From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Fri, 17 Aug 2001 12:00:11 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Fri, 17 Aug 2001 12:00:02 -0400 Received: from trained-monkey.org ([209.217.122.11]:24842 "EHLO trained-monkey.org") by vger.kernel.org with ESMTP id ; Fri, 17 Aug 2001 11:59:50 -0400 From: Jes Sorensen MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <15229.16241.39447.219387@trained-monkey.org> Date: Fri, 17 Aug 2001 11:59:45 -0400 To: "Maciej W. Rozycki" Cc: Ralf Baechle , linux-kernel@vger.kernel.org Subject: Re: [patch] dz.c 64 bit locking issues In-Reply-To: In-Reply-To: <20010817164031.A6083@bacchus.dhis.org> X-Mailer: VM 6.90 under Emacs 20.7.1 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org >>>>> "Maciej" == Maciej W Rozycki writes: Maciej> On Fri, 17 Aug 2001, Ralf Baechle wrote: >> > The dz.c driver has an instance where it does save_flags() to a 32 >> bit > type which isn't safe for 64 bit boxen. >> >> It's safe because a MIPS only driver. Maciej> Not necessarily. It might be safe now, but there is a Maciej> TURBOchannel serial card with DZ11-compatible chipset that could Maciej> be driven by the dz.c code after a few tweaks. And chances are Maciej> someone will finish writing support for TURBOchannel Alphas one Maciej> day... Maciej> The change is harmless anyway. Yep, I also think it's better for cosmetic reasons that we make all drivers use unsigned long. That way the chance of someone else copying bad code for a new driver is less likely. Jes