From: Dan Murphy Subject: Re: [PATCH v2] can: tcan4x5x: Turn on the power before parsing the config Date: Fri, 31 Jan 2020 11:27:44 -0600 Message-ID: References: <20191210163204.28225-1-dmurphy@ti.com> <4a2e80f0-13c5-df7b-65af-25f86ca48f2a@pengutronix.de> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8"; format=flowed Content-Transfer-Encoding: 8bit Return-path: In-Reply-To: <4a2e80f0-13c5-df7b-65af-25f86ca48f2a-bIcnvbaLZ9MEGnE8C9+IrQ@public.gmane.org> Content-Language: en-US Sender: devicetree-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Marc Kleine-Budde Cc: linux-can-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-Id: linux-can.vger.kernel.org Marc On 1/2/20 6:38 AM, Marc Kleine-Budde wrote: > On 12/10/19 5:32 PM, Dan Murphy wrote: >> The parse config function now performs action on the device either >> reading or writing and a reset. If the regulator is managed it needs >> to be turned on. So turn on the regulator if available if the parsing >> fails then turn off the regulator. > Another BTW: > Consider converting the switching of the vsup to runtime_pm. > > Yet another one: > Why do you disable the clocks in the error path of tcan4x5x_can_probe(), > but never enable them? > >> out_clk: >> if (!IS_ERR(mcan_class->cclk)) { >> clk_disable_unprepare(mcan_class->cclk); >> clk_disable_unprepare(mcan_class->hclk); >> } > - please move the clock handling from the m_can.c to the individual > driver > - please move the clock handling to runtime_pm in the individual driver > - remove the obsolete m_can_class_get_clocks() > - make runtime_pm mandatory > > regards, > Marc > I have separate the clock calls into pm runtime calls and moved the clock init into the respective children of the framework. Did you want me to submit 1 patch with all the changes or would you like 3 separate patches?  First 2 patches will abstract the clocks away into the children and the 3rd patch would be to remove the clocks API from the framework Dan