From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [PATCH net-next v3] rps: selective flow shedding during softnet overflow Date: Tue, 23 Apr 2013 13:46:49 -0700 Message-ID: <1366750009.8964.9.camel@edumazet-glaptop> References: <1366742817-8480-1-git-send-email-willemb@google.com> <1366744738.8964.6.camel@edumazet-glaptop> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org, David Miller , Stephen Hemminger To: Willem de Bruijn Return-path: Received: from mail-pa0-f50.google.com ([209.85.220.50]:58724 "EHLO mail-pa0-f50.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755216Ab3DWUqw (ORCPT ); Tue, 23 Apr 2013 16:46:52 -0400 Received: by mail-pa0-f50.google.com with SMTP id bg4so727427pad.37 for ; Tue, 23 Apr 2013 13:46:52 -0700 (PDT) In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Tue, 2013-04-23 at 16:30 -0400, Willem de Bruijn wrote: > On Tue, Apr 23, 2013 at 3:18 PM, Eric Dumazet wrote: > >> +write_unlock: > >> + synchronize_rcu(); > > > > I believe you do not need this synchronize_rcu() call. > > Because in this special case rcu_assign_pointer always replaces a > NULL value, correct? Thanks again for the feedback! I rebased, reran > the tests and will send v4 with these two changes (only). Well, there is no assignment or freeing after the synchronize_rcu(); (mask is a local var only) This looks as a leftover.