From mboxrd@z Thu Jan  1 00:00:00 1970
From: zhouqiao@marvell.com (Qiao Zhou)
Date: Fri, 15 Jun 2012 13:23:36 +0800
Subject: [PATCH 1/3] mfd: support 88pm80x in 80x driver
In-Reply-To: <20120614184356.GC30185@sirena.org.uk>
References: <1339578251-15009-1-git-send-email-zhouqiao@marvell.com>
 <1339578251-15009-2-git-send-email-zhouqiao@marvell.com>
 <201206141227.48867.arnd@arndb.de> <20120614184356.GC30185@sirena.org.uk>
Message-ID: <4FDAC6D8.6070202@marvell.com>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On 06/15/2012 02:43 AM, Mark Brown wrote:
> On Thu, Jun 14, 2012 at 12:27:48PM +0000, Arnd Bergmann wrote:
>> On Wednesday 13 June 2012, Qiao Zhou wrote:
>
>>> +static struct mfd_cell onkey_devs[] = {
>>> +	{
>>> +	 .name = "88pm80x-onkey",
>>> +	 .num_resources = 1,
>>> +	 .resources =&onkey_resources[0],
>>> +	 .id = -1,
>>> +	 },
>>> +};
>
>> I wonder if it really makes sense to use the mfd_cell abstraction here, when each
>> array only contains a single device. Why not just use
>> platform_device_register_simple()?
>
> mfd_cell does some other stuff like let you pass in bases for the various
> resource ranges when you register which looks to be being used here.
>
>>> +static irqreturn_t pm805_irq(int irq, void *data)
>>> +{
>>> +	struct pm80x_chip *chip = data;
>>> +	struct pm80x_subchip *pm805_chip = chip->pm805_chip;
>>> +	struct pm80x_irq_data *irq_data;
>>> +	struct i2c_client *i2c;
>>> +	int i, read_reg = -1, value = 0;
>
>> The functions for pm800 and pm805 look almost identical. Have you tried
>> consolidating them so you can use the same irqchip and code with different
>> init data structures?
>
> Since you use regmap perhaps you can use regmap_irq and just convert
> them into data?
Mark,

thanks for your suggestion, and would update accordingly.

-- 

Best Regards
Qiao