netdev - Re: [RFC PATCH net-next 0/3] TC: Introduce qevents

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: <877dwxvgzk.fsf@mellanox.com>
Date:   Wed, 27 May 2020 11:56:15 +0200
From:   Petr Machata <petrm@...lanox.com>
To:     Cong Wang <xiyou.wangcong@...il.com>
Cc:     Linux Kernel Network Developers <netdev@...r.kernel.org>,
        Jakub Kicinski <kuba@...nel.org>,
        Eric Dumazet <eric.dumazet@...il.com>,
        Jamal Hadi Salim <jhs@...atatu.com>,
        Jiri Pirko <jiri@...lanox.com>,
        Ido Schimmel <idosch@...lanox.com>
Subject: Re: [RFC PATCH net-next 0/3] TC: Introduce qevents


Cong Wang <xiyou.wangcong@...il.com> writes:

> On Tue, May 26, 2020 at 10:11 AM Petr Machata <petrm@...lanox.com> wrote:
>>
>> The Spectrum hardware allows execution of one of several actions as a
>> result of queue management events: tail-dropping, early-dropping, marking a
>> packet, or passing a configured latency threshold or buffer size. Such
>> packets can be mirrored, trapped, or sampled.
>>
>> Modeling the action to be taken as simply a TC action is very attractive,
>> but it is not obvious where to put these actions. At least with ECN marking
>> one could imagine a tree of qdiscs and classifiers that effectively
>> accomplishes this task, albeit in an impractically complex manner. But
>> there is just no way to match on dropped-ness of a packet, let alone
>> dropped-ness due to a particular reason.
>>
>> To allow configuring user-defined actions as a result of inner workings of
>> a qdisc, this patch set introduces a concept of qevents. Those are attach
>> points for TC blocks, where filters can be put that are executed as the
>> packet hits well-defined points in the qdisc algorithms. The attached
>> blocks can be shared, in a manner similar to clsact ingress and egress
>> blocks, arbitrary classifiers with arbitrary actions can be put on them,
>> etc.
>
> This concept does not fit well into qdisc, essentially you still want to
> install filters (and actions) somewhere on qdisc, but currently all filters
> are executed at enqueue, basically you want to execute them at other
> pre-defined locations too, for example early drop.
>
> So, perhaps adding a "position" in tc filter is better? Something like:
>
> tc qdisc add dev x root handle 1: ... # same as before
> tc filter add dev x parent 1:0 position early_drop matchall action....

Position would just be a chain index.

> And obviously default position must be "enqueue". Makes sense?

Chain 0.

So if I understand the proposal correctly: a qdisc has a classification
block (cl_ops->tcf_block). Qevents then reference a chain on that
classification block.

If the above is correct, I disagree that this is a better model. RED
does not need to classify anything, modelling this as a classification
block is wrong. It should be another block. (Also shareable, because as
an operator you likely want to treat all, say, early drops the same, and
therefore to add just one rule for all 128 or so of your qdiscs.)