Message-ID: <CAMDZJNV3-y5jkUAJJ--10PcicKpGMwKS_3gG9O7srjomO3begw@mail.gmail.com>
Date:   Sat, 11 Dec 2021 01:37:55 +0800
From:   Tonghao Zhang <xiangxia.m.yue@...il.com>
To:     John Fastabend <john.fastabend@...il.com>
Cc:     Linux Kernel Network Developers <netdev@...r.kernel.org>,
        "David S. Miller" <davem@...emloft.net>,
        Jakub Kicinski <kuba@...nel.org>,
        Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Andrii Nakryiko <andrii@...nel.org>,
        Martin KaFai Lau <kafai@...com>,
        Song Liu <songliubraving@...com>, Yonghong Song <yhs@...com>,
        KP Singh <kpsingh@...nel.org>,
        Eric Dumazet <edumazet@...gle.com>,
        Antoine Tenart <atenart@...nel.org>,
        Alexander Lobakin <alexandr.lobakin@...el.com>,
        Wei Wang <weiwan@...gle.com>, Arnd Bergmann <arnd@...db.de>
Subject: Re: [net v5 2/3] net: sched: add check tc_skip_classify in sch egress

On Sat, Dec 11, 2021 at 12:43 AM John Fastabend
<john.fastabend@...il.com> wrote:
>
> xiangxia.m.yue@ wrote:
> > From: Tonghao Zhang <xiangxia.m.yue@...il.com>
> >
> > Try to resolve the issues as below:
> > * We look up and then check tc_skip_classify flag in net
> >   sched layer, even though skb don't want to be classified.
> >   That case may consume a lot of cpu cycles. This patch
> >   is useful when there are a lot of filters with different
> >   prio. There is ~5 prio in in production, ~1% improvement.
> >
> >   Rules as below:
> >   $ for id in $(seq 1 5); do
> >   $       tc filter add ... egress prio $id ... action mirred egress redirect dev ifb0
> >   $ done
> >
> > * bpf_redirect may be invoked in egress path. If we don't
> >   check the flags and then return immediately, the packets
> >   will loopback.
>
> This would be the naive case right? Meaning the BPF program is
> doing a redirect without any logic or is buggy?
>
> Can you map out how this happens for me, I'm not fully sure I
> understand the exact concern. Is it possible for BPF programs
> that used to see packets no longer see the packet as expected?
>
> Is this the path you are talking about?
Hi John
Tx ethx -> __dev_queue_xmit -> sch_handle_egress
->  execute BPF program on ethx with bpf_redirect(ifb0) ->
-> ifb_xmit -> ifb_ri_tasklet -> dev_queue_xmit -> __dev_queue_xmit
the packets loopbacks, that means bpf_redirect doesn't work with ifb
netdev, right ?
so in sch_handle_egress, I add the check skb_skip_tc_classify().

>  rx ethx  ->
>    execute BPF program on ethx with bpf_redirect(ifb0) ->
>      __skb_dequeue @ifb tc_skip_classify = 1 ->
>        dev_queue_xmit() ->
>           sch_handle_egress() ->
>             execute BPF program again
>
> I can't see why you want to skip that second tc BPF program,
> or for that matter any tc filter there. In general how do you
> know that is the correct/expected behavior? Before the above
> change it would have been called, what if its doing useful
> work.
bpf_redirect works fine on ingress with ifb
__netif_receive_skb_core -> sch_handle_ingress -> bpf_redirect (ifb0)
-> ifb_xmit -> netif_receive_skb -> __netif_receive_skb_core
but
__netif_receive_skb_core --> skb_skip_tc_classify(so the packets will
execute the BPF progam again)

> Also its not clear how your ifb setup is built or used. That
> might help understand your use case. I would just remove the
> IFB altogether and the above discussion is mute.
tc filter add dev veth1 egress bpf direct-action obj
test_bpf_redirect_ifb.o sec redirect_ifb

the test_bpf_redirect_ifb  bpf progam:
+// SPDX-License-Identifier: GPL-2.0
+/* Copyright (c) 2021 DiDi Global */
+
+#include <linux/bpf.h>
+#include <bpf/bpf_helpers.h>
+
+SEC("redirect_ifb")
+int redirect(struct __sk_buff *skb)
+{
+       return bpf_redirect(skb->ifindex + 1 /* ifbX */, 0);
+}
+
+char __license[] SEC("license") = "GPL";

The 3/3 is selftest:
https://patchwork.kernel.org/project/netdevbpf/patch/20211208145459.9590-4-xiangxia.m.yue@gmail.com/

> Thanks,
> John



-- 
Best regards, Tonghao

Powered by blists - more mailing lists