Re: WARNING in strp_data_ready

From: Dmitry Vyukov
Date: Thu Dec 28 2017 - 13:35:29 EST


On Thu, Dec 28, 2017 at 7:21 PM, Ozgur <ozgur@xxxxxxxxxx> wrote:
>
>
> 28.12.2017, 19:33, "Dmitry Vyukov" <dvyukov@xxxxxxxxxx>:
>> On Thu, Dec 28, 2017 at 5:14 PM, Tom Herbert <tom@xxxxxxxxxxxxxxx> wrote:
>>> On Thu, Dec 28, 2017 at 12:59 AM, Ozgur <ozgur@xxxxxxxxxx> wrote:
>>>> 28.12.2017, 04:19, "Tom Herbert" <tom@xxxxxxxxxxxxxxx>:
>>>>> On Wed, Dec 27, 2017 at 12:20 PM, Ozgur <ozgur@xxxxxxxxxx> wrote:
>>>>>> 27.12.2017, 23:14, "Dmitry Vyukov" <dvyukov@xxxxxxxxxx>:
>>>>>>> On Wed, Dec 27, 2017 at 9:08 PM, Ozgur <ozgur@xxxxxxxxxx> wrote:
>>>>>>>> 27.12.2017, 22:21, "Dmitry Vyukov" <dvyukov@xxxxxxxxxx>:
>>>>>>>>> On Wed, Dec 27, 2017 at 8:09 PM, Tom Herbert <tom@xxxxxxxxxxxxxxx> wrote:
>>>>>>>>>> Did you try the patch I posted?
>>>>>>>>>
>>>>>>>>> Hi Tom,
>>>>>>>>
>>>>>>>> Hello Dmitry,
>>>>>>>>
>>>>>>>>> No. And I didn't know I need to. Why?
>>>>>>>>> If you think the patch needs additional testing, you can ask syzbot to
>>>>>>>>> test it. See https://github.com/google/syzkaller/blob/master/docs/syzbot.md#communication-with-syzbot
>>>>>>>>> Otherwise proceed with committing it. Or what are we waiting for?
>>>>>>>>>
>>>>>>>>> Thanks
>>>>>>>>
>>>>>>>> I think we need to fixed patch for crash, in fact check to patch code and test solve the bug.
>>>>>>>> How do test it because there is no patch in the following bug?
>>>>>>>
>>>>>>> Hi Ozgur,
>>>>>>>
>>>>>>> I am not sure I completely understand what you mean. But the
>>>>>>> reproducer for this bug (which one can use for testing) is here:
>>>>>>> https://groups.google.com/forum/#!topic/syzkaller-bugs/Kxs05ziCpgY
>>>>>>> Tom also mentions there is some patch for this, but I don't know where
>>>>>>> it is, it doesn't seem to be referenced from this thread.
>>>>>>
>>>>>> Hello Dmitry,
>>>>>>
>>>>>> Ah, I'm sorry I don't seen Tom mail and I don't have a patch not tested :)
>>>>>> I think Tom send patch to only you and are you tested?
>>>>>>
>>>>>> kcmsock.c will change and strp_data_ready I think locked.
>>>>>>
>>>>>> Tom, please send a patch for me? I can test and inform you.
>>>>>
>>>>> Hi Ozgur,
>>>>>
>>>>> I reposted the patches as RFC "kcm: Fix lockdep issue". Please test if you can!
>>>>>
>>>>> Thanks,
>>>>> Tom
>>>>
>>>> Hello Tom,
>>>>
>>>> Which are you use the repos? I pulled but I don't seen this patches.
>>> They are not in any public repo yet. I posted the patches to netdev
>>> list so they can be reviewed and tested by third parties. Posting
>>> patches to the list a normal path to get patches into the kernel
>>> (http://nickdesaulniers.github.io/blog/2017/05/16/submitting-your-first-patch-to-the-linux-kernel-and-responding-to-feedback/).
>>>
>>> These patches were applied to net-next but are simple enough that they
>>> should apply to other branches. I will repost and target to net per
>>> Dave's directive once they are verified to fix the issue.
>
> Hello,
>
> thanks Tom and I have tested the fixed patch for linux-next builds and don't have to kernel panic. when nocheck funcs call sk_lock.owned and kernel doesn't give a panic. I have compiled and uploaded next-kernel.
>
> Dmitry,
> could you test it on linux-next?

If you are trying to test how many times I can repeat this, I can
repeat this lots of times:

If you think the patch needs additional testing, you can ask syzbot to
test it. See https://github.com/google/syzkaller/blob/master/docs/syzbot.md#communication-with-syzbot