From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Warren Subject: Re: [PATCH v2 5/5] ARM: tegra: Add thermal reset (thermtrip) support to PMC Date: Thu, 21 Aug 2014 09:40:02 -0600 Message-ID: <53F612D2.8020803@wwwdotorg.org> References: <1407933685-12404-1-git-send-email-mperttunen@nvidia.com> <1407933685-12404-6-git-send-email-mperttunen@nvidia.com> <53F5042A.7020603@wwwdotorg.org> <53F5F019.10308@nvidia.com> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <53F5F019.10308-DDmLM1+adcrQT0dZR+AlfA@public.gmane.org> Sender: linux-tegra-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Mikko Perttunen , thierry.reding-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org Cc: linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org, linux-tegra-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, wni-DDmLM1+adcrQT0dZR+AlfA@public.gmane.org List-Id: linux-tegra@vger.kernel.org On 08/21/2014 07:11 AM, Mikko Perttunen wrote: > On 20/08/14 23:25, Stephen Warren wrote: >> On 08/13/2014 06:41 AM, Mikko Perttunen wrote: >>> This adds a device tree controlled option to enable PMC-based >>> thermal reset in overheating situations. Thermtrip is supported on >>> Tegra30, Tegra114 and Tegra124. The thermal reset only works when >>> the thermal sensors are calibrated, so a soctherm driver is also >>> required. >> >> If calibration is required, presumably the soctherm must initialize >> before this thermtrip code can initialize, or this thermtrip logic might >> be triggered by uncalibrated sensors? > > SOCTHERM requires that each sensor be explicitly enabled before it gives > readings. If a sensor is not enabled, the temperature given by the > register (and used to trigger thermtrip) will be zero. So in order for a > thermtrip shutdown to be caused before soctherm is initialized, the > thermtrip temperature would have to be programmed to below zero (the > default value is 105C), in which case an immediate shutdown would > probably be in order anyway (unless the user uses LN2 or something to > cool the soc below zero). > >> >> If so, then there needs to be some explicit mechanism to force the two >> drivers into probing in the right order. > > Because of the above, I think it isn't necessary to probe these in order. OK, that makes sense. Briefly mentioning this in the commit description could be useful.