From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Wed, 2 Apr 2003 14:10:34 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Wed, 2 Apr 2003 14:10:34 -0500 Received: from smtp1.cavium.com ([209.113.159.133]:8945 "EHLO smtp1.cavium.com") by vger.kernel.org with ESMTP id ; Wed, 2 Apr 2003 14:10:33 -0500 From: "Adam Khan" To: Cc: "'Adam Khan'" , Subject: Memory allocation differences between 2.4.18 and 2.5.52 Date: Wed, 2 Apr 2003 11:21:47 -0800 Message-ID: <000b01c2f94d$1be53780$7710a8c0@D6XKZR11> MIME-Version: 1.0 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook, Build 10.0.3416 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2600.0000 Importance: Normal X-OriginalArrivalTime: 02 Apr 2003 19:21:54.0595 (UTC) FILETIME=[1EE57F30:01C2F94D] Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org It seems that the kmalloc() function on 2.4.18 provides memory allocation blocks on 8-byte boundaries. 2.5.52 provides memory allocation blocks on 4-byte boundaries. We have run into an issue with a device driver developed for 2.4.18 and ported to 2.5.52. The driver requires buffers that are allocated on 8-byte boundaries. Can anyone explain why this difference in buffer boundaries exists? Any suggestions to ensure that the buffer allocation is 8-byte bounded? Can we write a my_kmalloc() function that would always return an 8-byte bounded buffer. Thanks in advance, Adam