From mboxrd@z Thu Jan 1 00:00:00 1970 From: Guenter Roeck Subject: Re: [PATCH v2 1/3] power: reset: Add Renesas reset driver Date: Wed, 22 Feb 2017 07:28:15 -0800 Message-ID: <20170222152815.GA23780@roeck-us.net> References: <20170216172320.10897-1-chris.brandt@renesas.com> <20170216172320.10897-2-chris.brandt@renesas.com> <20170216182644.GA17814@roeck-us.net> <20170216222014.GA29672@roeck-us.net> <55cf087f-aa70-3660-1186-af9b9b20376d@roeck-us.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: Sender: linux-pm-owner@vger.kernel.org To: Chris Brandt Cc: Geert Uytterhoeven , Wim Van Sebroeck , Sebastian Reichel , Rob Herring , Mark Rutland , Simon Horman , "linux-renesas-soc@vger.kernel.org" , "linux-pm@vger.kernel.org" , "devicetree@vger.kernel.org" , "linux-watchdog@vger.kernel.org" List-Id: devicetree@vger.kernel.org On Wed, Feb 22, 2017 at 01:35:12PM +0000, Chris Brandt wrote: > Hello Geert and Guenter, > > On Thursday, February 16, 2017, Guenter Roeck wrote: > > FWIW, the watchdog subsystem should support that easily, even with 125 ms > > hardware timeout. We added that capability for that very purpose. That > > would only fail if the system is stuck with interrupts disabled for more > > than 125 ms, which seems unlikely. I think the gpio watchdog on some > > systems has a similar low hardware timeout. > > > While I'm going to try that for the RZ/A1, I have a question for you guys > (or anyone else that has an opinion about end applications) > > > If I were going to make a request to the chip designers to make the timeout longer > for the next RZ/A chip, what would be a good max timeout for common Linux applications? > > Looking through the drivers in the watchdog directly, I see default timeouts of 20, > 30, 60, and 120 seconds. > 30 and 60 are pretty common for default timeouts, though I personally think they are a bit long. If you want direct HW support, a maximum of 120 seconds should be sufficient though not really necessary anymore since the core supports virtual timeouts. A maximum of at least 30 seconds would be needed if the watchdog is supposed to run at boot time (ie if it is enabled by ROMMON/BIOS and kept running by the kernel). Hope this helps, Guenter