[PATCH] can: m_can: Set pinmux into "default" state on start

[PATCH] can: m_can: Set pinmux into "default" state on start

[Marek Vasut]

On systems like stm32mp1 where pins have both "default" and "sleep" pinmux
states in DT, the pins are in "sleep" state by default. Switch the pins into
the "default" state when bringing the interface up in m_can_start(), because
otherwise no CAN communication is possible. This replicates the behavior of
the resume path, which does the same.

Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Alexandre Torgue <alexandre.torgue@st.com>
Cc: Marc Kleine-Budde <mkl@pengutronix.de>
Cc: Wolfgang Grandegger <wg@grandegger.com>
To: linux-can@vger.kernel.org
---
 drivers/net/can/m_can/m_can.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/net/can/m_can/m_can.c b/drivers/net/can/m_can/m_can.c
index 02c5795b7393..76fadf2b8ac2 100644
--- a/drivers/net/can/m_can/m_can.c
+++ b/drivers/net/can/m_can/m_can.c
@@ -1247,6 +1247,8 @@ static void m_can_start(struct net_device *dev)
 	/* basic m_can configuration */
 	m_can_chip_config(dev);
 
+	pinctrl_pm_select_default_state(cdev->dev);
+
 	cdev->can.state = CAN_STATE_ERROR_ACTIVE;
 
 	m_can_enable_all_interrupts(cdev);
-- 
2.25.1

Re: [PATCH] can: m_can: Set pinmux into "default" state on start

[Marek Vasut]

On 5/31/20 8:13 PM, Marek Vasut wrote:
> On systems like stm32mp1 where pins have both "default" and "sleep" pinmux
> states in DT, the pins are in "sleep" state by default. Switch the pins into
> the "default" state when bringing the interface up in m_can_start(), because
> otherwise no CAN communication is possible. This replicates the behavior of
> the resume path, which does the same.
> 
> Signed-off-by: Marek Vasut <marex@denx.de>
> Cc: Alexandre Torgue <alexandre.torgue@st.com>
> Cc: Marc Kleine-Budde <mkl@pengutronix.de>
> Cc: Wolfgang Grandegger <wg@grandegger.com>
> To: linux-can@vger.kernel.org
> ---
>  drivers/net/can/m_can/m_can.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/drivers/net/can/m_can/m_can.c b/drivers/net/can/m_can/m_can.c
> index 02c5795b7393..76fadf2b8ac2 100644
> --- a/drivers/net/can/m_can/m_can.c
> +++ b/drivers/net/can/m_can/m_can.c
> @@ -1247,6 +1247,8 @@ static void m_can_start(struct net_device *dev)
>  	/* basic m_can configuration */
>  	m_can_chip_config(dev);
>  
> +	pinctrl_pm_select_default_state(cdev->dev);
> +
>  	cdev->can.state = CAN_STATE_ERROR_ACTIVE;
>  
>  	m_can_enable_all_interrupts(cdev);
> 

Any news on this ?

Re: [PATCH] can: m_can: Set pinmux into "default" state on start

[Marek Vasut]

On 8/9/20 1:01 PM, Marek Vasut wrote:
> On 5/31/20 8:13 PM, Marek Vasut wrote:
>> On systems like stm32mp1 where pins have both "default" and "sleep" pinmux
>> states in DT, the pins are in "sleep" state by default. Switch the pins into
>> the "default" state when bringing the interface up in m_can_start(), because
>> otherwise no CAN communication is possible. This replicates the behavior of
>> the resume path, which does the same.
>>
>> Signed-off-by: Marek Vasut <marex@denx.de>
>> Cc: Alexandre Torgue <alexandre.torgue@st.com>
>> Cc: Marc Kleine-Budde <mkl@pengutronix.de>
>> Cc: Wolfgang Grandegger <wg@grandegger.com>
>> To: linux-can@vger.kernel.org
>> ---
>>  drivers/net/can/m_can/m_can.c | 2 ++
>>  1 file changed, 2 insertions(+)
>>
>> diff --git a/drivers/net/can/m_can/m_can.c b/drivers/net/can/m_can/m_can.c
>> index 02c5795b7393..76fadf2b8ac2 100644
>> --- a/drivers/net/can/m_can/m_can.c
>> +++ b/drivers/net/can/m_can/m_can.c
>> @@ -1247,6 +1247,8 @@ static void m_can_start(struct net_device *dev)
>>  	/* basic m_can configuration */
>>  	m_can_chip_config(dev);
>>  
>> +	pinctrl_pm_select_default_state(cdev->dev);
>> +
>>  	cdev->can.state = CAN_STATE_ERROR_ACTIVE;
>>  
>>  	m_can_enable_all_interrupts(cdev);
>>
> 
> Any news on this ?

Almost another month has passed by, ping ?

Re: [PATCH] can: m_can: Set pinmux into "default" state on start

[Marek Vasut]

On 9/1/20 8:56 PM, Marek Vasut wrote:
> On 8/9/20 1:01 PM, Marek Vasut wrote:
>> On 5/31/20 8:13 PM, Marek Vasut wrote:
>>> On systems like stm32mp1 where pins have both "default" and "sleep" pinmux
>>> states in DT, the pins are in "sleep" state by default. Switch the pins into
>>> the "default" state when bringing the interface up in m_can_start(), because
>>> otherwise no CAN communication is possible. This replicates the behavior of
>>> the resume path, which does the same.
>>>
>>> Signed-off-by: Marek Vasut <marex@denx.de>
>>> Cc: Alexandre Torgue <alexandre.torgue@st.com>
>>> Cc: Marc Kleine-Budde <mkl@pengutronix.de>
>>> Cc: Wolfgang Grandegger <wg@grandegger.com>
>>> To: linux-can@vger.kernel.org
>>> ---
>>>  drivers/net/can/m_can/m_can.c | 2 ++
>>>  1 file changed, 2 insertions(+)
>>>
>>> diff --git a/drivers/net/can/m_can/m_can.c b/drivers/net/can/m_can/m_can.c
>>> index 02c5795b7393..76fadf2b8ac2 100644
>>> --- a/drivers/net/can/m_can/m_can.c
>>> +++ b/drivers/net/can/m_can/m_can.c
>>> @@ -1247,6 +1247,8 @@ static void m_can_start(struct net_device *dev)
>>>  	/* basic m_can configuration */
>>>  	m_can_chip_config(dev);
>>>  
>>> +	pinctrl_pm_select_default_state(cdev->dev);
>>> +
>>>  	cdev->can.state = CAN_STATE_ERROR_ACTIVE;
>>>  
>>>  	m_can_enable_all_interrupts(cdev);
>>>
>>
>> Any news on this ?
> 
> Almost another month has passed by, ping ?

Bump ?