[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Thu, 8 Sep 2016 22:48:12 -0700
From: Alexei Starovoitov <alexei.starovoitov@...il.com>
To: Cong Wang <xiyou.wangcong@...il.com>
Cc: Eric Dumazet <eric.dumazet@...il.com>,
David Miller <davem@...emloft.net>,
Linux Kernel Network Developers <netdev@...r.kernel.org>,
Jamal Hadi Salim <jhs@...atatu.com>,
John Fastabend <john.fastabend@...il.com>,
Hadar Hen Zion <hadarh@...lanox.com>,
Amir Vadai <amirva@...lanox.com>
Subject: Re: [PATCH net] net_sched: act_mirred: full rcu conversion
On Thu, Sep 08, 2016 at 10:24:32PM -0700, Cong Wang wrote:
> On Thu, Sep 8, 2016 at 8:35 AM, Eric Dumazet <eric.dumazet@...il.com> wrote:
> > From: Eric Dumazet <edumazet@...gle.com>
> >
> > As reported by Cong Wang, I was lazy when I did initial RCU conversion
> > of tc_mirred, as I thought I could avoid allocation/freeing of a
> > parameter block.
>
> Quote from Eric Dumazet:
>
> https://www.mail-archive.com/netdev@vger.kernel.org/msg115482.html
>
> <Quote>
> Well, I added a READ_ONCE() to read tcf_action once.
>
> Adding rcu here would mean adding a pointer and extra cache line, to
> deref the values.
>
> IMHO the race here has no effect . You either read the old or new value.
> </Quote>
>
> Me with facepalm... ;-)
imo the deliberate small race in Eric's initial lock removal in mirred
was a good design choice. I think he did this patch only to
silence the complains with the code instead of arguing with words.
imo the initial code is good as-is. This patch is also a nice improvement,
but certainly not mandatory. 'face palm' is unnecessary.
Powered by blists - more mailing lists