From mboxrd@z Thu Jan 1 00:00:00 1970 From: Santosh Shilimkar Subject: Re: [PATCH 0/4] cpufreq: OMAP: if available, scale the iva coprocessor Date: Wed, 7 Nov 2012 08:42:38 -0600 Message-ID: <509A735E.4080103@ti.com> References: <1352252861-18384-1-git-send-email-jemele@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset="ISO-8859-1"; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1352252861-18384-1-git-send-email-jemele@gmail.com> Sender: cpufreq-owner@vger.kernel.org To: Joshua Emele Cc: Kevin Hilman , "Rafael J. Wysocki" , linux-omap@vger.kernel.org, cpufreq@vger.kernel.org, linux-pm@vger.kernel.org, linux-kernel@vger.kernel.org List-Id: linux-pm@vger.kernel.org Hi, On Tuesday 06 November 2012 07:47 PM, Joshua Emele wrote: > The iva coprocessor, available on some omap platforms, shares a voltage domain > with the mpu. If cpufreq is active and the mpu processor is scaled down, the iva > coprocessor should also be scaled. The goal is to make sure we do not ramp down > the voltage on the domain and affect clocking on the iva coprocessor leading to > a dsp crash. > > I only have access to an omap3evm-ish device, so I do not know what the iva > clock name is for omap24xx and omap44xx. This detail can be added later if the > general approach is approved. > > I have tested a version of this patch against the linux-3.3 kernel, so this my > attempt at a forward port against the current mainline. I have based my patch > series against linux-omap-pm/pm-next. > How about making use of CPUFREQ notifiers to handle it. I don't like this series since it adds multi-media co-processor scaling into the CPUFreq driver. Regards Santosh