From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754261Ab3BHAGY (ORCPT ); Thu, 7 Feb 2013 19:06:24 -0500 Received: from mail.linuxfoundation.org ([140.211.169.12]:41599 "EHLO mail.linuxfoundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751268Ab3BHAGX (ORCPT ); Thu, 7 Feb 2013 19:06:23 -0500 Date: Thu, 7 Feb 2013 16:06:20 -0800 From: Andrew Morton To: Zhang Yanfei Cc: "Eric W. Biederman" , benh@kernel.crashing.org, mahesh@linux.vnet.ibm.com, zhangyanfei@cn.fujitsu.com, linux-kernel@vger.kernel.org Subject: Re: [PATCH] fs/proc/vmcore.c: Put if tests in the top of the while loop to reduce duplication Message-Id: <20130207160620.6c8f93d4.akpm@linux-foundation.org> In-Reply-To: <510E6280.3070203@gmail.com> References: <510E6280.3070203@gmail.com> X-Mailer: Sylpheed 3.0.2 (GTK+ 2.20.1; x86_64-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Sun, 03 Feb 2013 21:13:36 +0800 Zhang Yanfei wrote: > From: Zhang Yanfei > > In function read_vmcore, two if tests are duplicate. Change the position > of them could reduce the duplication. This change does not affect > the behaviour of the function. > hm, yes. > --- a/fs/proc/vmcore.c > +++ b/fs/proc/vmcore.c > @@ -175,15 +175,16 @@ static ssize_t read_vmcore(struct file *file, char __user *buffer, > start = map_offset_to_paddr(*fpos, &vmcore_list, &curr_m); > if (!curr_m) > return -EINVAL; > - if ((tsz = (PAGE_SIZE - (start & ~PAGE_MASK))) > buflen) > - tsz = buflen; > - > - /* Calculate left bytes in current memory segment. */ > - nr_bytes = (curr_m->size - (start - curr_m->paddr)); > - if (tsz > nr_bytes) > - tsz = nr_bytes; > > while (buflen) { > + if ((tsz = (PAGE_SIZE - (start & ~PAGE_MASK))) > buflen) > + tsz = buflen; > + > + /* Calculate left bytes in current memory segment. */ > + nr_bytes = (curr_m->size - (start - curr_m->paddr)); > + if (tsz > nr_bytes) > + tsz = nr_bytes; > + > tmp = read_from_oldmem(buffer, tsz, &start, 1); > if (tmp < 0) > return tmp; > @@ -198,12 +199,6 @@ static ssize_t read_vmcore(struct file *file, char __user *buffer, > struct vmcore, list); > start = curr_m->paddr; `start' gets changed here > } > - if ((tsz = (PAGE_SIZE - (start & ~PAGE_MASK))) > buflen) > - tsz = buflen; > - /* Calculate left bytes in current memory segment. */ > - nr_bytes = (curr_m->size - (start - curr_m->paddr)); > - if (tsz > nr_bytes) > - tsz = nr_bytes; So this code can calcualte a different value of `tsz'. But that won't get used. Looks right to me. > } > return acc; > } While we're there, let's squish that assignment-in-a-if pestilence and use max_t(). --- a/fs/proc/vmcore.c~fs-proc-vmcorec-put-if-tests-in-the-top-of-the-while-loop-to-reduce-duplication-fix +++ a/fs/proc/vmcore.c @@ -178,8 +178,7 @@ static ssize_t read_vmcore(struct file * return -EINVAL; while (buflen) { - if ((tsz = (PAGE_SIZE - (start & ~PAGE_MASK))) > buflen) - tsz = buflen; + tsz = max_t(size_t, buflen, PAGE_SIZE - (start & ~PAGE_MASK)); /* Calculate left bytes in current memory segment. */ nr_bytes = (curr_m->size - (start - curr_m->paddr)); _