From mboxrd@z Thu Jan 1 00:00:00 1970 From: James Bottomley Subject: Re: [patch 12/17] ips.c: fix build warning Date: Sun, 30 Mar 2008 12:07:22 -0500 Message-ID: <1206896842.4224.46.camel@localhost.localdomain> References: <200803282148.m2SLmeDg012249@imap1.linux-foundation.org> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: Received: from accolon.hansenpartnership.com ([76.243.235.52]:48242 "EHLO accolon.hansenpartnership.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752732AbYC3RH0 (ORCPT ); Sun, 30 Mar 2008 13:07:26 -0400 In-Reply-To: <200803282148.m2SLmeDg012249@imap1.linux-foundation.org> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: akpm@linux-foundation.org Cc: linux-scsi@vger.kernel.org, Mark_Salyzyn@adaptec.com, aacraid@adaptec.com, fujita.tomonori@lab.ntt.co.jp, jeff@garzik.org On Fri, 2008-03-28 at 14:48 -0700, akpm@linux-foundation.org wrote: > From: Andrew Morton > > powerpc: > > drivers/scsi/ips.c: In function 'ips_issue_copperhead': > drivers/scsi/ips.c:5436: warning: large integer implicitly truncated to unsigned type > > it doesn't seem necessary to do outw(cpu_to_le32(...), ...) anyway. Actually, it's not only not necessary ... it's an actual bug on BE platforms. Both writeX and outX/inX automatically convert from native to bus endian (usually simply to LE since the bus is PCI). Therefore if you add a cpu_to_leX, you actually write the wrong value on a BE system. Mark, it looks like the correct fix is simply to strip all of these ... has this driver actually been tested on a BE platform (like parisc or PPC)? James