[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CANn89i+A2wBXmu8kPpJraBsaz3ndGYYuKz8=y_6n=_mL+HzhHg@mail.gmail.com>
Date: Fri, 9 Oct 2020 09:35:03 +0200
From: Eric Dumazet <edumazet@...gle.com>
To: Jakub Kicinski <kuba@...nel.org>
Cc: Eric Dumazet <eric.dumazet@...il.com>,
"David S . Miller" <davem@...emloft.net>,
netdev <netdev@...r.kernel.org>,
Cong Wang <xiyou.wangcong@...il.com>,
Allen Pais <allen.lkml@...il.com>
Subject: Re: [PATCH net-next] net/sched: get rid of qdisc->padded
On Fri, Oct 9, 2020 at 2:18 AM Jakub Kicinski <kuba@...nel.org> wrote:
>
> On Wed, 7 Oct 2020 09:51:11 -0700 Eric Dumazet wrote:
> > diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
> > index 6c762457122fd0091cb0f2bf41bda73babc4ac12..d8fd8676fc724110630904909f64d7789f3a4b47 100644
> > --- a/include/net/sch_generic.h
> > +++ b/include/net/sch_generic.h
> > @@ -91,7 +91,7 @@ struct Qdisc {
> > struct net_rate_estimator __rcu *rate_est;
> > struct gnet_stats_basic_cpu __percpu *cpu_bstats;
> > struct gnet_stats_queue __percpu *cpu_qstats;
> > - int padded;
> > + int pad;
> > refcount_t refcnt;
> >
> > /*
>
> Hi Eric!
>
> Why keep the pad field? the member to lines down is
> __cacheline_aligned, so we shouldn't have to manually
> push things out?
>
> struct gnet_stats_queue __percpu *cpu_qstats;
> int pad;
I usually prefer adding explicit fields to tell where the holes are,
for future reuse.
Not many developers are aware of pahole existence :/
I renamed it to pad, I could have used padding or something else.
Powered by blists - more mailing lists