netdev - Re: [PATCH bpf-next v2 0/8] Simplify xdp_do_redirect_map()/xdp_do_flush

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: <20191219081815.0b07de1a@carbon>
Date:   Thu, 19 Dec 2019 08:18:15 +0100
From:   Jesper Dangaard Brouer <jbrouer@...hat.com>
To:     Björn Töpel <bjorn.topel@...il.com>
Cc:     netdev@...r.kernel.org, ast@...nel.org, daniel@...earbox.net,
        bpf@...r.kernel.org, davem@...emloft.net,
        jakub.kicinski@...ronome.com, hawk@...nel.org,
        john.fastabend@...il.com, magnus.karlsson@...el.com,
        jonathan.lemon@...il.com
Subject: Re: [PATCH bpf-next v2 0/8] Simplify
 xdp_do_redirect_map()/xdp_do_flush_map() and XDP maps

On Thu, 19 Dec 2019 07:09:58 +0100
Björn Töpel <bjorn.topel@...il.com> wrote:

>   $ sudo ./xdp_redirect_cpu --dev enp134s0f0 --cpu 22 xdp_cpu_map0
                                                        ^^^^^^^^^^^^
>   
>   Running XDP/eBPF prog_name:xdp_cpu_map5_lb_hash_ip_pairs
                     ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>   XDP-cpumap      CPU:to  pps            drop-pps    extra-info
>   XDP-RX          20      7723038        0           0
>   XDP-RX          total   7723038        0
[...]

Talking about how to invoke the 'xdp_redirect_cpu' program, I notice
that you are using BPF-prog named: 'xdp_cpu_map5_lb_hash_ip_pairs'
(default) but on cmdline it looks like you want 'xdp_cpu_map0'. 
You need to use '--prog xdp_cpu_map0'.  It will make a HUGE performance
difference.

Like:
  sudo ./xdp_redirect_cpu --dev mlx5p1 --cpu 22 --prog xdp_cpu_map0



(p.s. The load-balance hash_ip_pairs is the default, because it is
usable for solving the most common issue, where the NIC is not
RSS distributing traffic correctly, e.g. ixgbe Q-in-Q. And it is close
to the Suricata approach.)
-- 
Best regards,
  Jesper Dangaard Brouer
  MSc.CS, Principal Kernel Engineer at Red Hat
  LinkedIn: http://www.linkedin.com/in/brouer