SF 1055820: weakref callback vs gc vs threads

In cyclic gc, clear weakrefs to unreachable objects before allowing any
Python code (weakref callbacks or __del__ methods) to run.

This is a critical bugfix, affecting all versions of Python since weakrefs
were introduced.  I'll backport to 2.3.

Include/weakrefobject.h

@@ -9,11 +9,31 @@ extern "C" {
 
 typedef struct _PyWeakReference PyWeakReference;
 
+/* PyWeakReference is the base struct for the Python ReferenceType, ProxyType,
+ * and CallableProxyType.
+ */
 struct _PyWeakReference {
     PyObject_HEAD
+
+    /* The object to which this is a weak reference, or Py_None if none.
+     * Note that this is a stealth reference:  wr_object's refcount is
+     * not incremented to reflect this pointer.
+     */
     PyObject *wr_object;
+
+    /* A callable to invoke when wr_object dies, or NULL if none. */
     PyObject *wr_callback;
+
+    /* A cache for wr_object's hash code.  As usual for hashes, this is -1
+     * if the hash code isn't known yet.
+     */
     long hash;
+
+    /* If wr_object is weakly referenced, wr_object has a doubly-linked NULL-
+     * terminated list of weak references to it.  These are the list pointers.
+     * If wr_object goes away, wr_object is set to Py_None, and these pointers
+     * have no meaning then.
+     */
     PyWeakReference *wr_prev;
     PyWeakReference *wr_next;
 };

Lib/test/test_gc.py

 from test.test_support import verify, verbose, TestFailed, vereq
 import sys
 import gc
+import weakref
 
 def expect(actual, expected, name):
     if actual != expected:
@@ -369,6 +370,191 @@ def test_get_referents():
 
     expect(gc.get_referents(1, 'a', 4j), [], "get_referents")
 
+# Bug 1055820 has several tests of longstanding bugs involving weakrefs and
+# cyclic gc.
+
+# An instance of C1055820 has a self-loop, so becomes cyclic trash when
+# unreachable.
+class C1055820(object):
+    def __init__(self, i):
+        self.i = i
+        self.loop = self
+
+class GC_Detector(object):
+    # Create an instance I.  Then gc hasn't happened again so long as
+    # I.gc_happened is false.
+
+    def __init__(self):
+        self.gc_happened = False
+
+        def it_happened(ignored):
+            self.gc_happened = True
+
+        # Create a piece of cyclic trash that triggers it_happened when
+        # gc collects it.
+        self.wr = weakref.ref(C1055820(666), it_happened)
+
+def test_bug1055820b():
+    # Corresponds to temp2b.py in the bug report.
+
+    ouch = []
+    def callback(ignored):
+        ouch[:] = [wr() for wr in WRs]
+
+    Cs = [C1055820(i) for i in range(2)]
+    WRs = [weakref.ref(c, callback) for c in Cs]
+    c = None
+
+    gc.collect()
+    expect(len(ouch), 0, "bug1055820b")
+    # Make the two instances trash, and collect again.  The bug was that
+    # the callback materialized a strong reference to an instance, but gc
+    # cleared the instance's dict anyway.
+    Cs = None
+    gc.collect()
+    expect(len(ouch), 2, "bug1055820b")  # else the callbacks didn't run
+    for x in ouch:
+        # If the callback resurrected one of these guys, the instance
+        # would be damaged, with an empty __dict__.
+        expect(x, None, "bug1055820b")
+
+def test_bug1055820c():
+    # Corresponds to temp2c.py in the bug report.  This is pretty elaborate.
+
+    c0 = C1055820(0)
+    # Move c0 into generation 2.
+    gc.collect()
+
+    c1 = C1055820(1)
+    c1.keep_c0_alive = c0
+    del c0.loop # now only c1 keeps c0 alive
+
+    c2 = C1055820(2)
+    c2wr = weakref.ref(c2) # no callback!
+
+    ouch = []
+    def callback(ignored):
+        ouch[:] = [c2wr()]
+
+    # The callback gets associated with a wr on an object in generation 2.
+    c0wr = weakref.ref(c0, callback)
+
+    c0 = c1 = c2 = None
+
+    # What we've set up:  c0, c1, and c2 are all trash now.  c0 is in
+    # generation 2.  The only thing keeping it alive is that c1 points to it.
+    # c1 and c2 are in generation 0, and are in self-loops.  There's a global
+    # weakref to c2 (c2wr), but that weakref has no callback.  There's also
+    # a global weakref to c0 (c0wr), and that does have a callback, and that
+    # callback references c2 via c2wr().
+    #
+    #               c0 has a wr with callback, which references c2wr
+    #               ^
+    #               |
+    #               |     Generation 2 above dots
+    #. . . . . . . .|. . . . . . . . . . . . . . . . . . . . . . . .
+    #               |     Generation 0 below dots
+    #               |
+    #               |
+    #            ^->c1   ^->c2 has a wr but no callback
+    #            |  |    |  |
+    #            <--v    <--v
+    #
+    # So this is the nightmare:  when generation 0 gets collected, we see that
+    # c2 has a callback-free weakref, and c1 doesn't even have a weakref.
+    # Collecting generation 0 doesn't see c0 at all, and c0 is the only object
+    # that has a weakref with a callback.  gc clears c1 and c2.  Clearing c1
+    # has the side effect of dropping the refcount on c0 to 0, so c0 goes
+    # away (despite that it's in an older generation) and c0's wr callback
+    # triggers.  That in turn materializes a reference to c2 via c2wr(), but
+    # c2 gets cleared anyway by gc.
+
+    # We want to let gc happen "naturally", to preserve the distinction
+    # between generations.
+    junk = []
+    i = 0
+    detector = GC_Detector()
+    while not detector.gc_happened:
+        i += 1
+        if i > 10000:
+            raise TestFailed("gc didn't happen after 10000 iterations")
+        expect(len(ouch), 0, "bug1055820c")
+        junk.append([])  # this will eventually trigger gc
+
+    expect(len(ouch), 1, "bug1055820c")  # else the callback wasn't invoked
+    for x in ouch:
+        # If the callback resurrected c2, the instance would be damaged,
+        # with an empty __dict__.
+        expect(x, None, "bug1055820c")
+
+def test_bug1055820d():
+    # Corresponds to temp2d.py in the bug report.  This is very much like
+    # test_bug1055820c, but uses a __del__ method instead of a weakref
+    # callback to sneak in a resurrection of cyclic trash.
+
+    ouch = []
+    class D(C1055820):
+        def __del__(self):
+            ouch[:] = [c2wr()]
+
+    d0 = D(0)
+    # Move all the above into generation 2.
+    gc.collect()
+
+    c1 = C1055820(1)
+    c1.keep_d0_alive = d0
+    del d0.loop # now only c1 keeps d0 alive
+
+    c2 = C1055820(2)
+    c2wr = weakref.ref(c2) # no callback!
+
+    d0 = c1 = c2 = None
+
+    # What we've set up:  d0, c1, and c2 are all trash now.  d0 is in
+    # generation 2.  The only thing keeping it alive is that c1 points to it.
+    # c1 and c2 are in generation 0, and are in self-loops.  There's a global
+    # weakref to c2 (c2wr), but that weakref has no callback.  There are no
+    # other weakrefs.
+    #
+    #               d0 has a __del__ method that references c2wr
+    #               ^
+    #               |
+    #               |     Generation 2 above dots
+    #. . . . . . . .|. . . . . . . . . . . . . . . . . . . . . . . .
+    #               |     Generation 0 below dots
+    #               |
+    #               |
+    #            ^->c1   ^->c2 has a wr but no callback
+    #            |  |    |  |
+    #            <--v    <--v
+    #
+    # So this is the nightmare:  when generation 0 gets collected, we see that
+    # c2 has a callback-free weakref, and c1 doesn't even have a weakref.
+    # Collecting generation 0 doesn't see d0 at all.  gc clears c1 and c2.
+    # Clearing c1 has the side effect of dropping the refcount on d0 to 0, so
+    # d0 goes away (despite that it's in an older generation) and d0's __del__
+    # triggers.  That in turn materializes a reference to c2 via c2wr(), but
+    # c2 gets cleared anyway by gc.
+
+    # We want to let gc happen "naturally", to preserve the distinction
+    # between generations.
+    detector = GC_Detector()
+    junk = []
+    i = 0
+    while not detector.gc_happened:
+        i += 1
+        if i > 10000:
+            raise TestFailed("gc didn't happen after 10000 iterations")
+        expect(len(ouch), 0, "bug1055820d")
+        junk.append([])  # this will eventually trigger gc
+
+    expect(len(ouch), 1, "bug1055820d")  # else __del__ wasn't invoked
+    for x in ouch:
+        # If __del__ resurrected c2, the instance would be damaged, with an
+        # empty __dict__.
+        expect(x, None, "bug1055820d")
+
+
 def test_all():
     gc.collect() # Delete 2nd generation garbage
     run_test("lists", test_list)
@@ -392,6 +578,19 @@ def test_all():
     run_test("boom_new", test_boom_new)
     run_test("boom2_new", test_boom2_new)
     run_test("get_referents", test_get_referents)
+    run_test("bug1055820b", test_bug1055820b)
+
+    gc.enable()
+    try:
+        run_test("bug1055820c", test_bug1055820c)
+    finally:
+        gc.disable()
+
+    gc.enable()
+    try:
+        run_test("bug1055820d", test_bug1055820d)
+    finally:
+        gc.disable()
 
 def test():
     if verbose:

Misc/NEWS

@@ -32,6 +32,17 @@ License Version 2.
 Core and builtins
 -----------------
 
+- Bug #1055820 Cyclic garbage collection was not protecting against that
+  calling a live weakref to a piece of cyclic trash could resurrect an
+  insane mutation of the trash if any Python code ran during gc (via
+  running a dead object's __del__ method, running another callback on a
+  weakref to a dead object, or via any Python code run in any other thread
+  that managed to obtain the GIL while a __del__ or callback was running
+  in the thread doing gc).  The most likely symptom was "impossible"
+  ``AttributeEror`` exceptions, appearing seemingly at random, on weakly
+  referenced objects.  The cure was to clear all weakrefs to unreachable
+  objects before allowing any callbacks to run.
+
 - Bug #1054139 _PyString_Resize() now invalidates its cached hash value.
 
 Extension Modules

Modules/gc_weakref.txt

+Intro
+=====
+
+The basic rule for dealing with weakref callbacks (and __del__ methods too,
+for that matter) during cyclic gc:
+
+    Once gc has computed the set of unreachable objects, no Python-level
+    code can be allowed to access an unreachable object.
+
+If that can happen, then the Python code can resurrect unreachable objects
+too, and gc can't detect that without starting over.  Since gc eventually
+runs tp_clear on all unreachable objects, if an unreachable object is
+resurrected then tp_clear will eventually be called on it (or may already
+have been called before resurrection).  At best (and this has been an
+historically common bug), tp_clear empties an instance's __dict__, and
+"impossible" AttributeErrors result.  At worst, tp_clear leaves behind an
+insane object at the C level, and segfaults result (historically, most
+often by setting a new-style class's mro pointer to NULL, after which
+attribute lookups performed by the class can segfault).
+
+OTOH, it's OK to run Python-level code that can't access unreachable
+objects, and sometimes that's necessary.  The chief example is the callback
+attached to a reachable weakref W to an unreachable object O.  Since O is
+going away, and W is still alive, the callback must be invoked.  Because W
+is still alive, everything reachable from its callback is also reachable,
+so it's also safe to invoke the callback (although that's trickier than it
+sounds, since other reachable weakrefs to other unreachable objects may
+still exist, and be accessible to the callback -- there are lots of painful
+details like this covered in the rest of this file).
+
+Python 2.4/2.3.5
+================
+
+The "Before 2.3.3" section below turned out to be wrong in some ways, but
+I'm leaving it as-is because it's more right than wrong, and serves as a
+wonderful example of how painful analysis can miss not only the forest for
+the trees, but also miss the trees for the aphids sucking the trees
+dry <wink>.
+
+The primary thing it missed is that when a weakref to a piece of cyclic
+trash (CT) exists, then any call to any Python code whatsoever can end up
+materializing a strong reference to that weakref's CT referent, and so
+possibly resurrect an insane object (one for which cyclic gc has called-- or
+will call before it's done --tp_clear()).  It's not even necessarily that a
+weakref callback or __del__ method does something nasty on purpose:  as
+soon as we execute Python code, threads other than the gc thread can run
+too, and they can do ordinary things with weakrefs that end up resurrecting
+CT while gc is running.
+
+    http://www.python.org/sf/1055820
+
+shows how innocent it can be, and also how nasty.  Variants of the three
+focussed test cases attached to that bug report are now part of Python's
+standard Lib/test/test_gc.py.
+
+Jim Fulton gave the best nutshell summary of the new (in 2.4 and 2.3.5)
+approach:
+
+    Clearing cyclic trash can call Python code.  If there are weakrefs to
+    any of the cyclic trash, then those weakrefs can be used to resurrect
+    the objects.  Therefore, *before* clearing cyclic trash, we need to
+    remove any weakrefs.  If any of the weakrefs being removed have
+    callbacks, then we need to save the callbacks and call them *after* all
+    of the weakrefs have been cleared.
+
+Alas, doing just that much doesn't work, because it overlooks what turned
+out to be the much subtler problems that were fixed earlier, and described
+below.  We do clear all weakrefs to CT now before breaking cycles, but not
+all callbacks encountered can be run later.  That's explained in horrid
+detail below.
+
+Older text follows, with a some later comments in [] brackets:
+
+Before 2.3.3
+============
+
 Before 2.3.3, Python's cyclic gc didn't pay any attention to weakrefs.
 Segfaults in Zope3 resulted.
 
@@ -19,12 +95,19 @@ segfaults) can happen right then, during the callback's execution, or can
 happen at any later time if the callback manages to resurrect an insane
 object.
 
+[That missed that, in addition, a weakref to CT can exist outside CT, and
+ any callback into Python can use such a non-CT weakref to resurrect its CT
+ referent.  The same bad kinds of things can happen then.]
+
 Note that if it's possible for the callback to get at objects in the trash
 cycles, it must also be the case that the callback itself is part of the
 trash cycles.  Else the callback would have acted as an external root to
 the current collection, and nothing reachable from it would be in cyclic
 trash either.
 
+[Except that a non-CT callback can also use a non-CT weakref to get at
+ CT objects.]
+
 More, if the callback itself is in cyclic trash, then the weakref to which
 the callback is attached must also be trash, and for the same kind of
 reason:  if the weakref acted as an external root, then the callback could
@@ -47,6 +130,13 @@ cyclic trash, it's defensible to first clear weakrefs with callbacks.  It's
 a feature of Python's weakrefs too that when a weakref goes away, the
 callback (if any) associated with it is thrown away too, unexecuted.
 
+[In 2.4/2.3.5, we first clear all weakrefs to CT objects, whether or not
+ those weakrefs are themselves CT, and whether or not they have callbacks.
+ The callbacks (if any) on non-CT weakrefs (if any) are invoked later,
+ after all weakrefs-to-CT have been cleared.  The callbacks (if any) on CT
+ weakrefs (if any) are never invoked, for the excruciating reasons
+ explained here.]
+
 Just that much is almost enough to prevent problems, by throwing away
 *almost* all the weakref callbacks that could get triggered by gc.  The
 problem remaining is that clearing a weakref with a callback decrefs the
@@ -56,7 +146,15 @@ weakrefs can trigger callbacks attached to other weakrefs, and those
 latter weakrefs may or may not be part of cyclic trash.
 
 So, to prevent any Python code from running while gc is invoking tp_clear()
-on all the objects in cyclic trash, it's not quite enough just to invoke
+on all the objects in cyclic trash,
+
+[That was always wrong:  we can't stop Python code from running when gc
+ is breaking cycles.  If an object with a __del__ method is not itself in
+ a cycle, but is reachable only from CT, then breaking cycles will, as a
+ matter of course, drop the refcount on that object to 0, and its __del__
+ will run right then.  What we can and must stop is running any Python
+ code that could access CT.]
+                                     it's not quite enough just to invoke
 tp_clear() on weakrefs with callbacks first.  Instead the weakref module
 grew a new private function (_PyWeakref_ClearRef) that does only part of
 tp_clear():  it removes the weakref from the weakly-referenced object's list
@@ -65,9 +163,13 @@ _PyWeakref_ClearRef(wr) ensures that wr's callback object will never
 trigger, and (unlike weakref's tp_clear()) also prevents any callback
 associated *with* wr's callback object from triggering.
 
+[Although we may trigger such callbacks later, as explained below.]
+
 Then we can call tp_clear on all the cyclic objects and never trigger
 Python code.
 
+[As above, not so:  it means never trigger Python code that can access CT.]
+
 After we do that, the callback objects still need to be decref'ed.  Callbacks
 (if any) *on* the callback objects that were also part of cyclic trash won't
 get invoked, because we cleared all trash weakrefs with callbacks at the
@@ -76,6 +178,10 @@ acted as external roots to everything reachable from them, so nothing
 reachable from them was part of cyclic trash, so gc didn't do any damage to
 objects reachable from them, and it's safe to call them at the end of gc.
 
+[That's so.  In addition, now we also invoke (if any) the callbacks on
+ non-CT weakrefs to CT objects, during the same pass that decrefs the
+ callback objects.]
+
 An alternative would have been to treat objects with callbacks like objects
 with __del__ methods, refusing to collect them, appending them to gc.garbage
 instead.  That would have been much easier.  Jim Fulton gave a strong
@@ -105,3 +211,9 @@ __del__ methods is to avoid running finalizers in an arbitrary order).
 However, a weakref callback on a weakref callback has got to be rare.
 It's possible to do such a thing, so gc has to be robust against it, but
 I doubt anyone has done it outside the test case I wrote for it.
+
+[The callbacks (if any) on non-CT weakrefs to CT objects are also executed
+ in an arbitrary order now.  But they were before too, depending on the
+ vagaries of when tp_clear() happened to break enough cycles to trigger
+ them.  People simply shouldn't try to use __del__ or weakref callbacks to
+ do fancy stuff.]

Objects/weakrefobject.c

@@ -850,7 +850,9 @@ PyWeakref_GetObject(PyObject *ref)
     return PyWeakref_GET_OBJECT(ref);
 }
 
-
+/* Note that there's an inlined copy-paste of handle_callback() in gcmodule.c's
+ * handle_weakrefs().
+ */
 static void
 handle_callback(PyWeakReference *ref, PyObject *callback)
 {