From mboxrd@z Thu Jan 1 00:00:00 1970 From: Lu Guanqun Subject: Re: [PATCH] ASoC: Remove special casing for registerless widgets Date: Sat, 2 Apr 2011 16:22:00 +0800 Message-ID: <20110402082200.GI14553@qtel.sh.intel.com> References: <1301730541-10767-1-git-send-email-broonie@opensource.wolfsonmicro.com> <20110402080755.GG14553@qtel.sh.intel.com> <20110402081244.GH14553@qtel.sh.intel.com> <20110402081801.GC28417@opensource.wolfsonmicro.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from mga14.intel.com (mga14.intel.com [143.182.124.37]) by alsa0.perex.cz (Postfix) with ESMTP id 57F721037FC for ; Sat, 2 Apr 2011 10:22:33 +0200 (CEST) Content-Disposition: inline In-Reply-To: <20110402081801.GC28417@opensource.wolfsonmicro.com> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: alsa-devel-bounces@alsa-project.org Errors-To: alsa-devel-bounces@alsa-project.org To: Mark Brown Cc: "alsa-devel@alsa-project.org" , Liam Girdwood List-Id: alsa-devel@alsa-project.org On Sat, Apr 02, 2011 at 04:18:02PM +0800, Mark Brown wrote: > On Sat, Apr 02, 2011 at 04:12:44PM +0800, Lu Guanqun wrote: > > On Sat, Apr 02, 2011 at 04:07:55PM +0800, Lu Guanqun wrote: > > > > But I still don't quite understand the write compression, it aims at > > > reducing the number of register writes, however in > > > dapm_seq_run_coalesced, it calls snd_soc_update_bits. > > > I mean, for write compression or not, as long as its widget register is > > valid, it will be updated. So I don't see why write compression would > > reduce the number... > > The idea is that since you often have all the enable bits together in a > small set of registers if you're enabling two of the same widget type at > once and they're both in the same register you can do it with one write. > For example, if you're doing a stereo playback this will usually mean > that enabling left and right DACs is one register write. This is also > good for pop/click performance as it ensures that both channels turn on > simultaneously. Thanks for the explaination. I get the idea. > You can see this quite easily if you turn on DAPM pop debugging with a > short delay - write to asoc/$CARD/dapm_pop_time in debugfs and start a > playback or something and you'll see logging showing the writes being > combined. I'll try this and see the differences. Thanks Mark! -- guanqun