From mboxrd@z Thu Jan 1 00:00:00 1970 From: Lee Jones Subject: Re: [PATCH 20/21] ASoC: codecs: Enable AB8500 CODEC for Device Tree Date: Thu, 26 Jul 2012 16:01:14 +0100 Message-ID: <50115BBA.2020400@linaro.org> References: <1343298534-13611-1-git-send-email-lee.jones@linaro.org> <1343298534-13611-21-git-send-email-lee.jones@linaro.org> <20120726115055.GD3099@opensource.wolfsonmicro.com> <50114D71.9060700@linaro.org> <20120726142825.GK3099@opensource.wolfsonmicro.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <20120726142825.GK3099@opensource.wolfsonmicro.com> Sender: linux-kernel-owner@vger.kernel.org To: Mark Brown Cc: linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, STEricsson_nomadik_linux@list.st.com, linus.walleij@stericsson.com, arnd@arndb.de, sameo@linux.intel.com, olalilja@yahoo.se, ola.o.lilja@stericsson.com, alsa-devel@alsa-project.org, lrg@ti.com List-Id: alsa-devel@alsa-project.org On 26/07/12 15:28, Mark Brown wrote: > On Thu, Jul 26, 2012 at 03:00:17PM +0100, Lee Jones wrote: >> On 26/07/12 12:50, Mark Brown wrote: > >>> Yet again no binding documentation.... > >> RFC. ;) > >> I'll write the documentation when/if the properties are accepted. > > No, write the documentation. It's way too much effort to reverse > engineer the bindings from the code. > >>>> + default : >>>> + codec->ear_cmv =3D EAR_CMV_UNKNOWN; >>>> + dev_err(dev, "Unsuitable earpiece voltage found in DT\n"); > >>> The platform data code picks a default, can't the DT code do the sa= me? > >> No, I don't think that it does? The original code returns -EINVAL >> unless a value is specified. > > The code doesn't specify values for the enumeration so it ought to > default to EAR_CMV_0_95V if nothing is specified. Ah, I see what you mean. I guess we could compromise and print a warnin= g=20 _and_ fall back to the 0th original emum. >> The original author is keen to have a clear error message in case >> users try to specify non-exact values. I'd rather we fail-out than >> use incorrect values which would be a great deal harder for a user >> to debug. > > By that argument all the properties should be mandatory but it's only > this one IIRC. This is the only value which the user can pick an obscure value, such a= s=20 913, thinking they can pick 913mV. I'm happy to fall-back, as long as=20 Ola is too. --=20 Lee Jones Linaro ST-Ericsson Landing Team Lead Linaro.org =E2=94=82 Open source software for ARM SoCs =46ollow Linaro: Facebook | Twitter | Blog