i2c: at91: switched to resume/suspend callbacks.

Message ID 1540203006-805-1-git-send-email-andrei.stefanescu@microchip.com
State New
Headers show
Series
  • i2c: at91: switched to resume/suspend callbacks.
Related show

Commit Message

Andrei.Stefanescu@microchip.com Oct. 22, 2018, 10:17 a.m.
In the previous version of the driver resume/suspend_noirq callbacks
were used. Because of this, when resuming from suspend-to-ram,
an I2C (belonging to a FLEXCOM) would resume before FLEXCOM.
The first read on the I2C bus would then result in a timeout.

This patch switches to resume/suspend callbacks which are 
called after FLEXCOM resumes. FLEXCOM, SPI and USART drivers use
resume/suspend callbacks.

Signed-off-by: Andrei Stefanescu <andrei.stefanescu@microchip.com>
---
 drivers/i2c/busses/i2c-at91.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

Comments

Ludovic Desroches Nov. 7, 2018, 8:39 a.m. | #1
On Mon, Oct 22, 2018 at 12:17:47PM +0200, Andrei Stefanescu - M50506 wrote:
> In the previous version of the driver resume/suspend_noirq callbacks
> were used. Because of this, when resuming from suspend-to-ram,
> an I2C (belonging to a FLEXCOM) would resume before FLEXCOM.
> The first read on the I2C bus would then result in a timeout.
> 
> This patch switches to resume/suspend callbacks which are 
> called after FLEXCOM resumes. FLEXCOM, SPI and USART drivers use
> resume/suspend callbacks.
> 
> Signed-off-by: Andrei Stefanescu <andrei.stefanescu@microchip.com>
I can't figure out why we use the _noirq variant. When patches for PM
stuff were sent, suspend/resume callbacks were used but in the latest
version it moved to the _noirq variant without explanation.

Excepting if someone has an argument to keep the _noirq variant,
Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com> 

Thanks

Regards

Ludovic

> ---
>  drivers/i2c/busses/i2c-at91.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/i2c/busses/i2c-at91.c b/drivers/i2c/busses/i2c-at91.c
> index bfd1fdf..81f7b94 100644
> --- a/drivers/i2c/busses/i2c-at91.c
> +++ b/drivers/i2c/busses/i2c-at91.c
> @@ -1174,7 +1174,7 @@ static int at91_twi_runtime_resume(struct device *dev)
>  	return clk_prepare_enable(twi_dev->clk);
>  }
>  
> -static int at91_twi_suspend_noirq(struct device *dev)
> +static int at91_twi_suspend(struct device *dev)
>  {
>  	if (!pm_runtime_status_suspended(dev))
>  		at91_twi_runtime_suspend(dev);
> @@ -1182,7 +1182,7 @@ static int at91_twi_suspend_noirq(struct device *dev)
>  	return 0;
>  }
>  
> -static int at91_twi_resume_noirq(struct device *dev)
> +static int at91_twi_resume(struct device *dev)
>  {
>  	struct at91_twi_dev *twi_dev = dev_get_drvdata(dev);
>  	int ret;
> @@ -1202,8 +1202,8 @@ static int at91_twi_resume_noirq(struct device *dev)
>  }
>  
>  static const struct dev_pm_ops at91_twi_pm = {
> -	.suspend_noirq	= at91_twi_suspend_noirq,
> -	.resume_noirq	= at91_twi_resume_noirq,
> +	.suspend	= at91_twi_suspend,
> +	.resume		= at91_twi_resume,
>  	.runtime_suspend	= at91_twi_runtime_suspend,
>  	.runtime_resume		= at91_twi_runtime_resume,
>  };
> -- 
> 2.7.4
>
Andrei.Stefanescu@microchip.com Nov. 15, 2018, 2:18 p.m. | #2
Hello Wolfram,

Could you please tell me if there are any changes that I should make
to the patch?

Thank you,
Andrei Stefanescu

On 07.11.2018 10:39, Ludovic Desroches wrote:
> On Mon, Oct 22, 2018 at 12:17:47PM +0200, Andrei Stefanescu - M50506 wrote:
>> In the previous version of the driver resume/suspend_noirq callbacks
>> were used. Because of this, when resuming from suspend-to-ram,
>> an I2C (belonging to a FLEXCOM) would resume before FLEXCOM.
>> The first read on the I2C bus would then result in a timeout.
>>
>> This patch switches to resume/suspend callbacks which are
>> called after FLEXCOM resumes. FLEXCOM, SPI and USART drivers use
>> resume/suspend callbacks.
>>
>> Signed-off-by: Andrei Stefanescu <andrei.stefanescu@microchip.com>
> I can't figure out why we use the _noirq variant. When patches for PM
> stuff were sent, suspend/resume callbacks were used but in the latest
> version it moved to the _noirq variant without explanation.
>
> Excepting if someone has an argument to keep the _noirq variant,
> Acked-by: Ludovic Desroches <ludovic.desroches@microchip.com>
>
> Thanks
>
> Regards
>
> Ludovic
>
>> ---
>>   drivers/i2c/busses/i2c-at91.c | 8 ++++----
>>   1 file changed, 4 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/i2c/busses/i2c-at91.c b/drivers/i2c/busses/i2c-at91.c
>> index bfd1fdf..81f7b94 100644
>> --- a/drivers/i2c/busses/i2c-at91.c
>> +++ b/drivers/i2c/busses/i2c-at91.c
>> @@ -1174,7 +1174,7 @@ static int at91_twi_runtime_resume(struct device *dev)
>>   	return clk_prepare_enable(twi_dev->clk);
>>   }
>>   
>> -static int at91_twi_suspend_noirq(struct device *dev)
>> +static int at91_twi_suspend(struct device *dev)
>>   {
>>   	if (!pm_runtime_status_suspended(dev))
>>   		at91_twi_runtime_suspend(dev);
>> @@ -1182,7 +1182,7 @@ static int at91_twi_suspend_noirq(struct device *dev)
>>   	return 0;
>>   }
>>   
>> -static int at91_twi_resume_noirq(struct device *dev)
>> +static int at91_twi_resume(struct device *dev)
>>   {
>>   	struct at91_twi_dev *twi_dev = dev_get_drvdata(dev);
>>   	int ret;
>> @@ -1202,8 +1202,8 @@ static int at91_twi_resume_noirq(struct device *dev)
>>   }
>>   
>>   static const struct dev_pm_ops at91_twi_pm = {
>> -	.suspend_noirq	= at91_twi_suspend_noirq,
>> -	.resume_noirq	= at91_twi_resume_noirq,
>> +	.suspend	= at91_twi_suspend,
>> +	.resume		= at91_twi_resume,
>>   	.runtime_suspend	= at91_twi_runtime_suspend,
>>   	.runtime_resume		= at91_twi_runtime_resume,
>>   };
>> -- 
>> 2.7.4
>>

Patch

diff --git a/drivers/i2c/busses/i2c-at91.c b/drivers/i2c/busses/i2c-at91.c
index bfd1fdf..81f7b94 100644
--- a/drivers/i2c/busses/i2c-at91.c
+++ b/drivers/i2c/busses/i2c-at91.c
@@ -1174,7 +1174,7 @@  static int at91_twi_runtime_resume(struct device *dev)
 	return clk_prepare_enable(twi_dev->clk);
 }
 
-static int at91_twi_suspend_noirq(struct device *dev)
+static int at91_twi_suspend(struct device *dev)
 {
 	if (!pm_runtime_status_suspended(dev))
 		at91_twi_runtime_suspend(dev);
@@ -1182,7 +1182,7 @@  static int at91_twi_suspend_noirq(struct device *dev)
 	return 0;
 }
 
-static int at91_twi_resume_noirq(struct device *dev)
+static int at91_twi_resume(struct device *dev)
 {
 	struct at91_twi_dev *twi_dev = dev_get_drvdata(dev);
 	int ret;
@@ -1202,8 +1202,8 @@  static int at91_twi_resume_noirq(struct device *dev)
 }
 
 static const struct dev_pm_ops at91_twi_pm = {
-	.suspend_noirq	= at91_twi_suspend_noirq,
-	.resume_noirq	= at91_twi_resume_noirq,
+	.suspend	= at91_twi_suspend,
+	.resume		= at91_twi_resume,
 	.runtime_suspend	= at91_twi_runtime_suspend,
 	.runtime_resume		= at91_twi_runtime_resume,
 };