Re: [PATCH] module: Fix race condition between load and unload module
From: Anatol Pomozov
Date: Fri Apr 12 2013 - 20:05:10 EST
On Fri, Apr 12, 2013 at 4:53 PM, Greg Kroah-Hartman
> On Fri, Apr 12, 2013 at 04:47:50PM -0700, Linus Torvalds wrote:
>> On Fri, Apr 12, 2013 at 3:32 PM, Anatol Pomozov
>> <anatol.pomozov@xxxxxxxxx> wrote:
>> > Here is timeline for the crash in case if kset_find_obj() searches for
>> > an object tht nobody holds and other thread is doing kobject_put()
>> > on the same kobject:
>> > THREAD A (calls kset_find_obj()) THREAD B (calls kobject_put())
>> > splin_lock()
>> > atomic_dec_return(kobj->kref), counter gets zero here
>> > ... starts kobject cleanup ....
>> > spin_lock() // WAIT thread A in kobj_kset_leave()
>> > iterate over kset->list
>> > atomic_inc(kobj->kref) (counter becomes 1)
>> > spin_unlock()
>> > spin_lock() // taken
>> > // it does not know that thread A increased counter so it
>> > remove obj from list
>> > spin_unlock()
>> > vfree(module) // frees module object with containing kobj
>> > // kobj points to freed memory area!!
>> > koubject_put(kobj) // OOPS!!!!
>> This is a much more generic bug in kobjects, and I would hate to add
>> some random workaround for just one case of this bug like you do. The
>> more fundamental bug needs to be fixed too.
>> I think the more fundamental bugfix is to just fix kobject_get() to
>> return NULL if the refcount was zero, because in that case the kobject
>> no longer really exists.
>> So instead of having
>> it should do
>> if (!atomic_inc_not_zero(&kobj->kref.refcount))
>> kobj = NULL;
>> and I think that should fix your race automatically, no? Proper patch
>> attached (but TOTALLY UNTESTED - it seems to compile, though).
>> The problem is that we lose the warning for when the refcount is zero
>> and somebody does a kobject_get(), but that is ok *assuming* that
>> people actually check the return value of kobject_get() rather than
>> just "know" that if they passed in a non-NULL kobj, they'll get it
>> right back.
>> Greg - please take a look... I'm adding Al to the discussion too,
>> because Al just *loooves* these kinds of races ;)
> We "should" have some type of "higher-up" lock to prevent the
> release/get races from happening, we have that in the driver core, and I
> thought we had such a lock already in the module subsystem as well,
> which will prevent any of this from being needed.
> Rusty, don't we have a lock for this somewhere?
> Linus, I think your patch will reduce the window the race could happen,
> but it should still be there, although testing with it would be
> interesting to see if the original problem can be triggered with it.
Linus patch should fix the module race condition. vfree(module) cannot
be called while we keep kobj->kset->lock. vfree() is called in
THREAD_B only after it acquires lock, removes kobj from list. So if
kobj is found by THREAD_A in kset->list and we did not release lock
then memory is not freed.
> I'll look at it some more tomorrow, about to go to dinner now...
> greg k-h
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/