From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756752AbZCTLGa (ORCPT ); Fri, 20 Mar 2009 07:06:30 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751779AbZCTLGV (ORCPT ); Fri, 20 Mar 2009 07:06:21 -0400 Received: from mailrelay003.isp.belgacom.be ([195.238.6.53]:64052 "EHLO mailrelay003.isp.belgacom.be" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751315AbZCTLGU (ORCPT ); Fri, 20 Mar 2009 07:06:20 -0400 X-Belgacom-Dynamic: yes X-IronPort-Anti-Spam-Filtered: true X-IronPort-Anti-Spam-Result: ApoEAPQVw0lR9Rxo/2dsb2JhbADUCAeDdgY Date: Fri, 20 Mar 2009 12:06:15 +0100 From: Wim Van Sebroeck To: Alan Cox Cc: Wim Van Sebroeck , linux-kernel@vger.kernel.org Subject: Re: [PATCH] watchdog: New library code Message-ID: <20090320110615.GC6160@infomag.iguana.be> References: <20081204155250.59a1d30c@lxorguk.ukuu.org.uk> <20081209161425.GE32662@infomag.infomag.iguana.be> <20081209165440.64501e00@lxorguk.ukuu.org.uk> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20081209165440.64501e00@lxorguk.ukuu.org.uk> User-Agent: Mutt/1.5.18 (2008-05-17) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi Alan, I put your new library code in a git branch called alan-generic-watchdog in the linux-2.6-watchdog-next tree. My old code is in the branch wim-generic-watchdog. I'm preparing a new tree with the best of both trees in a new branch called generic-watchdog. I will sent you the merged code for review in a couple of days. > > > - Handles the temperature device > > Why don't we treat the temperature part not as a hwmon? > > Because the temperature interface was there first and is part of the > existing kernel to user space API. Now that the device is abstracted I > agree entirely that it should also appear in hwmon and the support > library code can be extended to do this along with multiple watchdogs. Agree. > > > - It does all the reboot, nowayout and module locking and housekeeping > > Why don't we use the shutdown methods of our platform and other devices instead of the reboot notifier setup? > > As and when watchdogs are all platform devices we probably should. Right > now most of them are not and we don't really have a proper 'isa_device' > object yet. I see that as a migration path or an 'aspiration', but having > it in one place will make that easier to achieve after we have the common > API. I'd rather do the conversion to a platform device and the conversion to the new api in one go. Concerning the isa_device -> see pcwd.c . > Likewise sysfs, classes and multiple watchdogs all want addressing but > that gets a lot easier once everything is using a common framework and a > passed 'struct watchdog' to make multiple instances just work. Agree. Kind regards, Wim.