From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from bh-25.webhostbox.net ([208.91.199.152]:46061 "EHLO bh-25.webhostbox.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751553AbcBFQfM (ORCPT ); Sat, 6 Feb 2016 11:35:12 -0500 Subject: Re: [RFC 2/4] watchdog: ni9x3x_wdt: Add counter sysfs attribute To: Josh Cartwright , Kyle Roeschley References: <1454635683-13668-1-git-send-email-kyle.roeschley@ni.com> <1454635683-13668-2-git-send-email-kyle.roeschley@ni.com> <20160205195336.GB9579@jcartwri.amer.corp.natinst.com> Cc: wim@iguana.be, linux-watchdog@vger.kernel.org From: Guenter Roeck Message-ID: <56B620BA.8000009@roeck-us.net> Date: Sat, 6 Feb 2016 08:35:06 -0800 MIME-Version: 1.0 In-Reply-To: <20160205195336.GB9579@jcartwri.amer.corp.natinst.com> Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 8bit Sender: linux-watchdog-owner@vger.kernel.org List-Id: linux-watchdog@vger.kernel.org On 02/05/2016 11:53 AM, Josh Cartwright wrote: > On Thu, Feb 04, 2016 at 07:28:01PM -0600, Kyle Roeschley wrote: >> The NI 9x3x watchdog timer has 30.72 µs resolution, so expose this >> by allowing the user to set the timeout counter directly via a sysfs >> attribute. > > But _why_? That isn't clear to me. Why do we want a user to directly > muck with the counter value? (As opposed to working in units of time?). > In addition to that, this also messes up the watchdog core's notion of wdd->timeout, meaning the watchdog subsystem would no longer report the actually configured timeout value to user space. Guenter