netdev - [RFC Patch v5 0/5] net_sched: introduce eBPF based Qdisc

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>] [day] [month] [year] [list]

Message-ID: <95092865.3867028.1663832781195.JavaMail.zimbra@sjtu.edu.cn>
Date:   Thu, 22 Sep 2022 15:46:21 +0800 (CST)
From:   杨培灏 <yangpeihao@...u.edu.cn>
To:     sdf@...gle.com, YrYj3LPaHV7thgJW@...gle.com
Cc:     bpf@...r.kernel.org, cong.wang@...edance.com, jhs@...atatu.com,
        jiri@...nulli.us, netdev@...r.kernel.org, toke@...hat.com,
        xiyou.wangcong@...il.com
Subject: [RFC Patch v5 0/5] net_sched: introduce eBPF based Qdisc

On 06/24, sdf wrote:

> I feel like this patch series (even rfc), should also have a good example
to show that bpf qdisc is on par and can be used to at least implement
existing policies.

I recently build an example for the simplest `pfifo` qdisc based on Cong Wang's patches, 
see https://github.com/Forsworns/sch_bpf/
if you have interests. The sample is a little large, and involving modifications on libbpf, 
so I didn't attach it as a patch here. 

I also run a micro-benchmark with iperf3, see 
https://gist.github.com/Forsworns/71b70ed67d2ac53a4316bac862ce7d0f

I know it's unfair to evaluate the sch_bpf with `pfifo`, since sch_bpf is based on rbtree.
But I think it a good example for other classiful and more complicated qdiscs.
The `pfifo` is simple enough as a start :)