[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20160209100936.GA8273@gmail.com>
Date: Tue, 9 Feb 2016 11:09:36 +0100
From: Ingo Molnar <mingo@...nel.org>
To: Steven Rostedt <rostedt@...dmis.org>
Cc: 平松雅巳 / HIRAMATU,MASAMI
<masami.hiramatsu.pt@...achi.com>,
'Martin KaFai Lau' <kafai@...com>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
Ingo Molnar <mingo@...hat.com>,
Alexei Starovoitov <alexei.starovoitov@...il.com>,
Josef Bacik <jbacik@...com>, Kernel Team <kernel-team@...com>
Subject: Re: [PATCH RESEND] kprobe: Use percpu counter to collect nhit
statistics
* Steven Rostedt <rostedt@...dmis.org> wrote:
> On Thu, 4 Feb 2016 10:24:12 +0000
> 平松雅巳 / HIRAMATU,MASAMI <masami.hiramatsu.pt@...achi.com> wrote:
>
> > From: Martin KaFai Lau [mailto:kafai@...com]
> > >
> > >When doing ebpf+kprobe on some hot TCP functions (e.g.
> > >tcp_rcv_established), the kprobe_dispatcher shows up in 'perf report'.
> > >
> > >In kprobe_dispatcher(), there is a lot of cache bouncing
> > >in 'tk->nhit++'. 'tk->nhit' and 'tk->tp.flags' also share
> > >the same cacheline.
> > >
> >
> > Ingo, could you consider to merge this?
> >
>
> Ingo, you can also add my
>
> Acked-by: Steven Rostedt <rostedt@...dmis.org>
Yeah, nice optimization, picked it up - thanks guys!
Ingo
Powered by blists - more mailing lists