From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [PATCH] net: Support specifying the network namespace upon device creation. Date: Tue, 03 Nov 2009 11:16:59 +0100 Message-ID: <4AF0031B.4060708@gmail.com> References: <4AEFF7BB.4090301@gmail.com> <20091103.020555.156050455.davem@davemloft.net> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: ebiederm@xmission.com, netdev@vger.kernel.org To: David Miller Return-path: Received: from gw1.cosmosbay.com ([212.99.114.194]:57084 "EHLO gw1.cosmosbay.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750933AbZKCKRZ (ORCPT ); Tue, 3 Nov 2009 05:17:25 -0500 In-Reply-To: <20091103.020555.156050455.davem@davemloft.net> Sender: netdev-owner@vger.kernel.org List-ID: David Miller a =E9crit : > It can also be argued that for functions, wrapping the args is > worse because it makes grep output less useful. In fact that's, > I believe, Linus's most recent recommendation in this area :) Yes, true. One thing I usually miss is the { next to "struct some_name" declaratio= ns, to ease games based on "grep"=20 Would you accept one boring cleanup patch in include/net like following= ? diff --git a/include/net/route.h b/include/net/route.h index 40f6346..1524b75 100644 --- a/include/net/route.h +++ b/include/net/route.h @@ -49,8 +49,7 @@ =20 struct fib_nh; struct inet_peer; -struct rtable -{ +struct rtable { union { struct dst_entry dst; @@ -77,16 +76,14 @@ struct rtable struct inet_peer *peer; /* long-living peer info */ }; =20 -struct ip_rt_acct -{ +struct ip_rt_acct { __u32 o_bytes; __u32 o_packets; __u32 i_bytes; __u32 i_packets; }; =20 -struct rt_cache_stat=20 -{ +struct rt_cache_stat { unsigned int in_hit; unsigned int in_slow_tot; unsigned int in_slow_mc;