From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH 3/5] netdev: bfin_mac: fix performance issue found by netperf Date: Fri, 29 May 2009 02:05:54 -0700 (PDT) Message-ID: <20090529.020554.107005632.davem@davemloft.net> References: <1243404615-25879-1-git-send-email-vapier@gentoo.org> <1243404615-25879-4-git-send-email-vapier@gentoo.org> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org, uclinux-dist-devel@blackfin.uclinux.org, michael.hennerich@analog.com, cooloney@kernel.org To: vapier@gentoo.org Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:38943 "EHLO sunset.davemloft.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757140AbZE2JFy (ORCPT ); Fri, 29 May 2009 05:05:54 -0400 In-Reply-To: <1243404615-25879-4-git-send-email-vapier@gentoo.org> Sender: netdev-owner@vger.kernel.org List-ID: From: Mike Frysinger Date: Wed, 27 May 2009 02:10:13 -0400 > @@ -606,86 +606,28 @@ static int bfin_mac_hard_start_xmit(struct sk_buff *skb, > struct net_device *dev) > { > u16 *data; > - > + u32 data_align = (u32)(skb->data) & 0x3; A pointer is not castable directly to a u32 everywhere. It can be 64-bits and in which case GCC will warn here. I know this is a blackfin specific driver, but please write portable code. Use "unsigned long" or similar. I'm not applying this as it is now.