From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750928AbVI1VLS (ORCPT ); Wed, 28 Sep 2005 17:11:18 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750931AbVI1VLR (ORCPT ); Wed, 28 Sep 2005 17:11:17 -0400 Received: from fmr13.intel.com ([192.55.52.67]:12935 "EHLO fmsfmr001.fm.intel.com") by vger.kernel.org with ESMTP id S1750923AbVI1VLQ (ORCPT ); Wed, 28 Sep 2005 17:11:16 -0400 Subject: earlier allocation of order 0 pages in __alloc_pages From: Rohit Seth To: linux-kernel@vger.kernel.org, linux-mm@kvack.org Content-Type: text/plain Organization: Intel Date: Wed, 28 Sep 2005 14:18:36 -0700 Message-Id: <1127942316.5046.37.camel@akash.sc.intel.com> Mime-Version: 1.0 X-Mailer: Evolution 2.2.2 (2.2.2-5) Content-Transfer-Encoding: 7bit X-OriginalArrivalTime: 28 Sep 2005 21:11:06.0502 (UTC) FILETIME=[240BAA60:01C5C471] Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org I'm wondering if it is a good idea in __alloc_pages to first try to see if a order 0 request can be serviced by cpu's pcp before checking the low water marks for the zone. The is useful if a request can be serviced by a free page on the pcp then there is no reason to check the zone's limits. This early allocation should be without any replenishing of pcps from zone free list thanks, -rohit