Armadaxp GPIO interrupts

Armadaxp GPIO interrupts

[Andrew Lunn]

On Tue, Aug 04, 2015 at 08:52:17PM +0530, raghu MG wrote:
> Hello,
> 
> I am working on a card which as GPIOs connected to external I/O's. The
> board consists of ARMADAXP 78460 host cpu.
> 
> Board currently runs Linux-4.1 with modified  armada-xp-gp.dtb for ArmadaXP.
> I enabled "orion-gpio" driver to initialize GPIOs as given in
> armada-xp-mv78460.

What kernel configuration are you using? Do you have the same problem
with multi_v7_defconfig and mvebu_v7_defconfig?
 
> The driver while initializing calls irq_domain_add_simple which throws up
> following warning
> "irq: Cannot allocate irq_descs @ IRQ47, assuming pre-allocated"
> The warning repeats for next set(32-63  & 64-66).
>
> Also the GPIO IRQs are not getting listed in cat /proc/interrupts

irq_domain_add_simple() returning an error is fatal for the probe. The
driver will not be loaded, so more than interrupts will be missing,
all the gpios will be missing.

    Andrew

Armadaxp GPIO interrupts

[raghu MG]

Hi Andrew,

Thanks for your quick response,dont mind for the delay(India,US timings).
I checked with both multi_v7_defconfig & mvebu_v7_defconfig , the result is same
"irq: Cannot allocate irq_descs @ IRQ47, assuming pre-allocated"

Pasting here the GPIO driver probe debug statements

armada-xp-pinctrl f1018000.pin-ctrl: registered pinctrl driver
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 680
mvebu_gpio_probe 682 match->name=
mvebu_gpio_probe 683 match->compatible=marvell,orion-gpio
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 727
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 734 res->start=f1018100
mvchip->membase=cf8c0100
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 747 soc_variant=1
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 785
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 788
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 805
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 840
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 844 mvchip->irqbase=47
irq: Cannot allocate irq_descs @ IRQ47, assuming pre-allocated
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 850 mvchip->domain->name=(null)
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 859
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 680
mvebu_gpio_probe 682 match->name=
mvebu_gpio_probe 683 match->compatible=marvell,orion-gpio
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 727
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 734 res->start=f1018140
mvchip->membase=cf8c2140
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 747 soc_variant=1
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 785
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 788
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 805
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 840
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 844 mvchip->irqbase=79
irq: Cannot allocate irq_descs @ IRQ79, assuming pre-allocated
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 850 mvchip->domain->name=(null)
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 859
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 680
mvebu_gpio_probe 682 match->name=
mvebu_gpio_probe 683 match->compatible=marvell,orion-gpio
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 727
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 734 res->start=f1018180
mvchip->membase=cf8c4180
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 747 soc_variant=1
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 785
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 788
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 805
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 840
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 844 mvchip->irqbase=111
irq: Cannot allocate irq_descs @ IRQ111, assuming pre-allocated
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 850 mvchip->domain->name=(null)
drivers/gpio/gpio-mvebu.c mvebu_gpio_probe 859



I will try to dig in more information in the probe & irq_domain_add_simple


On Tue, Aug 4, 2015 at 9:04 PM, Andrew Lunn <andrew@lunn.ch> wrote:
> On Tue, Aug 04, 2015 at 08:52:17PM +0530, raghu MG wrote:
>> Hello,
>>
>> I am working on a card which as GPIOs connected to external I/O's. The
>> board consists of ARMADAXP 78460 host cpu.
>>
>> Board currently runs Linux-4.1 with modified  armada-xp-gp.dtb for ArmadaXP.
>> I enabled "orion-gpio" driver to initialize GPIOs as given in
>> armada-xp-mv78460.
>
> What kernel configuration are you using? Do you have the same problem
> with multi_v7_defconfig and mvebu_v7_defconfig?
>
>> The driver while initializing calls irq_domain_add_simple which throws up
>> following warning
>> "irq: Cannot allocate irq_descs @ IRQ47, assuming pre-allocated"
>> The warning repeats for next set(32-63  & 64-66).
>>
>> Also the GPIO IRQs are not getting listed in cat /proc/interrupts
>
> irq_domain_add_simple() returning an error is fatal for the probe. The
> driver will not be loaded, so more than interrupts will be missing,
> all the gpios will be missing.
>
>     Andrew

Armadaxp GPIO interrupts

[Andrew Lunn]

On Wed, Aug 05, 2015 at 08:31:27PM +0530, raghu MG wrote:
> Hi Andrew,
> 
> The issue I found was  irq_alloc_descs is called twice
> 1st time inside   mvebu_gpio_probe & 2nd time inside irq_domain_add_simple.

So i'm somewhat confused. I just tested on my Kirkwood and 370
platforms, and an Armada XP WRT1900AC using 4.2.0-rc5, all which use
the same gpio driver.

They boot fine, the gpio driver is loaded, and the interrupt for the
gpio connected to a button are listed in /proc/interrupts. The
gpio-key driver is also able to use the gpio and the input-events
program does show events when i press the button.

So what is different with your platform?

   Andrew

Armadaxp GPIO interrupts

[raghu MG]

Hi Andrew,

My platform/board is ATCA carrier card which is getting interface to I/O card.
The I/O card is hot-pluggable  or Jack-in or Jack out.
Also I/O card consists of hot-plugable SFP interface.
These events are triggered using GPIO's

So the events the carrier card recieves are
1) If I/O card is either Jacked-in or Jacked-out ,respective GPIO
pin-x triggers an event
2) If SFP is pushed into I/O card the respective GPIO pin-y triggers an event.




Another observation about  "irq: Cannot allocate irq_descs @ IRQ47/79/111".
This is warning & irq_descs are already allocated , the overall
initialization/probe function doesn't get effected.
mvebu_gpio_irq_handler is only called if I register a another handler
at irq=82/83/84/85/87/88/89/90/92. I am registering this handler using
minimal kernel module.
My understanding of Linux IRQ/interrupt mechanism is limited to
request_irq, but registering domain & then IRQ is bit not clear.

Regards
Raghu




On Wed, Aug 5, 2015 at 9:51 PM, Andrew Lunn <andrew@lunn.ch> wrote:
> On Wed, Aug 05, 2015 at 08:31:27PM +0530, raghu MG wrote:
>> Hi Andrew,
>>
>> The issue I found was  irq_alloc_descs is called twice
>> 1st time inside   mvebu_gpio_probe & 2nd time inside irq_domain_add_simple.
>
> So i'm somewhat confused. I just tested on my Kirkwood and 370
> platforms, and an Armada XP WRT1900AC using 4.2.0-rc5, all which use
> the same gpio driver.
>
> They boot fine, the gpio driver is loaded, and the interrupt for the
> gpio connected to a button are listed in /proc/interrupts. The
> gpio-key driver is also able to use the gpio and the input-events
> program does show events when i press the button.
>
> So what is different with your platform?
>
>    Andrew

Armadaxp GPIO interrupts

[Andrew Lunn]

> mvebu_gpio_irq_handler is only called if I register a another handler
> at irq=82/83/84/85/87/88/89/90/92. I am registering this handler using
> minimal kernel module.

This is totally wrong. The gpio driver needs these interrupts, and
will register a chained interrupt handle for these. Don't mess around
with them. Here is the code in the driver:

        /* Setup the interrupt handlers. Each chip can have up to 4
         * interrupt handlers, with each handler dealing with 8 GPIO
         * pins. */
        for (i = 0; i < 4; i++) {
                int irq = platform_get_irq(pdev, i);

                if (irq < 0)
                        continue;
                irq_set_handler_data(irq, mvchip);
                irq_set_chained_handler(irq, mvebu_gpio_irq_handler);
        }

	Andrew

Armadaxp GPIO interrupts

[raghu MG]

Ok,I think I need to understand more about this gpio driver.

As you said its registering chained handler,but why are they(IRQs) not
visible in cat /proc/interrupts.
What could be the reason.?

Do I need to further initialize marvell GPIO registers to trigger
these events. The driver is unmasking all interrupts in probe function
                writel_relaxed(0, mvchip->membase + GPIO_EDGE_CAUSE_OFF);
                writel_relaxed(0, mvchip->membase + GPIO_EDGE_MASK_OFF);
                writel_relaxed(0, mvchip->membase + GPIO_LEVEL_MASK_OFF);


Do I need to change the polarity in polarity register to suit the
board requirements in probe function.

I will try this in the morning.


Regards
Raghu

On Thu, Aug 6, 2015 at 6:59 PM, Andrew Lunn <andrew@lunn.ch> wrote:
>> mvebu_gpio_irq_handler is only called if I register a another handler
>> at irq=82/83/84/85/87/88/89/90/92. I am registering this handler using
>> minimal kernel module.
>
> This is totally wrong. The gpio driver needs these interrupts, and
> will register a chained interrupt handle for these. Don't mess around
> with them. Here is the code in the driver:
>
>         /* Setup the interrupt handlers. Each chip can have up to 4
>          * interrupt handlers, with each handler dealing with 8 GPIO
>          * pins. */
>         for (i = 0; i < 4; i++) {
>                 int irq = platform_get_irq(pdev, i);
>
>                 if (irq < 0)
>                         continue;
>                 irq_set_handler_data(irq, mvchip);
>                 irq_set_chained_handler(irq, mvebu_gpio_irq_handler);
>         }
>
>         Andrew

Armadaxp GPIO interrupts

[Andrew Lunn]

On Thu, Aug 06, 2015 at 10:28:12PM +0530, raghu MG wrote:
> Ok,I think I need to understand more about this gpio driver.
> 
> As you said its registering chained handler,but why are they(IRQs) not
> visible in cat /proc/interrupts.

chained interrupts are never visible there. All you see are leaf
interrupts, i.e. the device interrupts. Here is my Armada XP based
WRT1900AC:

$ cat /proc/interrupts 
            CPU0       CPU1       
 16:    8959833   12783597  armada_370_xp_irq   5 Level     armada_370_xp_per_cpu_tick
 19:          0          0  armada_370_xp_irq  31 Level     mv64xxx_i2c
 20:       1453          0  armada_370_xp_irq  41 Level     serial
 26:          0          0  armada_370_xp_irq  45 Level     ehci_hcd:usb3
 27:       9682          0  armada_370_xp_irq   8 Level     mvneta
 28:          4          0  armada_370_xp_irq  10 Level     mvneta
 29:      35078          0  armada_370_xp_irq  55 Level     f10a0000.sata
 30:       3839          0  armada_370_xp_irq 113 Level     f10d0000.nand
 70:          0          0  f1018140.gpio   0 Edge      WPS
 71:         20          0  f1018140.gpio   1 Edge      Factory Reset Button
 90:          2          0  armada_370_xp_irq  51 Level     f1060900.xor
 91:          2          0  armada_370_xp_irq  52 Level     f1060900.xor
 92:          2          0  armada_370_xp_irq  94 Level     f10f0900.xor
 93:          2          0  armada_370_xp_irq  95 Level     f10f0900.xor
 94:          0          0  armada_370_xp_msi_irq   0 Edge      xhci_hcd

Notice the two f1018140.gpio, which are the buttons.

> Do I need to further initialize marvell GPIO registers to trigger
> these events.

Nope, just the normal gpio API and all will work, as demonstrated by
the two buttons on my board.

> Do I need to change the polarity in polarity register to suit the
> board requirements in probe function.

Nope, you put that in the DT file:

        gpio_keys {
                compatible = "gpio-keys";
                #address-cells = <1>;
                #size-cells = <0>;
                pinctrl-0 = <&keys_pin>;
                pinctrl-names = "default";

                button at 1 {
                        label = "WPS";
                        linux,code = <KEY_WPS_BUTTON>;
                        gpios = <&gpio1 0 GPIO_ACTIVE_HIGH>;
                };

                button at 2 {
                        label = "Factory Reset Button";
                        linux,code = <KEY_RESTART>;
                        gpios = <&gpio1 1 GPIO_ACTIVE_HIGH>;
                };
        };

You can change GPIO_ACTIVE_HIGH to GPIO_ACTIVE_LOW if that is what you
need.

	Andrew
 

Armadaxp GPIO interrupts

[raghu MG]

Ok...
my bad never realised DT entries are necessary.
will try this & post the results.

Regards
Raghu

On Thu, Aug 6, 2015 at 10:43 PM, Andrew Lunn <andrew@lunn.ch> wrote:
> On Thu, Aug 06, 2015 at 10:28:12PM +0530, raghu MG wrote:
>> Ok,I think I need to understand more about this gpio driver.
>>
>> As you said its registering chained handler,but why are they(IRQs) not
>> visible in cat /proc/interrupts.
>
> chained interrupts are never visible there. All you see are leaf
> interrupts, i.e. the device interrupts. Here is my Armada XP based
> WRT1900AC:
>
> $ cat /proc/interrupts
>             CPU0       CPU1
>  16:    8959833   12783597  armada_370_xp_irq   5 Level     armada_370_xp_per_cpu_tick
>  19:          0          0  armada_370_xp_irq  31 Level     mv64xxx_i2c
>  20:       1453          0  armada_370_xp_irq  41 Level     serial
>  26:          0          0  armada_370_xp_irq  45 Level     ehci_hcd:usb3
>  27:       9682          0  armada_370_xp_irq   8 Level     mvneta
>  28:          4          0  armada_370_xp_irq  10 Level     mvneta
>  29:      35078          0  armada_370_xp_irq  55 Level     f10a0000.sata
>  30:       3839          0  armada_370_xp_irq 113 Level     f10d0000.nand
>  70:          0          0  f1018140.gpio   0 Edge      WPS
>  71:         20          0  f1018140.gpio   1 Edge      Factory Reset Button
>  90:          2          0  armada_370_xp_irq  51 Level     f1060900.xor
>  91:          2          0  armada_370_xp_irq  52 Level     f1060900.xor
>  92:          2          0  armada_370_xp_irq  94 Level     f10f0900.xor
>  93:          2          0  armada_370_xp_irq  95 Level     f10f0900.xor
>  94:          0          0  armada_370_xp_msi_irq   0 Edge      xhci_hcd
>
> Notice the two f1018140.gpio, which are the buttons.
>
>> Do I need to further initialize marvell GPIO registers to trigger
>> these events.
>
> Nope, just the normal gpio API and all will work, as demonstrated by
> the two buttons on my board.
>
>> Do I need to change the polarity in polarity register to suit the
>> board requirements in probe function.
>
> Nope, you put that in the DT file:
>
>         gpio_keys {
>                 compatible = "gpio-keys";
>                 #address-cells = <1>;
>                 #size-cells = <0>;
>                 pinctrl-0 = <&keys_pin>;
>                 pinctrl-names = "default";
>
>                 button at 1 {
>                         label = "WPS";
>                         linux,code = <KEY_WPS_BUTTON>;
>                         gpios = <&gpio1 0 GPIO_ACTIVE_HIGH>;
>                 };
>
>                 button at 2 {
>                         label = "Factory Reset Button";
>                         linux,code = <KEY_RESTART>;
>                         gpios = <&gpio1 1 GPIO_ACTIVE_HIGH>;
>                 };
>         };
>
> You can change GPIO_ACTIVE_HIGH to GPIO_ACTIVE_LOW if that is what you
> need.
>
>         Andrew
>

Armadaxp GPIO interrupts

[raghu MG]

Hi Andrew,
I added these entries into armada-xp-gp.dtb
 gpio_rtm {
                compatible = "gpio-rtm";
                #address-cells = <1>;
                #size-cells = <0>;
                pinctrl-0 = <&keys_pin>;
                pinctrl-names = "default";

                 rtm at 57 {
                        label = "RTM JACKINOUT";
                        gpios = <&gpio1 26 GPIO_ACTIVE_LOW>; /* GPIO 57 */
                };
        };

So I understand the leaf nodes are specific purpose of
GPIOs(input:keyboard,mouse,touchpad etc).
1)For the board I am working needs different implementation for GPIO
57 unlike input devices drivers driving GPIOs.
2)To implement the functionality I need to write driver specific to my
board which will implement event specific to requirements.
I hope I am going right here, can you name a simple driver(kernel
source) where it would make my driver development easy.

Thanks & Regards
Raghu

Armadaxp GPIO interrupts

[Andrew Lunn]

On Sat, Aug 08, 2015 at 02:24:04AM +0530, raghu MG wrote:
> Hi Andrew,
> I added these entries into armada-xp-gp.dtb
>  gpio_rtm {
>                 compatible = "gpio-rtm";
>                 #address-cells = <1>;
>                 #size-cells = <0>;
>                 pinctrl-0 = <&keys_pin>;
>                 pinctrl-names = "default";
> 
>                  rtm at 57 {
>                         label = "RTM JACKINOUT";
>                         gpios = <&gpio1 26 GPIO_ACTIVE_LOW>; /* GPIO 57 */
>                 };
>         };
> 
> So I understand the leaf nodes are specific purpose of
> GPIOs(input:keyboard,mouse,touchpad etc).
> 1)For the board I am working needs different implementation for GPIO
> 57 unlike input devices drivers driving GPIOs.
> 2)To implement the functionality I need to write driver specific to my
> board which will implement event specific to requirements.
> I hope I am going right here, can you name a simple driver(kernel
> source) where it would make my driver development easy.

Given the name "RTM JACKINOUT", how about

sound/soc/soc-jack.c

	Andrew

Armadaxp GPIO interrupts

[raghu MG]

Hi Andrew,

The issue I found was  irq_alloc_descs is called twice
1st time inside   mvebu_gpio_probe & 2nd time inside irq_domain_add_simple.

The warning is thrown up when irq_alloc_descs is called 2nd
time(irq_domain_add_simple) & bitmap_find_next_zero_area gives new
value of start which is not equal to irq & the following condition
returns with error.

irq_alloc_descs(....)
.................
....................
..............
start = bitmap_find_next_zero_area(
allocated_irqs, IRQ_BITMAP_BITS,
                                           from, cnt, 0);
ret=-EEXIST;
if (irq >=0 && start != irq){
                goto err;
        }
...........................
................
err:
       mutex_unlock(&sparse_irq_lock);
        return ret;


I could bypass the error/warning by giving calling
irq_domain_add_simple(np, mvchip->chip.ngpio,
                                               0,
                                               &irq_domain_simple_ops,
                                               mvchip);
I  passed zero instead of mvchip->irqbase as first_irq .
The warning is not appearing.& I think driver is registered in kernel.

But driver & gpio entry is not found in /proc/interrupts.

The GPIO ISR is not getting invoked by driver.
But I wrote a simple module where I register simple handler & then
both mvebu_gpio_irq_handler & modules registered handler is called.

My doubt is whether the gpio-driver in the kernel needs bit more
tweaking to suit the board & devices connected to gpio.

I would appreciate bit of guidance here

On Tue, Aug 4, 2015 at 9:04 PM, Andrew Lunn <andrew@lunn.ch> wrote:
> On Tue, Aug 04, 2015 at 08:52:17PM +0530, raghu MG wrote:
>> Hello,
>>
>> I am working on a card which as GPIOs connected to external I/O's. The
>> board consists of ARMADAXP 78460 host cpu.
>>
>> Board currently runs Linux-4.1 with modified  armada-xp-gp.dtb for ArmadaXP.
>> I enabled "orion-gpio" driver to initialize GPIOs as given in
>> armada-xp-mv78460.
>
> What kernel configuration are you using? Do you have the same problem
> with multi_v7_defconfig and mvebu_v7_defconfig?
>
>> The driver while initializing calls irq_domain_add_simple which throws up
>> following warning
>> "irq: Cannot allocate irq_descs @ IRQ47, assuming pre-allocated"
>> The warning repeats for next set(32-63  & 64-66).
>>
>> Also the GPIO IRQs are not getting listed in cat /proc/interrupts
>
> irq_domain_add_simple() returning an error is fatal for the probe. The
> driver will not be loaded, so more than interrupts will be missing,
> all the gpios will be missing.
>
>     Andrew