netdev - Re: [PATCH net-next v3 0/7] Allow offloading of UDP NEW connections via act

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <87fsc5bsh9.fsf@nvidia.com>
Date:   Fri, 20 Jan 2023 08:57:16 +0200
From:   Vlad Buslov <vladbu@...dia.com>
To:     Marcelo Ricardo Leitner <marcelo.leitner@...il.com>
CC:     <davem@...emloft.net>, <kuba@...nel.org>, <pabeni@...hat.com>,
        <pablo@...filter.org>, <netdev@...r.kernel.org>,
        <netfilter-devel@...r.kernel.org>, <jhs@...atatu.com>,
        <xiyou.wangcong@...il.com>, <jiri@...nulli.us>, <ozsh@...dia.com>,
        <simon.horman@...igine.com>
Subject: Re: [PATCH net-next v3 0/7] Allow offloading of UDP NEW connections
 via act_ct

On Fri 20 Jan 2023 at 08:38, Vlad Buslov <vladbu@...dia.com> wrote:
> On Thu 19 Jan 2023 at 18:37, Marcelo Ricardo Leitner <marcelo.leitner@...il.com> wrote:
>> On Thu, Jan 19, 2023 at 08:50:57PM +0100, Vlad Buslov wrote:
>>> Currently only bidirectional established connections can be offloaded
>>> via act_ct. Such approach allows to hardcode a lot of assumptions into
>>> act_ct, flow_table and flow_offload intermediate layer codes. In order
>>> to enabled offloading of unidirectional UDP NEW connections start with
>>> incrementally changing the following assumptions:
>>> 
>>> - Drivers assume that only established connections are offloaded and
>>>   don't support updating existing connections. Extract ctinfo from meta
>>>   action cookie and refuse offloading of new connections in the drivers.
>>
>> Hi Vlad,
>>
>> Regarding ct_seq_show(). When dumping the CT entries today, it will do
>> things like:
>>
>>         if (!test_bit(IPS_OFFLOAD_BIT, &ct->status))
>>                 seq_printf(s, "%ld ", nf_ct_expires(ct)  / HZ);
>>
>> omit the timeout, which is okay with this new patchset, but then:
>>
>>         if (test_bit(IPS_HW_OFFLOAD_BIT, &ct->status))
>>                 seq_puts(s, "[HW_OFFLOAD] ");
>>         else if (test_bit(IPS_OFFLOAD_BIT, &ct->status))
>>                 seq_puts(s, "[OFFLOAD] ");
>>         else if (test_bit(IPS_ASSURED_BIT, &ct->status))
>>                 seq_puts(s, "[ASSURED] ");
>>
>> Previously, in order to be offloaded, it had to be Assured. But not
>> anymore after this patchset. Thoughts?
>
> Hi Marcelo,
>
> I know that for some reason offloaded entries no longer display
> 'assured' flag in the dump. This could be changed, but I don't have a
> preference either way and this patch set doesn't modify the behavior.
> Up to you and maintainers I guess.

BTW after checking the log I don't think the assumption that all
offloaded connections are always assured is true. As far as I understand
act_ct originally offloaded established connections and change to
offload assured was made relatively recently in 43332cf97425
("net/sched: act_ct: Offload only ASSURED connections") without
modifying the prints you mentioned.