From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Thu, 16 Aug 2001 13:39:15 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Thu, 16 Aug 2001 13:39:05 -0400 Received: from femail13.sdc1.sfba.home.com ([24.0.95.140]:3521 "EHLO femail13.sdc1.sfba.home.com") by vger.kernel.org with ESMTP id ; Thu, 16 Aug 2001 13:38:55 -0400 Message-ID: <3B7C112F.1F2D183B@home.com> Date: Thu, 16 Aug 2001 11:30:07 -0700 From: kernelkracker X-Mailer: Mozilla 4.75 [en] (X11; U; SunOS 5.8 sun4u) X-Accept-Language: en MIME-Version: 1.0 To: linux-kernel@vger.kernel.org Subject: global_irq_count vs global_irq_lock Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hello, This is probably a kernel newbie question so my apologies in advance. I have a design/implementation question. global_irq_count is defined as atomic_t whereas globall_irq_lock is defined as unsigned volatiile int. Is this by design? or by chance? Why would you be concerned about reading global_irq_count atomically and not global_irq_count as it is done is several functions in linux/arch/i386/irq.c in kernel 2.2.18. If it is by design, what am I missing. Please cc me on your respone as I am not on the list. Take care!