Re: [PATCH 04/14] usb: phy: nop: Handle power supply regulator for the PHY

[Roger Quadros <rogerq@ti.com>]

On 01/11/2013 07:17 PM, Russell King - ARM Linux wrote:
> On Thu, Jan 10, 2013 at 06:51:24PM +0200, Roger Quadros wrote:
>> We use "vcc" as the supply name for the PHY's power supply.
>> The power supply will be enabled during .init() and disabled
>> during .shutdown()
>>
>> Signed-off-by: Roger Quadros <rogerq@ti.com>
>> ---
>>  drivers/usb/otg/nop-usb-xceiv.c |   18 ++++++++++++++++++
>>  1 files changed, 18 insertions(+), 0 deletions(-)
>>
>> diff --git a/drivers/usb/otg/nop-usb-xceiv.c b/drivers/usb/otg/nop-usb-xceiv.c
>> index 163f972..1c6db10 100644
>> --- a/drivers/usb/otg/nop-usb-xceiv.c
>> +++ b/drivers/usb/otg/nop-usb-xceiv.c
>> @@ -33,11 +33,13 @@
>>  #include <linux/usb/nop-usb-xceiv.h>
>>  #include <linux/slab.h>
>>  #include <linux/clk.h>
>> +#include <linux/regulator/consumer.h>
>>  
>>  struct nop_usb_xceiv {
>>  	struct usb_phy		phy;
>>  	struct device		*dev;
>>  	struct clk		*clk;
>> +	struct regulator	*vcc;
>>  };
>>  
>>  static struct platform_device *pd;
>> @@ -70,6 +72,11 @@ static int nop_init(struct usb_phy *phy)
>>  {
>>  	struct nop_usb_xceiv *nop = dev_get_drvdata(phy->dev);
>>  
>> +	if (nop->vcc) {
>> +		if (regulator_enable(nop->vcc))
>> +			dev_err(phy->dev, "Failed to enable power\n");
>> +	}
>> +
>>  	if (nop->clk)
>>  		clk_enable(nop->clk);
>>  
>> @@ -82,6 +89,11 @@ static void nop_shutdown(struct usb_phy *phy)
>>  
>>  	if (nop->clk)
>>  		clk_disable(nop->clk);
>> +
>> +	if (nop->vcc) {
>> +		if (regulator_disable(nop->vcc))
>> +			dev_err(phy->dev, "Failed to disable power\n");
>> +	}
>>  }
>>  
>>  static int nop_set_peripheral(struct usb_otg *otg, struct usb_gadget *gadget)
>> @@ -157,6 +169,12 @@ static int nop_usb_xceiv_probe(struct platform_device *pdev)
>>  		}
>>  	}
>>  
>> +	nop->vcc = devm_regulator_get(&pdev->dev, "vcc");
>> +	if (IS_ERR(nop->vcc)) {
>> +		dev_dbg(&pdev->dev, "Error getting vcc regulator\n");
>> +		nop->vcc = NULL;
>> +	}
> 
> Is it really appropriate for drivers to do this kind of thing with
> pointer-returning functions (I mean, setting the pointer to NULL on
> error, rather than just using a test for IS_ERR() in the above
> locations).  You are imposing driver-local assumptions on an API.
> 
> Practically it probably doesn't make much difference but given the
> amount of mistakes that we have with IS_ERR_OR_NULL()...
> 
Makes sense. I'll convert it to use IS_ERR_OR_NULL() throughout.

--
cheers,
-roger