From mboxrd@z Thu Jan 1 00:00:00 1970 From: Nicolin Chen Subject: Re: [PATCH] ASoC: simple-card: overwrite DAIFMT_MASTER of cpu_dai->fmt Date: Wed, 12 Mar 2014 10:01:24 +0800 Message-ID: <20140312020123.GC1177@MrMyself> References: <1394542472-16580-1-git-send-email-Guangyu.Chen@freescale.com> <20140311132553.GR28112@sirena.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Return-path: Content-Disposition: inline In-Reply-To: <20140311132553.GR28112@sirena.org.uk> Sender: linux-kernel-owner@vger.kernel.org To: Mark Brown Cc: alsa-devel@alsa-project.org, linux-kernel@vger.kernel.org, kuninori.morimoto.gx@renesas.com, moinejf@free.fr, Li.Xiubo@freescale.com, lgirdwood@gmail.com, Jyri Sarha List-Id: alsa-devel@alsa-project.org On Tue, Mar 11, 2014 at 01:25:53PM +0000, Mark Brown wrote: > On Tue, Mar 11, 2014 at 08:54:32PM +0800, Nicolin Chen wrote: > > Adding Jyri who's been looking at this as well but not added anyone else > working on simple-card so you might've missed his mails. > > > It's pretty fair to do it for DAIFMT_INV since two dais might have differnt > > definitions in their drivers even though it'd be better to keep it same if > > we could. > > No, that's not at all OK - anything that requires this is broken. The > same DAI format should be usable by both ends of the link unless the > board itself is inverting one of the signals or something. That makes sense to me. Okay, I'll overwrite them all... > > > Thus this patch fixes it by overwriting the DAIFMT_MASTER bit of cpu_dai->fmt > > with the one of codec_dai->fmt since we defined DAIFMT_MASTER basing on CODEC > > at the first place. > > This seems closer to what I'd expect for something like this but it does > mean that any format settings on the CPU DAI will be ignored (rather > than say warning or something). I'm not sure this is a bad thing > though, probably wants the binding documenting at least. ...along with the binding doc.