From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752222AbXCKN33 (ORCPT ); Sun, 11 Mar 2007 09:29:29 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752226AbXCKN33 (ORCPT ); Sun, 11 Mar 2007 09:29:29 -0400 Received: from web36615.mail.mud.yahoo.com ([209.191.85.32]:38410 "HELO web36615.mail.mud.yahoo.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1752209AbXCKN33 (ORCPT ); Sun, 11 Mar 2007 09:29:29 -0400 DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=s1024; d=yahoo.com; h=X-YMail-OSG:Received:Date:From:Subject:To:MIME-Version:Content-Type:Content-Transfer-Encoding:Message-ID; b=3pdt/SOVYnjAI8usCWfdHXLoPth2zcaC2BpQoh7icM95o32TOXmpscNyYp3zC4l5WqOAtd24a6+KYM3fIoojQdtJKzP/kY+8iA3i420/Q3sjLKlOz7XXjXUNv/D22H0AIsfmSvfedP+HpWF/z0pJl6skVdKJ5Uo8boW6O8zsJYk=; X-YMail-OSG: aftGpHgVM1msVtsQvSVuoiqFuCdK.yHI0C4wFEkCw6.VVvSKS9Fhmg1YvafxMZUfXuE7H9FTSO2hjdjOg4_KPmpFImKUUecHh7mV3ZIG2SAWrvSkJVOHJG6VMjv84sow9oTrs6fR3F9Y_XU- Date: Sun, 11 Mar 2007 06:29:28 -0700 (PDT) From: Parav Pandit Subject: Locking interrupt handler in L1 cache To: linux-kernel@vger.kernel.org MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT Message-ID: <265904.11493.qm@web36615.mail.mud.yahoo.com> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hi, I have MPC 8548 Linux 2.6.x based firewall which will mostly do packet processing for 80% time. So obviously most of the time it will RX and TX packets through gianfar ethernet driver. I want to lock my interrupt handler of this driver in the L1 cache. 1. Is there any kernel API for locking function and data to lock them in the L1/L2 cache? 2. How can I use "icbtls" - Instruction Cache Block Touch and Lock Set" for locking my interrupt handler? 3. Is "icbtls" is the correct instruction at which I am looking at? 4. How do I find end address of the interrupt handler function and how do we pass it to cache locking instructions? (Because it can happen that interrupt handler size is more than a cache line, not aligned etc)? 5. Can we enhance request_irq() function to take an additional parameter to lock the interrupt handler in the cache? I understand that if my interrupt handler is going to be called most of the time then it is very likely to happen that OS will flush the same, but there is no guarantee for it. Regards, Parav Pandit ____________________________________________________________________________________ Get your own web address. Have a HUGE year through Yahoo! Small Business. http://smallbusiness.yahoo.com/domains/?p=BESTDEAL