From mboxrd@z Thu Jan  1 00:00:00 1970
From: broonie@opensource.wolfsonmicro.com (Mark Brown)
Date: Wed, 9 Dec 2009 15:20:27 +0000
Subject: [PATCH v2 09/09] regulator: refresh 88pm8607 driver with
	updated api
In-Reply-To: <771cded00912090518s29c295a9v7f555a543f3bc4e6@mail.gmail.com>
References: <771cded00912090518s29c295a9v7f555a543f3bc4e6@mail.gmail.com>
Message-ID: <20091209152026.GD32223@rakim.wolfsonmicro.main>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On Wed, Dec 09, 2009 at 08:18:18AM -0500, Haojian Zhuang wrote:

> Since i2c API of mfd 88pm860x driver is changed, refresh 88pm8607 driver.
> Remove the support of 88PM8607 A0/A1 stepping.

> Signed-off-by: Haojian Zhuang <haojian.zhuang@marvell.com>

Might be nice to split the removal of the old devices into a separate
patch but that is definitely a win.  One blocker below, though.

> +		ret = (index < 3) ? (index * 50000 + 1800000) :
> +			((index < 7) ? (index * 50000 + 2550000) :
> +			3300000);

I know the previous code did this but I'm really not loving the multiple
ternery operators here for legibility.

>  	return 0;
>  }
> -subsys_initcall(pm8607_regulator_init);
> +module_init(pm8607_regulator_init);

This should be left at subsys_initcall() to try to ensure that the
regulators are initialsied before the things that use them, otherwise
the handling of init sequencing gets too complex.