From mboxrd@z Thu Jan 1 00:00:00 1970 From: Daniel Borkmann Subject: Re: [PATCH] samples/bpf: Add xdp_sample_pkts example Date: Mon, 4 Jun 2018 14:31:34 +0200 Message-ID: <672f2d99-f44d-7605-7c07-e9b6315f0bcd@iogearbox.net> References: <20180530164524.21760-1-toke@toke.dk> <87a7sgcg7i.fsf@toke.dk> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Cc: Networking To: =?UTF-8?Q?Toke_H=c3=b8iland-J=c3=b8rgensen?= , Song Liu Return-path: Received: from www62.your-server.de ([213.133.104.62]:52504 "EHLO www62.your-server.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752399AbeFDMbh (ORCPT ); Mon, 4 Jun 2018 08:31:37 -0400 In-Reply-To: Content-Language: en-US Sender: netdev-owner@vger.kernel.org List-ID: On 06/02/2018 06:22 AM, Daniel Borkmann wrote: > On 05/31/2018 11:44 AM, Toke Høiland-Jørgensen wrote: >> Song Liu writes: >> >>> On Wed, May 30, 2018 at 9:45 AM, Toke Høiland-Jørgensen wrote: >>>> This adds an example program showing how to sample packets from XDP using >>>> the perf event buffer. The example userspace program just prints the >>>> ethernet header for every packet sampled. >>>> >>>> Most of the userspace code is borrowed from other examples, most notably >>>> trace_output. >>>> >>>> Note that the example only works when everything runs on CPU0; so >>>> suitable smp_affinity needs to be set on the device. Some drivers seem >>>> to reset smp_affinity when loading an XDP program, so it may be >>>> necessary to change it after starting the example userspace program. >>> >>> Why does this only works when everything runs on CPU0? Is this >>> something we can improve? >> >> Yeah, good question. Basically, the call from XDP to >> bpf_perf_event_output() will fail with -EOPNOTSUPP. I tracked this down >> to this if statement in __bpf_perf_event_output() in bpf_trace.c: >> >>> if (unlikely(event->oncpu != cpu)) >>> return -EOPNOTSUPP; >> >> I *think* that the way to fix this is for the userspace program to open >> a perf file descriptor for each CPU in the system and poll all of them, >> in which case the XDP program can pass the BPF_F_CURRENT_CPU flag to >> access the right one. > That is correct, you need one perf fd per cpu, and map them accordingly > into the map slots when you use BPF_F_CURRENT_CPU. Given this is a sample that users are likely to copy from, I think it would be great if you could fix this up so you can just pass in BPF_F_CURRENT_CPU eventually. Thanks for working on this, Toke!