From: Kuniyuki Iwashima <kuniyu@amazon.com>
To: <lmb@isovalent.com>
Cc: <andrii@kernel.org>, <ast@kernel.org>, <bpf@vger.kernel.org>,
<daniel@iogearbox.net>, <davem@davemloft.net>,
<dsahern@kernel.org>, <edumazet@google.com>, <haoluo@google.com>,
<hemanthmalla@gmail.com>, <joe@wand.net.nz>,
<john.fastabend@gmail.com>, <jolsa@kernel.org>,
<kpsingh@kernel.org>, <kuba@kernel.org>, <kuniyu@amazon.com>,
<linux-kernel@vger.kernel.org>, <linux-kselftest@vger.kernel.org>,
<martin.lau@linux.dev>, <mykolal@fb.com>,
<netdev@vger.kernel.org>, <pabeni@redhat.com>, <sdf@google.com>,
<shuah@kernel.org>, <song@kernel.org>,
<willemdebruijn.kernel@gmail.com>, <yhs@fb.com>
Subject: Re: [PATCH bpf-next v2 3/6] net: remove duplicate reuseport_lookup functions
Date: Tue, 20 Jun 2023 11:31:23 -0700 [thread overview]
Message-ID: <20230620183123.74585-1-kuniyu@amazon.com> (raw)
In-Reply-To: <CAN+4W8ge-ZQjins-E1=GHDnsi9myFqt7pwNqMkUQHZOPHQhFvQ@mail.gmail.com>
From: Lorenz Bauer <lmb@isovalent.com>
Date: Tue, 20 Jun 2023 15:26:05 +0100
> On Tue, Jun 13, 2023 at 7:57 PM Kuniyuki Iwashima <kuniyu@amazon.com> wrote:
> >
> > The assignment to result below is buggy. Let's say SO_REUSEPROT group
> > have TCP_CLOSE and TCP_ESTABLISHED sockets.
> >
> > 1. Find TCP_CLOSE sk and do SO_REUSEPORT lookup
> > 2. result is not NULL, but the group has TCP_ESTABLISHED sk
> > 3. result = result
> > 4. Find TCP_ESTABLISHED sk, which has a higher score
> > 5. result = result (TCP_CLOSE) <-- should be sk.
> >
> > Same for v6 function.
>
> Thanks for your explanation, I think I get it now. I misunderstood
> that you were worried about returning TCP_ESTABLISHED instead of
> TCP_CLOSE, but it's exactly the other way around.
>
> I have a follow up question regarding the existing code:
>
> result = lookup_reuseport(net, sk, skb,
> saddr, sport, daddr, hnum);
> /* Fall back to scoring if group has connections */
> if (result && !reuseport_has_conns(sk))
> return result;
>
> result = result ? : sk;
> badness = score;
>
> Assuming that result != NULL but reuseport_has_conns() == true, we use
> the reuseport socket as the result, but assign the score of sk to
> badness. Shouldn't we use the score of the reuseport socket?
Good point. This is based on an assumption that all SO_REUSEPORT
sockets have the same score, which is wrong for two corner cases
if reuseport_has_conns() == true :
1) SO_INCOMING_CPU is set
-> selected sk might have +1 score
2) BPF prog returns ESTABLISHED and/or SO_INCOMING_CPU sk
-> selected sk will have more than 8
Using the old score could trigger more lookups depending on the
order that sockets are created.
sk -> sk (SO_INCOMING_CPU) -> sk (ESTABLISHED)
| |
`-> select the next SO_INCOMING_CPU sk
|
`-> select itself (We should save this lookup)
So, yes, we should update badness like
if (unlikely(result)) {
badness = compute_score(result, ...);
} else {
result = sk;
badness = score;
}
next prev parent reply other threads:[~2023-06-20 18:31 UTC|newest]
Thread overview: 22+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-06-13 10:14 [PATCH bpf-next v2 0/6] Add SO_REUSEPORT support for TC bpf_sk_assign Lorenz Bauer
2023-06-13 10:14 ` [PATCH bpf-next v2 1/6] net: export inet_lookup_reuseport and inet6_lookup_reuseport Lorenz Bauer
2023-06-13 10:14 ` [PATCH bpf-next v2 2/6] net: document inet[6]_lookup_reuseport sk_state requirements Lorenz Bauer
2023-06-13 10:14 ` [PATCH bpf-next v2 3/6] net: remove duplicate reuseport_lookup functions Lorenz Bauer
2023-06-13 15:32 ` Simon Horman
2023-06-14 15:42 ` Lorenz Bauer
2023-06-15 7:21 ` Simon Horman
2023-06-13 17:26 ` kernel test robot
2023-06-13 18:56 ` Kuniyuki Iwashima
2023-06-14 15:25 ` Lorenz Bauer
2023-06-14 16:52 ` Kuniyuki Iwashima
2023-06-20 14:26 ` Lorenz Bauer
2023-06-20 18:31 ` Kuniyuki Iwashima [this message]
2023-06-21 8:01 ` Lorenz Bauer
2023-06-21 13:49 ` Lorenz Bauer
2023-06-21 15:00 ` Kuniyuki Iwashima
2023-06-13 10:14 ` [PATCH bpf-next v2 4/6] net: remove duplicate sk_lookup helpers Lorenz Bauer
2023-06-13 19:03 ` Kuniyuki Iwashima
2023-06-13 19:41 ` kernel test robot
2023-06-13 10:15 ` [PATCH bpf-next v2 5/6] bpf, net: Support SO_REUSEPORT sockets with bpf_sk_assign Lorenz Bauer
2023-06-13 17:26 ` kernel test robot
2023-06-13 10:15 ` [PATCH bpf-next v2 6/6] selftests/bpf: Test that SO_REUSEPORT can be used with sk_assign helper Lorenz Bauer
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230620183123.74585-1-kuniyu@amazon.com \
--to=kuniyu@amazon.com \
--cc=andrii@kernel.org \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=davem@davemloft.net \
--cc=dsahern@kernel.org \
--cc=edumazet@google.com \
--cc=haoluo@google.com \
--cc=hemanthmalla@gmail.com \
--cc=joe@wand.net.nz \
--cc=john.fastabend@gmail.com \
--cc=jolsa@kernel.org \
--cc=kpsingh@kernel.org \
--cc=kuba@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=lmb@isovalent.com \
--cc=martin.lau@linux.dev \
--cc=mykolal@fb.com \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
--cc=sdf@google.com \
--cc=shuah@kernel.org \
--cc=song@kernel.org \
--cc=willemdebruijn.kernel@gmail.com \
--cc=yhs@fb.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).