From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Lunn Subject: Re: [PATCH 1/4] cpufreq: Add a cpufreq driver for Marvell Dove Date: Mon, 28 Oct 2013 12:31:04 +0100 Message-ID: <20131028113104.GE27015@lunn.ch> References: <1382533461-31021-1-git-send-email-andrew@lunn.ch> <20131023135116.GC31126@lunn.ch> <20131023143657.GA31536@lunn.ch> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from vps0.lunn.ch ([178.209.37.122]:44536 "EHLO vps0.lunn.ch" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755773Ab3J1Lfz (ORCPT ); Mon, 28 Oct 2013 07:35:55 -0400 Content-Disposition: inline In-Reply-To: Sender: linux-pm-owner@vger.kernel.org List-Id: linux-pm@vger.kernel.org To: Viresh Kumar Cc: Andrew Lunn , Sebastian Hesselbarth , Jason Cooper , "Rafael J. Wysocki" , "linux-pm@vger.kernel.org" , linux ARM On Wed, Oct 23, 2013 at 08:30:02PM +0530, Viresh Kumar wrote: > On 23 October 2013 20:06, Andrew Lunn wrote: > >> So, you have actually tested your code without interrupt handler? > > > > No. > > It would be better if you atleast try this and confirm that this dummy > handler is required. Hi Viresh I have tested this now. The dummy handler is required. Without it, it seems like the WFI exits at the next timer tick, not immediately once the cpufreq change has finished. I get poor numbers out of cpufreq-bench without the handler. Andrew From mboxrd@z Thu Jan 1 00:00:00 1970 From: andrew@lunn.ch (Andrew Lunn) Date: Mon, 28 Oct 2013 12:31:04 +0100 Subject: [PATCH 1/4] cpufreq: Add a cpufreq driver for Marvell Dove In-Reply-To: References: <1382533461-31021-1-git-send-email-andrew@lunn.ch> <20131023135116.GC31126@lunn.ch> <20131023143657.GA31536@lunn.ch> Message-ID: <20131028113104.GE27015@lunn.ch> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wed, Oct 23, 2013 at 08:30:02PM +0530, Viresh Kumar wrote: > On 23 October 2013 20:06, Andrew Lunn wrote: > >> So, you have actually tested your code without interrupt handler? > > > > No. > > It would be better if you atleast try this and confirm that this dummy > handler is required. Hi Viresh I have tested this now. The dummy handler is required. Without it, it seems like the WFI exits at the next timer tick, not immediately once the cpufreq change has finished. I get poor numbers out of cpufreq-bench without the handler. Andrew