From mboxrd@z Thu Jan 1 00:00:00 1970 From: Cong Wang Subject: Re: [Patch net-next v3 1/3] igmp: fix return value of some functions Date: Mon, 10 Jun 2013 11:07:33 +0800 Message-ID: <1370833653.11543.3.camel@cr0> References: <1370689003.868.3.camel@cr0> , <1370485211-11498-1-git-send-email-amwang@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: "David S. Miller" , netdev-owner@vger.kernel.org, netdev@vger.kernel.org, Stephen Hemminger To: David Stevens Return-path: Received: from mx1.redhat.com ([209.132.183.28]:27369 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751026Ab3FJDHo (ORCPT ); Sun, 9 Jun 2013 23:07:44 -0400 In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On Sat, 2013-06-08 at 20:12 -0600, David Stevens wrote: > >From: Cong Wang > > > Good point! I will convert the such return value to > >documented ones, for example, EINVAL. Thanks! > > You're also taking things that were not treated as errors > before, and now returning an error. For example, removing > a source not in the list. > > Unless POSIX requires these to be treated as errors, and > that should be quoted, I think the 8 or 9 years of returning > "0" for these cases means we must continue to. Existing > programs that have always done this with a zero return > will start getting errors where they didn't before. > > Also, though it's been a long time, I believe the original code > did return errors in those cases (which is why the code is there) > and was changed to return 0, presumably for POSIX compliance. Are you suggesting to just make ip_mc_leave_src() void? Thanks.