Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753424AbZGNDz3 (ORCPT ); Mon, 13 Jul 2009 23:55:29 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753350AbZGNDz3 (ORCPT ); Mon, 13 Jul 2009 23:55:29 -0400 Received: from fgwmail6.fujitsu.co.jp ([192.51.44.36]:46116 "EHLO fgwmail6.fujitsu.co.jp" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753255AbZGNDz2 (ORCPT ); Mon, 13 Jul 2009 23:55:28 -0400 X-SecurityPolicyCheck-FJ: OK by FujitsuOutboundMailChecker v1.3.1 Date: Tue, 14 Jul 2009 12:53:34 +0900 From: KAMEZAWA Hiroyuki To: Paul Menage Cc: Benjamin Blum , Li Zefan , linux-kernel@vger.kernel.org, containers@lists.linux-foundation.org, akpm@linux-foundation.org, serue@us.ibm.com Subject: Re: [PATCH 3/3] Quick vmalloc vs kmalloc fix to the case where array size is too large Message-Id: <20090714125334.b476aa78.kamezawa.hiroyu@jp.fujitsu.com> In-Reply-To: <6599ad830907132050n226aecb3ucab6746a4d0e81fa@mail.gmail.com> References: <20090710230043.16778.29656.stgit@hastromil.mtv.corp.google.com> <20090710230205.16778.11707.stgit@hastromil.mtv.corp.google.com> <4A5AA3E7.9070800@cn.fujitsu.com> <20090713150303.70ab5176.kamezawa.hiroyu@jp.fujitsu.com> <2f86c2480907130827u7d2b062bw26bbb80a8e3de657@mail.gmail.com> <20090714084950.2401d9a6.kamezawa.hiroyu@jp.fujitsu.com> <6599ad830907132050n226aecb3ucab6746a4d0e81fa@mail.gmail.com> Organization: FUJITSU Co. LTD. X-Mailer: Sylpheed 2.5.0 (GTK+ 2.10.14; i686-pc-mingw32) 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 Content-Length: 897 Lines: 24 On Mon, 13 Jul 2009 20:50:33 -0700 Paul Menage wrote: > On Mon, Jul 13, 2009 at 4:49 PM, KAMEZAWA > Hiroyuki wrote: > > IIUC, this place, .../procs interface is not so important for performance > > as to being allowed this ugly conding. > > > > It's not just the calling thread that suffers from the overhead - > we've seen performance hits on other processes on a machine due to the > TLB-shootdown overhead associated with vmalloc()/vfree(). > Hmm, ok...then, if too much pids, you hit vmalloc()/vfree() problem again. So, it's not good idea to use vmalloc/vfree here after all. Thanks, -Kame -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/