Re: [RFC PATCH v2 2/4] serial: mxs-auart: Use helpers for gpio irqs

["Janusz Użycki" <j.uzycki@elproma.com.pl>]

W dniu 2015-01-13 o 10:35, Uwe Kleine-König pisze:
> Hello,
>
> On Tue, Jan 13, 2015 at 10:29:44AM +0100, Janusz Użycki wrote:
>> W dniu 2015-01-13 o 09:08, Uwe Kleine-König pisze:
>>> Hello,
>>>
>>> On Sat, Jan 10, 2015 at 03:32:44PM +0100, Janusz Uzycki wrote:
>>>> The patch updates mxs-auart driver to use new mctrl_gpio helpers for
>>>> gpio irqs. The code is much simpler now.
>>>>
>>>> Signed-off-by: Janusz Uzycki <j.uzycki@elproma.com.pl>
>>>> ---
>>>>
>>>> There is no changes since v1 (rebased only).
>>>>
>>>> ---
>>>>   drivers/tty/serial/mxs-auart.c | 133 ++++-------------------------------------
>>>>   1 file changed, 13 insertions(+), 120 deletions(-)
>>> Very nice!
>>>
>>>> diff --git a/drivers/tty/serial/mxs-auart.c b/drivers/tty/serial/mxs-auart.c
>>>> index ec553f8..2ddba69 100644
>>>> --- a/drivers/tty/serial/mxs-auart.c
>>>> +++ b/drivers/tty/serial/mxs-auart.c
>>>> [...]
>>>> @@ -483,18 +458,9 @@ static void mxs_auart_enable_ms(struct uart_port *port)
>>>>   	s->ms_irq_enabled = true;
>>>> -	if (s->gpio_irq[UART_GPIO_CTS] >= 0)
>>>> -		enable_irq(s->gpio_irq[UART_GPIO_CTS]);
>>>> -	/* TODO: enable AUART_INTR_CTSMIEN otherwise */
>>>> -
>>>> -	if (s->gpio_irq[UART_GPIO_DSR] >= 0)
>>>> -		enable_irq(s->gpio_irq[UART_GPIO_DSR]);
>>>> -
>>>> -	if (s->gpio_irq[UART_GPIO_RI] >= 0)
>>>> -		enable_irq(s->gpio_irq[UART_GPIO_RI]);
>>>> -
>>>> -	if (s->gpio_irq[UART_GPIO_DCD] >= 0)
>>>> -		enable_irq(s->gpio_irq[UART_GPIO_DCD]);
>>>> +	mctrl_gpio_enable_ms(s->gpios);
>>>> +	/* TODO: enable AUART_INTR_CTSMIEN
>>>> +	 * if s->gpios->irq[UART_GPIO_CTS] == 0 */
>>> What is the problem here? For the other lines nothing needs to be done?
>>> This comment doesn't match the coding style.
>> Right, the comment should be rather:
>> /* TODO: enable AUART_INTR_CTSMIEN
>>   * if (!mctrl_gpio_is_gpio(atmel_port->gpios, UART_GPIO_CTS)) */
> I'd say:
>
> 	/*
> 	 * TODO: enable AUART_INTR_CTSMIEN if CTS isn't handled by
> 	 * mctrl_gpio.
> 	 */

exactly, thanks

>
>> In this place I marked that CTSMIEN should be switched on
>> enable/disable_ms if CTS
>> is not a gpio. The driver enables CTSMIEN forever what is wrong but
>> I can't test it
>> and I don't need it so I've just marked the fact in the comment.
> That's what I thought. You're not affected because CTS is a gpio for
> you (or not?)?  What would be the effect otherwise?

Yes, my CTS is a gpio.
CTSMIEN control CTS signal of auart block. There is a choice in DT:
- use auart block's CTS: hardware flow control works for all baud rates, 
DMA can be used
- use gpio as CTS: hardware flow control is limited, DMA disabled but 
CTS line is
   not limited by hardware pinmux
Both options can't be set at once. I workarounded auart block's CTS irq 
handler in condition:
"if (CTS_AT_AUART() && s->ms_irq_enabled)". Support by _ms would be more 
elegance
but as I wrote I couldn't test all cases. Therefore the code for auart 
block's CTS is preserved.

best regards
Janusz

>
> Best regards
> Uwe
>

--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html