[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20190701150443.1e43e818@cakuba.netronome.com>
Date: Mon, 1 Jul 2019 15:04:43 -0700
From: Jakub Kicinski <jakub.kicinski@...ronome.com>
To: Andrii Nakryiko <andriin@...com>
Cc: <andrii.nakryiko@...il.com>, <ast@...com>, <daniel@...earbox.net>,
<kernel-team@...com>, <bpf@...r.kernel.org>,
<netdev@...r.kernel.org>
Subject: Re: [PATCH v4 bpf-next 4/4] tools/bpftool: switch map event_pipe to
libbpf's perf_buffer
On Sat, 29 Jun 2019 23:51:09 -0700, Andrii Nakryiko wrote:
> Switch event_pipe implementation to rely on new libbpf perf buffer API
> (it's raw low-level variant).
>
> Signed-off-by: Andrii Nakryiko <andriin@...com>
> -int do_event_pipe(int argc, char **argv)
> -{
> - int i, nfds, map_fd, index = -1, cpu = -1;
> struct bpf_map_info map_info = {};
> - struct event_ring_info *rings;
> - size_t tmp_buf_sz = 0;
> - void *tmp_buf = NULL;
> - struct pollfd *pfds;
> + struct perf_buffer_raw_opts opts;
I'm slightly worried we don't init the ops, but we can wait and see if
it bites us or not.
Acked-by: Jakub Kicinski <jakub.kicinski@...ronome.com>
Thanks!
Powered by blists - more mailing lists