From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stefan Tomanek Subject: Re: [PATCH] fib_rules: add minimum prefix length Date: Fri, 26 Jul 2013 00:11:16 +0200 Message-ID: <20130725221116.GC10216@zirkel.wertarbyte.de> References: <20130723220221.GP10216@zirkel.wertarbyte.de> <20130724021420.GA10670@order.stressinduktion.org> <20130725162931.GX10216@zirkel.wertarbyte.de> <20130725181712.GB24007@order.stressinduktion.org> <20130725182818.GZ10216@zirkel.wertarbyte.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Andrew Collins To: netdev@vger.kernel.org Return-path: Received: from zirkel.wertarbyte.de ([188.40.44.137]:55967 "EHLO zirkel.wertarbyte.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754659Ab3GYWLR (ORCPT ); Thu, 25 Jul 2013 18:11:17 -0400 Content-Disposition: inline In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: Dies schrieb Andrew Collins (bsderandrew@gmail.com): > > I've just completed the patch and sent it to the list, and I also placed the > > changes in my github account, just in case anyone wants to take a quick peek: > > https://github.com/wertarbyte/linux/compare/prefixlength > > > + if (arg->result) { ... > > This NULL check seems unnecessary (fib_table_lookup doesn't check). I wasn't sure if it was possible for fib_table_lookup to put a NULL pointer there, so I wanted to make sure following it with ->prefixlen was possible. > I like the idea, but I agree with Hannes that it'd be nice to split > this out into a fib constraint callback. I'm still not convinced that creating such a callback that will never be called from outside fib{4,}_rule_action() is worth the effort, but my insight into the code is still limited; if it's the right thing to do [tm], I won't stand in the way :-)