Re: [RFC Patch net-next] net_sched: introduce eBPF based Qdisc

From: Martin KaFai Lau <kafai@fb.com>
To: "Toke Høiland-Jørgensen" <toke@redhat.com>
Cc: John Fastabend <john.fastabend@gmail.com>,
	Cong Wang <xiyou.wangcong@gmail.com>,
	Linux Kernel Network Developers <netdev@vger.kernel.org>,
	bpf <bpf@vger.kernel.org>, Cong Wang <cong.wang@bytedance.com>,
	Jamal Hadi Salim <jhs@mojatatu.com>,
	Jiri Pirko <jiri@resnulli.us>
Subject: Re: [RFC Patch net-next] net_sched: introduce eBPF based Qdisc
Date: Thu, 9 Sep 2021 23:55:35 -0700	[thread overview]
Message-ID: <20210910065535.vtwafxy2a7boipqg@kafai-mbp.dhcp.thefacebook.com> (raw)
In-Reply-To: <87zgstra6j.fsf@toke.dk>

On Fri, Sep 03, 2021 at 04:44:04PM +0200, Toke Høiland-Jørgensen wrote:
> Martin KaFai Lau <kafai@fb.com> writes:
> 
> > On Fri, Sep 03, 2021 at 12:27:52AM +0200, Toke Høiland-Jørgensen wrote:
> >> >> The question is if it's useful to provide the full struct_ops for
> >> >> qdiscs? Having it would allow a BPF program to implement that interface
> >> >> towards userspace (things like statistics, classes etc), but the
> >> >> question is if anyone is going to bother with that given the wealth of
> >> >> BPF-specific introspection tools already available?
> > Instead of bpftool can only introspect bpf qdisc and the existing tc
> > can only introspect kernel qdisc,  it will be nice to have bpf
> > qdisc work as other qdisc and showing details together with others
> > in tc.  e.g. a bpf qdisc export its data/stats with its btf-id
> > to tc and have tc print it out in a generic way?
> 
> I'm not opposed to the idea, certainly. I just wonder if people who go
> to the trouble of writing a custom qdisc in BPF will feel it's worth it
> to do the extra work to make this available via a second API. We could
> certainly encourage it, and some things are easy (drop and pkt counters,
> etc), but other things (like class stats) will depend on the semantics
> of the qdisc being implemented, so will require extra work from the BPF
> qdisc developer...
Right, different qdisc has different stats, I think it is currently
stored in qdisc_priv()?  When a qdisc is created, a separate priv is
created together.

Yes, the bpf qdisc prog can store its stats to a bpf map, but then when the
same prog attached to different qdiscs, it has to create different stats maps?

Also, instead of ->enqueue() itself is a bpf prog,
having an ->enqueue() preparing a bpf ctx (zeroing, assigning...etc) and
then make another call to a bpf prog will all add some costs.

That said, I still think it needs a bpf skb map that can queue/dequeue
skb first.  Then it will become possible to prototype different interface
ideas.