一个被纳入垃圾回收机制监控的对象：

funcobject.c

static void
func_dealloc(PyFunctionObject *op)
{
    _PyObject_GC_UNTRACK(op);
    if (op->func_weakreflist != NULL)
        PyObject_ClearWeakRefs((PyObject *) op);
    Py_DECREF(op->func_code);
    Py_DECREF(op->func_globals);
    Py_XDECREF(op->func_module);
    Py_DECREF(op->func_name);
    Py_XDECREF(op->func_defaults);
    Py_XDECREF(op->func_kwdefaults);
    Py_XDECREF(op->func_doc);
    Py_XDECREF(op->func_dict);
    Py_XDECREF(op->func_closure);
    Py_XDECREF(op->func_annotations);
    Py_XDECREF(op->func_qualname);
    PyObject_GC_Del(op);
}

Modules\gcmodule.c

void
PyObject_GC_Del(void *op)
{
    PyGC_Head *g = AS_GC(op);
    if (IS_TRACKED(op))
        gc_list_remove(g);
    if (_PyRuntime.gc.generations[0].count > 0) {
        _PyRuntime.gc.generations[0].count--;
    }
    PyObject_FREE(g);
}

• 如果PyFunctionObject对象因为正常的引用计数维护到达引用计数为0的状态，就会调用func_dealloc。
• 在这里，PyFunctionObject对象主动将自己从垃圾收集监控的链表中摘除，然后调用PyObject_GC_Del释放内存。
• 之所以调用PyObject_GC_Del，主要为了将指向PyObject的指针调整为指向PyGC_Head的指针，以释放正确的内存。
• 所以，虽然有很多对象挂在垃圾收集机制监控的链表上，但更多时候是引用计数机制在维护这些对象。
• 只有面对引用计数无能为力的循环引用，垃圾收集机制才会生效。
• 实际上面对除了循环引用之外的对象，垃圾收集是无能为力的：

• 因为挂在垃圾收集机制上的对象都是引用计数不为0的，如果为0则已经被引用计数处理了。
• 而引用计数不为0的对象有两种情况，一是被程序使用的对象，二是循环用用中的对象。
• 被程序使用的对象不能回收，所以垃圾回收面对的只有循环引用中的对象。

• 另外在大多数情况下，Python都在使用内存池，所以垃圾收集和内存管理是融为一体的。