netdev - Re: FW: [PATCH bpf-next 2/4] xsk: allow AF_XDP sockets to receive packets directly from a queue

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]

Date:   Fri, 18 Oct 2019 21:22:52 +0200
From:   Toke Høiland-Jørgensen <toke@...hat.com>
To:     "Samudrala\, Sridhar" <sridhar.samudrala@...el.com>,
        Alexei Starovoitov <alexei.starovoitov@...il.com>,
        Jakub Kicinski <jakub.kicinski@...ronome.com>
Cc:     "Karlsson\, Magnus" <magnus.karlsson@...el.com>,
        Björn Töpel 
        <bjorn.topel@...el.com>, Netdev <netdev@...r.kernel.org>,
        "bpf\@vger.kernel.org" <bpf@...r.kernel.org>,
        intel-wired-lan <intel-wired-lan@...ts.osuosl.org>,
        "Fijalkowski\, Maciej" <maciej.fijalkowski@...el.com>,
        "Herbert\, Tom" <tom.herbert@...el.com>
Subject: Re: FW: [PATCH bpf-next 2/4] xsk: allow AF_XDP sockets to receive packets directly from a queue

"Samudrala, Sridhar" <sridhar.samudrala@...el.com> writes:

> Performance Results
> ===================
> Only 1 core is used in all these testcases as the app and the queue irq are pinned to the same core.
>
> ----------------------------------------------------------------------------------
>                                 mitigations ON                mitigations OFF
>    Testcase              ----------------------------------------------------------
>                          no patches    with patches       no patches   with patches
> ----------------------------------------------------------------------------------
> AF_XDP default rxdrop        X             X                   Y            Y

Is this really exactly the same with and without patches? You're adding
an extra check to xdp_do_redirect(); are you really saying that the
impact of that is zero?

> AF_XDP direct rxdrop        N/A          X+46%                N/A         Y+25%
> Kernel rxdrop              X+61%         X+61%               Y+53%        Y+53%
> ----------------------------------------------------------------------------------
>
> Here Y is pps with CPU security mitigations turned OFF and it is 26%
> better than X.

Any particular reason you're not sharing the values of X and Y?

-Toke