Re: [PATCH 2/3] irqchip/stm32-exti: add STM32MP13 support

[Alexandre TORGUE <alexandre.torgue@foss.st.com>]

Hi Marc

On 12/8/21 2:41 PM, Marc Zyngier wrote:
> On Wed, 08 Dec 2021 13:04:55 +0000,
> Alexandre Torgue <alexandre.torgue@foss.st.com> wrote:
>>
>> Enhance stm32-exti driver to support STM32MP13 SoC. This SoC uses the same
>> hardware version than STM32MP15. Only EXTI line mapping is changed and
>> following EXTI lines are supported: GPIO, RTC, I2C[1-5], UxART[1-8],
>> USBH_EHCI, USBH_OHCI, USB_OTG, LPTIM[1-5], ETH[1-2].
>>
>> Signed-off-by: Alexandre Torgue <alexandre.torgue@foss.st.com>
>>
>> diff --git a/drivers/irqchip/irq-stm32-exti.c b/drivers/irqchip/irq-stm32-exti.c
>> index b7cb2da71888..9d18f47040eb 100644
>> --- a/drivers/irqchip/irq-stm32-exti.c
>> +++ b/drivers/irqchip/irq-stm32-exti.c
>> @@ -214,6 +214,48 @@ static const struct stm32_desc_irq stm32mp1_desc_irq[] = {
>>   	{ .exti = 73, .irq_parent = 129, .chip = &stm32_exti_h_chip },
>>   };
>>   
>> +static const struct stm32_desc_irq stm32mp13_desc_irq[] = {
>> +	{ .exti = 0, .irq_parent = 6, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 1, .irq_parent = 7, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 2, .irq_parent = 8, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 3, .irq_parent = 9, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 4, .irq_parent = 10, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 5, .irq_parent = 24, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 6, .irq_parent = 65, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 7, .irq_parent = 66, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 8, .irq_parent = 67, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 9, .irq_parent = 68, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 10, .irq_parent = 41, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 11, .irq_parent = 43, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 12, .irq_parent = 77, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 13, .irq_parent = 78, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 14, .irq_parent = 106, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 15, .irq_parent = 109, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 16, .irq_parent = 1, .chip = &stm32_exti_h_chip },
>> +	{ .exti = 19, .irq_parent = 3, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 21, .irq_parent = 32, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 22, .irq_parent = 34, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 23, .irq_parent = 73, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 24, .irq_parent = 93, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 25, .irq_parent = 114, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 26, .irq_parent = 38, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 27, .irq_parent = 39, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 28, .irq_parent = 40, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 29, .irq_parent = 72, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 30, .irq_parent = 53, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 31, .irq_parent = 54, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 32, .irq_parent = 83, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 33, .irq_parent = 84, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 44, .irq_parent = 96, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 47, .irq_parent = 92, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 48, .irq_parent = 116, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 50, .irq_parent = 117, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 52, .irq_parent = 118, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 53, .irq_parent = 119, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 68, .irq_parent = 63, .chip = &stm32_exti_h_chip_direct },
>> +	{ .exti = 70, .irq_parent = 98, .chip = &stm32_exti_h_chip_direct },
>> +};
> 
> Why does the driver need to carry these tables? This sort of
> information should really come from DT, instead of being hardcoded in
> the driver and bloating it for no reason. This all has a funny taste
> of the board files we used to have pre-DT.
>

There are absolutely no reason to have it in driver. Honestly It has 
been done in this way to have minimal changes adding this new SoC 
support (and it's not smart, I agree).

I think it is better to abandon this series. I will create a new one 
which moves mapping table for MP15 and adds MP13 support to.

thanks
Alex



> 	M.
>