[PATCH V2 1/3] gpio: pca953x: make the register access by GPIO bank

[Gregory CLEMENT]

On 01/08/2013 09:58 AM, Gregory CLEMENT wrote:
> On 01/08/2013 09:32 AM, Maxime Ripard wrote:
>> Hi Gregory,
> 
> Hi Maxime,
> 
> thanks for testing
>>
>> On 07/01/2013 23:51, Gregory CLEMENT wrote:
>>> -static int pca953x_write_reg(struct pca953x_chip *chip, int reg, u32 val)
>>> +static int pca953x_read_single(struct pca953x_chip *chip, int reg, u32 *val,
>>> +				int off)
>>> +{
>>> +	int ret;
>>> +	int bank_shift = fls((chip->gpio_chip.ngpio - 1) / BANK_SZ);
>>> +	int offset = off / BANK_SZ;
>>> +
>>> +	ret = i2c_smbus_read_byte_data(chip->client,
>>> +				(reg << bank_shift) + offset);
>>> +	*val = ret;
>>> +
>>> +	if (ret < 0) {
>>> +		dev_err(&chip->client->dev, "failed reading register\n", reg);
>>
>> This triggers a warning, since you have an argument but nothing corresponding to it in your format string.
> 
> OK this one is easy, I didn't finished to clean up the message.
> 
>>
>>>  static irqreturn_t pca953x_irq_handler(int irq, void *devid)
>>>  {
>>>  	struct pca953x_chip *chip = devid;
>>> -	u32 pending;
>>> -	u32 level;
>>> +	u8 pending[MAX_BANK];
>>> +	u8 level;
>>> +	int i;
>>>  
>>> -	pending = pca953x_irq_pending(chip);
>>> -
>>> -	if (!pending)
>>> +	if (!pca953x_irq_pending(chip, pending))
>>>  		return IRQ_HANDLED;
>>>  
>>> -	do {
>>> -		level = __ffs(pending);
>>> -		handle_nested_irq(irq_find_mapping(chip->domain, level));
>>> -
>>> -		pending &= ~(1 << level);
>>> -	} while (pending);
>>> +	for (i = 0; i < NBANK(chip); i++) {
>>> +		do {
>>> +			level = __ffs(pending[i]);
>>> +			handle_nested_irq(irq_find_mapping(chip->domain,
>>> +							level + (BANK_SZ * i)));
>>> +			pending[i] &= ~(1 << level);
>>> +		} while (pending[i]);
>>> +	}
>>>  
>>>  	return IRQ_HANDLED;
>>>  }
>>
>> This triggers the following warning when an interrupt is raised:
>>
>> [   30.773500] ------------[ cut here ]------------
>> [   30.778843] WARNING: at /home/tmp/linux/kernel/irq/irqdomain.c:137 irq_domain_legacy_revmap+0x2c/0x48()
>> [   30.788375] Modules linked in:
>> [   30.791531] [<c0014100>] (unwind_backtrace+0x0/0xf0) from [<c001c878>] (warn_slowpath_common+0x4c/0x64)
>> [   30.801125] [<c001c878>] (warn_slowpath_common+0x4c/0x64) from [<c001c8ac>] (warn_slowpath_null+0x1c/0x24)
>> [   30.810968] [<c001c8ac>] (warn_slowpath_null+0x1c/0x24) from [<c006b5e0>] (irq_domain_legacy_revmap+0x2c/0x48)
>> [   30.821187] [<c006b5e0>] (irq_domain_legacy_revmap+0x2c/0x48) from [<c01c3734>] (pca953x_irq_handler+0x16c/0x1ac)
>> [   30.831656] [<c01c3734>] (pca953x_irq_handler+0x16c/0x1ac) from [<c0068688>] (irq_thread+0xd0/0x124)
>> [   30.841000] [<c0068688>] (irq_thread+0xd0/0x124) from [<c003c238>] (kthread+0xa4/0xb0)
>> [   30.849125] [<c003c238>] (kthread+0xa4/0xb0) from [<c000ed48>] (ret_from_fork+0x14/0x2c)
>> [   30.857375] ---[ end trace 09584b7a73100a49 ]---
>>
> 
> Humm it seems that this warning is caused by attempting to use an out of range hwirq.
> I need to investigate it more in details.

Unfortunately I didn't receive information on Mirabox to be able to use
interrupt. So I will need you help to debug this part.

Could you try the branch gpio-pca9505-debug  at
https://github.com/MISL-EBU-System-SW/mainline-public.git ?

I suspect that we use a GPIO above 8 to generate the interrupt, am I right?

Thanks,

Gregory