From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751903AbbD3PBf (ORCPT ); Thu, 30 Apr 2015 11:01:35 -0400 Received: from the.earth.li ([46.43.34.31]:54414 "EHLO the.earth.li" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751476AbbD3PBd (ORCPT ); Thu, 30 Apr 2015 11:01:33 -0400 Date: Thu, 30 Apr 2015 16:01:31 +0100 From: Jonathan McDowell To: Takashi Iwai Cc: alsa-devel@alsa-project.org, linux-kernel@vger.kernel.org Subject: Re: Audio crackles with 4.1-rc1 Message-ID: <20150430150131.GA10148@earth.li> References: <20150428112156.GQ10148@earth.li> <20150428123518.GU10148@earth.li> <20150429112859.GY10148@earth.li> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Apr 29, 2015 at 01:54:57PM +0200, Takashi Iwai wrote: > At Wed, 29 Apr 2015 12:28:59 +0100, > Jonathan McDowell wrote: > > > > On Tue, Apr 28, 2015 at 04:43:00PM +0200, Takashi Iwai wrote: > > > At Tue, 28 Apr 2015 15:05:20 +0200, > > > Takashi Iwai wrote: > > > > At Tue, 28 Apr 2015 13:35:18 +0100, > > > > Jonathan McDowell wrote: > > > > > On Tue, Apr 28, 2015 at 02:00:17PM +0200, Takashi Iwai wrote: > > > > > > At Tue, 28 Apr 2015 12:21:57 +0100, > > > > > > Jonathan McDowell wrote: > > > > > > > > > > > > > Having upgraded to 4.1-rc1 from 4.0 I'm now hearing audio > > > > > > > crackles at regular intervals. I'm fairly sure this is due to > > > > > > > the HDA power save as once audio is playing things are fine, > > > > > > > it's just when starting to play audio that I hear the crackle. > > > > > > > > > > > > > > System is a Dell Latitude E7240. I haven't tried a bisect yet > > > > > > > but will attempt to find some time to do so in the next few > > > > > > > days. It looks like there have been some changes in sound/hda/ > > > > > > > between 4.0 + 4.1-rc1 so I'll concentrate on those first. > > > > > > > > > > > > > There are lots of code changes and enhancements wrt power saving > > > > > > in 4.1, and bisection won't help so much, I'm afraid. > > > > > > > > > > > > First off, check the device status while you hear crackles. Is > > > > > > the codec in runtime suspend (aka power save)? This can be seen > > > > > > in /sys/bus/hdaudio/devices/*/power/runtime_status. Then check > > > > > > the runtime status of the controller, too, found in > > > > > > /sys/class/sound/card?/device/power/runtime_status. > > > > > > > > > > The cracking is definitely happening with the transition from suspended > > > > > to active: > > > > > > > > > > /sys/class/sound/card0/device/power/runtime_status:suspended > > > > > /sys/class/sound/card1/device/power/runtime_status:active > > > > > /sys/bus/hdaudio/devices/hdaudioC0D0/power/runtime_status:suspended > > > > > /sys/bus/hdaudio/devices/hdaudioC1D0/power/runtime_status:active > > > > > > > > > > card1 + C1D0 were suspended, I hit delete in a terminal to force a > > > > > terminal bell, got the crackle and the state changed to active as above. > > > > > > > > Ah, so it's a click noise that happens only once per runtime PM > > > > transition? I thought it were constant crackling noises from your > > > > description ("regular intervals"). > > > > > > > > In anyway, please give alsa-info.sh output on both 4.0 and 4.1. > > > > > > Also, does the click noise occur only at powering up, and not at > > > powering down? > > > > The click is only on the transition from suspended to active; not > > constantly during playback and not when transitioning from active to > > suspended. It doesn't appear 4.0 is doing the suspending at all; there > > are no /sys/bus/hdaudio/devices/*/power/runtime_status files and the > > card runtime_status is always active. alsa-info for 4.0 + 4.1-rc1 > > attached. > > OK, thanks for clarification. > > One patch you can try (with or without power_save_node disablement) is > below, it squashes the verb sequences at (runtime) PM resume as we did > for 4.0. Let me know if this changes the behavior. Patch alone has no effect. Patch with power_save_node disablement solves the problem. Didn't try power_save_node disablement without the patch. J. -- This .sig brought to you by the letter M and the number 7 Product of the Republic of HuggieTag