Re: [RFC v2] gpio/omap: auto-setup a GPIO when used as an IRQ

[Javier Martinez Canillas <javier.martinez@collabora.co.uk>]

On 09/24/2013 05:40 PM, Tony Lindgren wrote:
> * Javier Martinez Canillas <javier.martinez@collabora.co.uk> [130924 01:06]:
>> The OMAP GPIO controller HW requires a pin to be configured in GPIO
>> input mode in order to operate as an interrupt input. Since drivers
>> should not be aware of whether an interrupt pin is also a GPIO or not,
>> the HW should be fully configured/enabled as an IRQ if a driver solely
>> uses IRQ APIs such as request_irq(), and never calls any GPIO-related
>> APIs. As such, add the missing HW setup to the OMAP GPIO controller's
>> irq_chip driver.
>> 
>> Since this bypasses the GPIO subsystem we have to ensure that another
>> caller won't be able to request the same GPIO pin that is used as an
>> IRQ and set its direction as output. Requesting the GPIO and setting
>> its direction as input is allowed though.
> 
> Also please mention the regression that this fixes. So far we know
> that smsc911x for tobi and igep boards in mainline, and also the
> MMC card detect for omap4 boards.
>  

Ok, I'll mention that on the next post.

>> --- a/drivers/gpio/gpio-omap.c
>> +++ b/drivers/gpio/gpio-omap.c
>> @@ -63,6 +63,7 @@ struct gpio_bank {
>>  	struct gpio_chip chip;
>>  	struct clk *dbck;
>>  	u32 mod_usage;
>> +	u32 irq_usage;
>>  	u32 dbck_enable_mask;
>>  	bool dbck_enabled;
>>  	struct device *dev;
>> @@ -86,6 +87,9 @@ struct gpio_bank {
>>  #define GPIO_BIT(bank, gpio) (1 << GPIO_INDEX(bank, gpio))
>>  #define GPIO_MOD_CTRL_BIT	BIT(0)
>>  
>> +#define BANK_USED(bank) (bank->mod_usage || bank->irq_usage)
>> +#define LINE_USED(line, offset) (line & (1 << offset))
> 
> Hmm this patch is hard to read, maybe break it into two patches?
> 
> First you could do a patch to prepare thing by introducing
> BANK_USED and LINE_USED.
> 
>> +static int gpio_is_input(struct gpio_bank *bank, int mask)
>> +{
>> +	void __iomem *reg = bank->base + bank->regs->direction;
>> +
>> +	return __raw_readl(reg) & mask;
>> +}
> 
> And also move gpio_is_input() around in the first patch.
> 
> Then the second patch for the fix would probably be much
> easier to read.
>

Sure will split in more patches, I just wanted to keep in one patch since it was
a RFC but it seems that the change makes sense so I'll post it as a proper
patch-set.

> Regards,
> 
> Tony
> 

Thanks a lot for your feedback and best regards,
Javier