From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Message-ID: <1505721917.25945.275.camel@linux.intel.com> Subject: Re: [PATCH v3] clk: x86: Do not gate clocks enabled by the firmware From: Andy Shevchenko To: Pierre-Louis Bossart , Carlo Caione , alan@linux.intel.com Cc: Michael Turquette , "open list:COMMON CLK FRAMEWORK" , Darren Hart , Stephen Boyd , Linux Upstreaming Team , Enric Balletbo Serra , Carlo Caione Date: Mon, 18 Sep 2017 11:05:17 +0300 In-Reply-To: <58ef7321-a16d-8974-6a37-db473b6bf48e@linux.intel.com> References: <20170714082356.28117-1-carlo@caione.org> <58ef7321-a16d-8974-6a37-db473b6bf48e@linux.intel.com> Content-Type: text/plain; charset="UTF-8" Mime-Version: 1.0 List-ID: On Thu, 2017-09-07 at 16:51 -0500, Pierre-Louis Bossart wrote: > > > > > > Looks like the same code that was initially used as a reference for > > support of PMC clocks in the clock framework, so now we have 2 > > drivers > > programming the same PMC hardware, not so good. We'd probably have > > to > > move the atomisp driver to move to clk_get/prepare/enable instead > > of > > the old vlv2_clck_get? > > See attached an untested change to show the idea on moving to the clk > API and get feedback (I am not a camera guy, just trying to help - > compile-tested/checkpatch only). > This code *should* enable the same functionality using existing hooks > in > the kernel (and handle the BYT/CHT clocking difference as a bonus). > I am sure Andy will have plenty of comments :-) I have one comment: nice clean up! AtomISP is a *horrible* mess. -- Andy Shevchenko Intel Finland Oy