[PATCH] gpio: mvebu: Fix cause computation in irq handler

[PATCH] gpio: mvebu: Fix cause computation in irq handler

[Gregory CLEMENT]

When switching to regmap, the way to compute the irq cause was
reorganized. However while doing it, a typo was introduced: a 'xor'
replaced a 'and'.

This lead to wrong behavior in the interrupt handler ans one of the
symptom was wrong irq handler called on the Armada 388 GP:
"->handle_irq():  c016303c,
handle_bad_irq+0x0/0x278
->irq_data.chip(): c0b0ec0c,
0xc0b0ec0c
->action():   (null)
   IRQ_NOPROBE set
 IRQ_NOREQUEST set
unexpected IRQ trap at vector 00
irq 0, desc: ee804800, depth: 1, count: 0, unhandled: 0"

Fixes: 2233bf7a92e7 ("gpio: mvebu: switch to regmap for register access")
Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
---
 drivers/gpio/gpio-mvebu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpio/gpio-mvebu.c b/drivers/gpio/gpio-mvebu.c
index e338c3743562..45c65f805fd6 100644
--- a/drivers/gpio/gpio-mvebu.c
+++ b/drivers/gpio/gpio-mvebu.c
@@ -557,7 +557,7 @@ static void mvebu_gpio_irq_handler(struct irq_desc *desc)
 	edge_cause = mvebu_gpio_read_edge_cause(mvchip);
 	edge_mask  = mvebu_gpio_read_edge_mask(mvchip);
 
-	cause = (data_in ^ level_mask) | (edge_cause & edge_mask);
+	cause = (data_in & level_mask) | (edge_cause & edge_mask);
 
 	for (i = 0; i < mvchip->chip.ngpio; i++) {
 		int irq;
-- 
2.13.2

Re: [PATCH] gpio: mvebu: Fix cause computation in irq handler

[Gregory CLEMENT]

Hi Linus,
 
 On mer., juil. 12 2017, Gregory CLEMENT <gregory.clement@free-electrons.com> wrote:

> When switching to regmap, the way to compute the irq cause was
> reorganized. However while doing it, a typo was introduced: a 'xor'
> replaced a 'and'.
>
> This lead to wrong behavior in the interrupt handler ans one of the
> symptom was wrong irq handler called on the Armada 388 GP:
> "->handle_irq():  c016303c,
> handle_bad_irq+0x0/0x278
> ->irq_data.chip(): c0b0ec0c,
> 0xc0b0ec0c
> ->action():   (null)
>    IRQ_NOPROBE set
>  IRQ_NOREQUEST set
> unexpected IRQ trap at vector 00
> irq 0, desc: ee804800, depth: 1, count: 0, unhandled: 0"
>
> Fixes: 2233bf7a92e7 ("gpio: mvebu: switch to regmap for register access")
> Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com>

Would it be possible to push this fix on the current kernel (v4.13-rc) ?

Indeed currently due to this bug, the Armada 388 GP board does not boot
anymore because this board uses a gpio interrupt. So it prevents to have
continuous integration on this one with kci for instance.

Thanks,

Gregory


> ---
>  drivers/gpio/gpio-mvebu.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpio/gpio-mvebu.c b/drivers/gpio/gpio-mvebu.c
> index e338c3743562..45c65f805fd6 100644
> --- a/drivers/gpio/gpio-mvebu.c
> +++ b/drivers/gpio/gpio-mvebu.c
> @@ -557,7 +557,7 @@ static void mvebu_gpio_irq_handler(struct irq_desc *desc)
>  	edge_cause = mvebu_gpio_read_edge_cause(mvchip);
>  	edge_mask  = mvebu_gpio_read_edge_mask(mvchip);
>  
> -	cause = (data_in ^ level_mask) | (edge_cause & edge_mask);
> +	cause = (data_in & level_mask) | (edge_cause & edge_mask);
>  
>  	for (i = 0; i < mvchip->chip.ngpio; i++) {
>  		int irq;
> -- 
> 2.13.2
>

-- 
Gregory Clement, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

Re: [PATCH] gpio: mvebu: Fix cause computation in irq handler

[Gregory CLEMENT]

Hi Linus,
 
 On lun., juil. 24 2017, Gregory CLEMENT <gregory.clement@free-electrons.com> wrote:

> Hi Linus,
>  
>  On mer., juil. 12 2017, Gregory CLEMENT <gregory.clement@free-electrons.com> wrote:
>
>> When switching to regmap, the way to compute the irq cause was
>> reorganized. However while doing it, a typo was introduced: a 'xor'
>> replaced a 'and'.
>>
>> This lead to wrong behavior in the interrupt handler ans one of the
>> symptom was wrong irq handler called on the Armada 388 GP:
>> "->handle_irq():  c016303c,
>> handle_bad_irq+0x0/0x278
>> ->irq_data.chip(): c0b0ec0c,
>> 0xc0b0ec0c
>> ->action():   (null)
>>    IRQ_NOPROBE set
>>  IRQ_NOREQUEST set
>> unexpected IRQ trap at vector 00
>> irq 0, desc: ee804800, depth: 1, count: 0, unhandled: 0"
>>
>> Fixes: 2233bf7a92e7 ("gpio: mvebu: switch to regmap for register access")
>> Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
>
> Would it be possible to push this fix on the current kernel (v4.13-rc) ?
>
> Indeed currently due to this bug, the Armada 388 GP board does not boot
> anymore because this board uses a gpio interrupt. So it prevents to have
> continuous integration on this one with kci for instance.

I saw that yesterday you appplied patches on gpio/fixes but it seems
that you missed this one, because most of the patch you're applied were
actually posted after this one.

Thanks,

Gregory
>
>
>> ---
>>  drivers/gpio/gpio-mvebu.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpio/gpio-mvebu.c b/drivers/gpio/gpio-mvebu.c
>> index e338c3743562..45c65f805fd6 100644
>> --- a/drivers/gpio/gpio-mvebu.c
>> +++ b/drivers/gpio/gpio-mvebu.c
>> @@ -557,7 +557,7 @@ static void mvebu_gpio_irq_handler(struct irq_desc *desc)
>>  	edge_cause = mvebu_gpio_read_edge_cause(mvchip);
>>  	edge_mask  = mvebu_gpio_read_edge_mask(mvchip);
>>  
>> -	cause = (data_in ^ level_mask) | (edge_cause & edge_mask);
>> +	cause = (data_in & level_mask) | (edge_cause & edge_mask);
>>  
>>  	for (i = 0; i < mvchip->chip.ngpio; i++) {
>>  		int irq;
>> -- 
>> 2.13.2
>>
>
> -- 
> Gregory Clement, Free Electrons
> Kernel, drivers, real-time and embedded Linux
> development, consulting, training and support.
> http://free-electrons.com
>
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

-- 
Gregory Clement, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

Re: [PATCH] gpio: mvebu: Fix cause computation in irq handler

[Linus Walleij]

On Wed, Jul 12, 2017 at 1:22 PM, Gregory CLEMENT
<gregory.clement@free-electrons.com> wrote:

> When switching to regmap, the way to compute the irq cause was
> reorganized. However while doing it, a typo was introduced: a 'xor'
> replaced a 'and'.
>
> This lead to wrong behavior in the interrupt handler ans one of the
> symptom was wrong irq handler called on the Armada 388 GP:
> "->handle_irq():  c016303c,
> handle_bad_irq+0x0/0x278
> ->irq_data.chip(): c0b0ec0c,
> 0xc0b0ec0c
> ->action():   (null)
>    IRQ_NOPROBE set
>  IRQ_NOREQUEST set
> unexpected IRQ trap at vector 00
> irq 0, desc: ee804800, depth: 1, count: 0, unhandled: 0"
>
> Fixes: 2233bf7a92e7 ("gpio: mvebu: switch to regmap for register access")
> Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com>

Patch applied for fixes.

Yours,
Linus Walleij