Re: [PATCH 2/5] mfd: arizona: Fix and factor out read of device tree GPIOs

From: Lee Jones <lee.jones@linaro.org>
To: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
Cc: broonie@kernel.org, robh+dt@kernel.org, pawel.moll@arm.com,
	mark.rutland@arm.com, ijc+devicetree@hellion.org.uk,
	galak@codeaurora.org, rob@landley.net, sameo@linux.intel.com,
	lgirdwood@gmail.com, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 2/5] mfd: arizona: Fix and factor out read of device tree GPIOs
Date: Fri, 21 Mar 2014 10:52:18 +0000	[thread overview]
Message-ID: <20140321105218.GH15213@lee--X1> (raw)
In-Reply-To: <1395224338-3709-3-git-send-email-ckeepax@opensource.wolfsonmicro.com>

On Wed, 19 Mar 2014, Charles Keepax wrote:

> On non-DT systems with device tree built in the current device tree GPIO
> reads will overwrite the pdata with zero when they fail. This patch
> factors out the reading of GPIOs for the Arizona devices into a helper
> function, and ensures that the pdata version will be preserved if the
> device tree read fails.
> 
> Signed-off-by: Charles Keepax <ckeepax@opensource.wolfsonmicro.com>
> ---
>  drivers/mfd/arizona-core.c       |   33 ++++++++++++++++++++++++---------
>  include/linux/mfd/arizona/core.h |    3 +++
>  2 files changed, 27 insertions(+), 9 deletions(-)
> 
> diff --git a/drivers/mfd/arizona-core.c b/drivers/mfd/arizona-core.c
> index a45aab9..333621e 100644
> --- a/drivers/mfd/arizona-core.c
> +++ b/drivers/mfd/arizona-core.c
> @@ -512,19 +512,34 @@ int arizona_of_get_type(struct device *dev)
>  }
>  EXPORT_SYMBOL_GPL(arizona_of_get_type);
>  
> +int arizona_of_get_named_gpio(struct arizona *arizona, const char *prop,
> +			      bool mandatory, int *gpio)
> +{
> +	int ret;
> +
> +	ret = of_get_named_gpio(arizona->dev->of_node, prop, 0);
> +	if (ret >= 0) {
> +		*gpio = ret;
> +		return ret;
> +	}
> +
> +	/* Warn if GPIO is mandatory and not specified */
> +	if (mandatory && *gpio <= 0)
> +		dev_err(arizona->dev,
> +				"Mandatory DT gpio %s missing/malformed: %d\n",
> +				prop, ret);
> +
> +	return ret;
> +}
> +EXPORT_SYMBOL_GPL(arizona_of_get_named_gpio);
> +
>  static int arizona_of_get_core_pdata(struct arizona *arizona)
>  {
> +	struct arizona_pdata *pdata = &arizona->pdata;
>  	int ret, i;
>  
> -	arizona->pdata.reset = of_get_named_gpio(arizona->dev->of_node,
> -						 "wlf,reset", 0);
> -	if (arizona->pdata.reset < 0)
> -		arizona->pdata.reset = 0;
> -
> -	arizona->pdata.ldoena = of_get_named_gpio(arizona->dev->of_node,
> -						  "wlf,ldoena", 0);
> -	if (arizona->pdata.ldoena < 0)
> -		arizona->pdata.ldoena = 0;
> +	arizona_of_get_named_gpio(arizona, "wlf,reset", true, &pdata->reset);
> +	arizona_of_get_named_gpio(arizona, "wlf,ldoena", true, &pdata->ldoena);
>  
>  	ret = of_property_read_u32_array(arizona->dev->of_node,
>  					 "wlf,gpio-defaults",

This doesn't look correct to me at all. Have you tested it in the DT
and !DT case? By the looks of it arizona_of_get_core_pdata() is only
called if no platform data is present. Yet you are dereferencing it
here without allocated any memory.

So how were any platform data passed variables being overwritten in
the first place?

> diff --git a/include/linux/mfd/arizona/core.h b/include/linux/mfd/arizona/core.h
> index 5cf8b91..16150cc 100644
> --- a/include/linux/mfd/arizona/core.h
> +++ b/include/linux/mfd/arizona/core.h
> @@ -124,4 +124,7 @@ int wm5102_patch(struct arizona *arizona);
>  int wm5110_patch(struct arizona *arizona);
>  int wm8997_patch(struct arizona *arizona);
>  
> +extern int arizona_of_get_named_gpio(struct arizona *arizona, const char *prop,
> +				     bool mandatory, int *gpio);
> +
>  #endif

-- 
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog