Re: kernel stack trace using conntrack

[Pablo Neira Ayuso <pablo@netfilter.org>]

Patrick McHardy wrote:
> Ramblewski David wrote:
>> Hi Eric,
>>
>> The conntrack patch works successfully.
>>
>>>> diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c
>>>> index 0ffe689..d2657aa 100644
>>>> --- a/net/netfilter/nf_conntrack_netlink.c
>>>> +++ b/net/netfilter/nf_conntrack_netlink.c
>>>> @@ -923,7 +923,7 @@ ctnetlink_change_status(struct nf_conn *ct, const struct nlattr * const cda[])
>>>>     unsigned int status = ntohl(nla_get_be32(cda[CTA_STATUS]));
>>>>     d = ct->status ^ status;
>>>>
>>>> -   if (d & (IPS_EXPECTED|IPS_CONFIRMED|IPS_DYING))
>>>> +   if (d & (IPS_EXPECTED|IPS_DYING))
>>>>             /* unchangeable */
>>>>             return -EBUSY;
>>> I think that we should explicitly report if the user unsets
>>> IPS_CONFIRMED. Please, don't change this.
>>>
>>> Apart from that, the patch seems fine to me. Thanks!
>> Problem is we now (I mean after my patch) enter
>> ctnetlink_change_status() with ct->status being null (or at least,
>> IPS_CONFIRMED not set)
> 
> Pablo, please let me know whether you want me to apply this.

ctnetlink_change_helper() also calls nf_ct_ext_add() for conntracks that
are confirmed (in case of a helper update for an existing conntrack).
That would also trigger the assertion. If we want to support helper
assignation via ctnetlink for existing conntracks, we will need to add
locking to the conntrack extension infrastructure to avoid races.

I don't see a clear solution for this yet.