netdev - Re: [Patch bpf-next v5 06/11] sock: introduce sk->sk_prot->psock_update_sk

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: <CAM_iQpVYqcYCA97KTx6bRAEkkO4gpy_t8YCGTUQ2XRDnJ=-sFw@mail.gmail.com>
Date:   Thu, 18 Mar 2021 09:40:40 -0700
From:   Cong Wang <xiyou.wangcong@...il.com>
To:     Alexander Lobakin <alobakin@...me>
Cc:     bpf <bpf@...r.kernel.org>, duanxiongchun@...edance.com,
        Dongdong Wang <wangdongdong.6@...edance.com>,
        Jiang Wang <jiang.wang@...edance.com>,
        Cong Wang <cong.wang@...edance.com>,
        John Fastabend <john.fastabend@...il.com>,
        Daniel Borkmann <daniel@...earbox.net>,
        Jakub Sitnicki <jakub@...udflare.com>,
        Lorenz Bauer <lmb@...udflare.com>,
        Linux Kernel Network Developers <netdev@...r.kernel.org>
Subject: Re: [Patch bpf-next v5 06/11] sock: introduce sk->sk_prot->psock_update_sk_prot()

On Thu, Mar 18, 2021 at 5:09 AM Alexander Lobakin <alobakin@...me> wrote:
> Regarding that both {tcp,udp}_bpf_update_proto() is global and
> for now they are the only two implemented callbacks, wouldn't it
> be worthy to straighten the calls here? Like
>
>         return INDIRECT_CALL_2(sk->sk_prot->psock_update_sk_prot,
>                                tcp_bpf_update_proto,
>                                udp_bpf_update_proto,
>                                sk, false);

I get your point, but AF_UNIX will implement this in the next patchset,
and my colleague is working on vsock support too, so it will go beyond
INET very soon.

>
> (the same in sk_psock_restore_proto() then)
>
> Or this code path is not performance-critical?

It is not on the hot path, updating proto happens when we insert
the socket to the map or remove it from there.

Thanks.