From mboxrd@z Thu Jan  1 00:00:00 1970
From: subaparts@yandex.ru (Alexander Sverdlin)
Date: Wed, 02 Feb 2011 13:48:08 +0300
Subject: [PATCH] EDB93xx: Add support for CS4271 CODEC on EDB93xx boards
In-Reply-To: <0D753D10438DA54287A00B027084269764CEF59B2B@AUSP01VMBX24.collaborationhost.net>
References: <1296603653.1504.9.camel@r60e>
	<0D753D10438DA54287A00B027084269764CEF59B2B@AUSP01VMBX24.collaborationhost.net>
Message-ID: <1296643688.1504.23.camel@r60e>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

Dear Hartley, Dear Mark,

On Tue, 2011-02-01 at 18:02 -0600, H Hartley Sweeten wrote:
> > +static int edb93xx_cs4271_hw_setup(struct spi_device *spi)
> > +{
> > +	int gpio_nreset;
> > +	int err;
> > +
> > +	if (machine_is_edb9301() || machine_is_edb9302()) {
> > +		gpio_nreset = EP93XX_GPIO_LINE_EGPIO1;
> 
> Are you planning on removing gpio_nreset and gpio_disable from struct cs4271_platform_data?
> If not, you should just setup the gpio mapping in edb93xx_register_spi before you actually
> register the spi_board_info.  Then this function and the following two can just do something
> like the following and not worry about the if ... else if ... etc.

Initially I had GPIO code in machine driver, but during the
conversations in alsa-devel list it has been moved to CODEC driver:
http://mailman.alsa-project.org/pipermail/alsa-devel/2010-October/032358.html

I'm really confused now, where should it be? If I will init
cs4271_platform_data with GPIO values, CODEC driver will manage it.
Otherwise CODEC driver should be rewritten without GPIO support and I
can add it to platform code.

Mark Brown, could you please comment on this, as I've moved GPIO things
to CODEC part based on your review.

Best regards,
	Alexander Sverdlin.