drivers/gpio: Altera soft IP GPIO driver
Subsystems (3)
PATCH 1/1 drivers/gpio: Altera soft IP GPIO driver
From: Tien Hock Loh <[email protected]>
Adds a new driver for Altera soft GPIO IP. The driver is able to do read/write and allows GPIO to be a interrupt controller.
Tested on Altera GHRD on interrupt handling and IO.
--- .../devicetree/bindings/gpio/gpio-altera.txt | 37 +++ drivers/gpio/Kconfig | 7 + drivers/gpio/Makefile | 1 + drivers/gpio/gpio-altera.c | 363 +++++++++++++++++++++ 4 files changed, 408 insertions(+) create mode 100644 Documentation/devicetree/bindings/gpio/gpio-altera.txt create mode 100644 drivers/gpio/gpio-altera.c
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +37 -0
@@ -0,0 +1,37 @@ +Altera GPIO controller bindings + +Required properties: +- compatible: + - "altr,pio-1.0" +- reg: Physical base address and length of the controller's registers. +- #gpio-cells : Should be two. + - first cell is the pin number + - second cell is used to specify optional parameters (unused) +- gpio-controller : Marks the device node as a GPIO controller. +- #interrupt-cells : Should be 1. +- interrupts: Specify the interrupt. +- interrupt-controller: Mark the device node as an interrupt controller + The first cell is the GPIO number. + +Altera GPIO specific properties: +- width: Width of the GPIO bank, range from 1-32 +- level_trigger: Specifies whether the GPIO interrupt is level trigger. + This field is required if the Altera GPIO controller used has IRQ enabled. + +Note: If the Altera GPIO is set to be built as a module, peripherals that uses +Altera GPIO as interrupt-parent should be a module so that the peripheral +doesn't get initialized before Altera GPIO is initialized. + +Example: + +gpio_altr: gpio_altr { + compatible = "altr,pio-1.0"; + reg = <0xff200000 0x10>; + interrupts = <0 45 4>; + width = <32>; + level_trigger = <0>; + #gpio-cells = <2>; + gpio-controller; + #interrupt-cells = <1>; + interrupt-controller; +};
drivers/gpio/Kconfig +7 -0
@@ -119,6 +119,13 @@ config GPIO_GENERIC_PLATFORM help Say yes here to support basic platform_device memory-mapped GPIO controllers. +config GPIO_ALTERA + tristate "Altera GPIO" + select IRQ_DOMAIN + depends on OF_GPIO + help + Say yes here to support the Altera PIO device. + config GPIO_IT8761E tristate "IT8761E GPIO support" depends on X86 # unconditional access to IO space.
drivers/gpio/Makefile +1 -0
@@ -14,6 +14,7 @@ obj-$(CONFIG_GPIO_74X164) += gpio-74x164.o obj-$(CONFIG_GPIO_ADNP) += gpio-adnp.o obj-$(CONFIG_GPIO_ADP5520) += gpio-adp5520.o obj-$(CONFIG_GPIO_ADP5588) += gpio-adp5588.o +obj-$(CONFIG_GPIO_ALTERA) += gpio-altera.o obj-$(CONFIG_GPIO_AMD8111) += gpio-amd8111.o obj-$(CONFIG_GPIO_ARIZONA) += gpio-arizona.o obj-$(CONFIG_GPIO_BT8XX) += gpio-bt8xx.o
drivers/gpio/gpio-altera.c new file +363 -0
363 lines not on this page · raw message
Signature · 1 line
1.7.11.GIT
Hi,
Are the codes good for upstream?
Thanks.
-----Original Message-----
From: Tien Hock Loh
Sent: Tuesday, March 12, 2013 1:38 PM
To: Linus Walleij; Rob Landley
Cc: [email protected]; [email protected]; Tien Hock Loh
Subject: [PATCH 1/1] drivers/gpio: Altera soft IP GPIO driver
From: Tien Hock Loh <[email protected]>
Adds a new driver for Altera soft GPIO IP. The driver is able to do read/write and allows GPIO to be a interrupt controller.
Tested on Altera GHRD on interrupt handling and IO.
--- .../devicetree/bindings/gpio/gpio-altera.txt | 37 +++ drivers/gpio/Kconfig | 7 + drivers/gpio/Makefile | 1 + drivers/gpio/gpio-altera.c | 363 +++++++++++++++++++++ 4 files changed, 408 insertions(+) create mode 100644 Documentation/devicetree/bindings/gpio/gpio-altera.txt create mode 100644 drivers/gpio/gpio-altera.c
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +37 -0
@@ -0,0 +1,37 @@ +Altera GPIO controller bindings + +Required properties: +- compatible: + - "altr,pio-1.0" +- reg: Physical base address and length of the controller's registers. +- #gpio-cells : Should be two. + - first cell is the pin number + - second cell is used to specify optional parameters (unused) +- gpio-controller : Marks the device node as a GPIO controller. +- #interrupt-cells : Should be 1. +- interrupts: Specify the interrupt. +- interrupt-controller: Mark the device node as an interrupt controller + The first cell is the GPIO number. + +Altera GPIO specific properties: +- width: Width of the GPIO bank, range from 1-32 +- level_trigger: Specifies whether the GPIO interrupt is level trigger. + This field is required if the Altera GPIO controller used has IRQ enabled. + +Note: If the Altera GPIO is set to be built as a module, peripherals +that uses Altera GPIO as interrupt-parent should be a module so that +the peripheral doesn't get initialized before Altera GPIO is initialized. + +Example: + +gpio_altr: gpio_altr { + compatible = "altr,pio-1.0"; + reg = <0xff200000 0x10>; + interrupts = <0 45 4>; + width = <32>; + level_trigger = <0>; + #gpio-cells = <2>; + gpio-controller; + #interrupt-cells = <1>; + interrupt-controller; +};
drivers/gpio/Kconfig +7 -0
@@ -119,6 +119,13 @@ config GPIO_GENERIC_PLATFORM help Say yes here to support basic platform_device memory-mapped GPIO controllers. +config GPIO_ALTERA + tristate "Altera GPIO" + select IRQ_DOMAIN + depends on OF_GPIO + help + Say yes here to support the Altera PIO device. + config GPIO_IT8761E tristate "IT8761E GPIO support" depends on X86 # unconditional access to IO space.
drivers/gpio/Makefile +1 -0
@@ -14,6 +14,7 @@ obj-$(CONFIG_GPIO_74X164) += gpio-74x164.o obj-$(CONFIG_GPIO_ADNP) += gpio-adnp.o obj-$(CONFIG_GPIO_ADP5520) += gpio-adp5520.o obj-$(CONFIG_GPIO_ADP5588) += gpio-adp5588.o +obj-$(CONFIG_GPIO_ALTERA) += gpio-altera.o obj-$(CONFIG_GPIO_AMD8111) += gpio-amd8111.o obj-$(CONFIG_GPIO_ARIZONA) += gpio-arizona.o obj-$(CONFIG_GPIO_BT8XX) += gpio-bt8xx.o
drivers/gpio/gpio-altera.c new file +348 -0
348 lines not on this page · raw message
--
1.7.11.GIT
Confidentiality Notice.
This message may contain information that is confidential or otherwise protected from disclosure. If you are not the intended recipient, you are hereby notified that any use, disclosure, dissemination, distribution, or copying of this message, or any attachments, is strictly prohibited. If you have received this message in error, please advise the sender by reply e-mail, and delete the message and any attachments. Thank you.
Quoted 10 lines · On Tue, Mar 12, 2013 at 6:37 AM, <[email protected]> wrote:
From: Tien Hock Loh <[email protected]>
Adds a new driver for Altera soft GPIO IP. The driver is able to do read/write and allows GPIO to be a interrupt controller.
Tested on Altera GHRD on interrupt handling and IO.
Signed-off-by Tien Hock Loh <[email protected]>--- .../devicetree/bindings/gpio/gpio-altera.txt | 37 +++
Patches containing device tree bindings shall be CC:ed to devicetree-discuss (added on CC here, remember at next posting).
(...)
Quoted 6 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +0
-0
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 8
Quoted 8 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +8
-0@@ -0,0 +1,37 @@ +Altera GPIO controller bindings + +Required properties: +- compatible: + - "altr,pio-1.0" +- reg: Physical base address and length of the controller's registers. +- #gpio-cells : Should be two. + - first cell is the pin number
What does that mean? The GPIO subsystem is not referring to pins, it's about "gpios" which are just a handler or number.
It can be a local offset into the numberspace representiong the GPIO lines at this block, I guess that is what is meant here.
(...)
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 19
Quoted 4 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +4
-0+Altera GPIO specific properties: +- width: Width of the GPIO bank, range from 1-32 +- level_trigger: Specifies whether the GPIO interrupt is level trigger. + This field is required if the Altera GPIO controller used has IRQ enabled.
This is a configuration for the irc_chip portion of the driver I guess, isn't the usual custom that you don't set up things like this from the device tree, but instead use the kernels facilities on request_irq() such as:
#define IRQF_TRIGGER_NONE 0x00000000 #define IRQF_TRIGGER_RISING 0x00000001 #define IRQF_TRIGGER_FALLING 0x00000002 #define IRQF_TRIGGER_HIGH 0x00000004 #define IRQF_TRIGGER_LOW 0x00000008
?
I.e. that this is something you do at runtime and not a static config from the device tree?
Or shall this be percieved as some kind of default setting?
(...)
Documentation/devicetree/bindings/gpio/gpio-altera.txt
Quoted 7 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +6
-0+++ b/drivers/gpio/gpio-altera.c +struct altera_gpio_chip { + struct of_mm_gpio_chip mmchip; + struct irq_domain *irq; /* GPIO controller IRQ number */ + spinlock_t gpio_lock; /* Lock used for synchronization */ + int level_trigger;
I doubt this member of the struct. The irq core shall keep track of stuff like this.
Documentation/devicetree/bindings/gpio/gpio-altera.txt
Quoted 2 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +2
-0+ int hwirq; +};
(then follows some real nice code using irqdomain in a professional way, thanks for doing this!)
Documentation/devicetree/bindings/gpio/gpio-altera.txt
Quoted 7 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +7
-0+static int altera_gpio_irq_set_type(struct irq_data *d, + unsigned int type) +{ + if (type == IRQ_TYPE_NONE) + return 0; + return -EINVAL; +}
So if the chip supports setting edge vs level trigging, this is where you implement it, not as device tree properties.
Documentation/devicetree/bindings/gpio/gpio-altera.txt
Quoted 12 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +12
-0+static void altera_gpio_irq_handler(unsigned int irq, struct irq_desc *desc) +{ + struct altera_gpio_chip *altera_gc = irq_desc_get_handler_data(desc); + struct irq_chip *chip = irq_desc_get_chip(desc); + struct of_mm_gpio_chip *mm_gc = &altera_gc->mmchip; + unsigned long status; + + int base; + chip->irq_mask(&desc->irq_data); + + if (altera_gc->level_trigger) + status = readl(mm_gc->regs + ALTERA_GPIO_DATA);
So on the level trigged variant you read the ALTERA_GPIO_DATA and discard the result? Is it some kind of latch register?
Documentation/devicetree/bindings/gpio/gpio-altera.txt
Quoted 4 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +4
-0+ else { + status = readl(mm_gc->regs + ALTERA_GPIO_EDGE_CAP); + writel(status, mm_gc->regs + ALTERA_GPIO_EDGE_CAP); + }
So you need to get the affected register depending on the type set in the irq descriptor instead.
Documentation/devicetree/bindings/gpio/gpio-altera.txt
Quoted 7 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +7
-0+ status &= readl(mm_gc->regs + ALTERA_GPIO_IRQ_MASK); + for (base = 0; base < mm_gc->gc.ngpio; base++) { + if ((1 << base) & status) { + generic_handle_irq( + irq_linear_revmap(altera_gc->irq, base)); + } + }
When we looked at this for some IRQ controllers we realized we had to write the loop like so:
while ((stat = readl_relaxed(vic->base + VIC_IRQ_STATUS))) { (...)
}
Only to avoid missing transient IRQs occuring while we were processing another IRQ. (Like, you handle IRQ0, then IRQ4, and while you're handling IRQ4, IRQ0 triggers again and you miss it.
Make sure this does not apply to you...
Documentation/devicetree/bindings/gpio/gpio-altera.txt
Quoted 7 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +7
-0+ if (of_property_read_u32(node, "level_trigger", ®)) { + ret = -EINVAL; + pr_err("%s: level_trigger value not set in device tree\n", + node->full_name); + goto teardown; + } + altera_gc->level_trigger = reg;
So I'm suspicious about this.
Yours,
Linus Walleij
Quoted 7 lines · Linus Walleij
Quoted 5 lines · On Tue, Mar 12, 2013 at 6:37 AM, <[email protected]> wrote:
From: Tien Hock Loh <[email protected]>
Adds a new driver for Altera soft GPIO IP. The driver is able to do read/write and allows GPIO to be a interrupt controller.
Quoted 8 lines
Quoted 5 lines
Tested on Altera GHRD on interrupt handling and IO.
Signed-off-by Tien Hock Loh <[email protected]>--- .../devicetree/bindings/gpio/gpio-altera.txt | 37 +++Patches containing device tree bindings shall be CC:ed to devicetree-discuss (added on CC here, remember at next posting).
Noted. Thanks for the info.
Quoted 20 lines
(...)
Quoted 6 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +0
-0Documentation/devicetree/bindings/gpio/gpio-altera.txt line 8
Quoted 8 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +8
-0@@ -0,0 +1,37 @@ +Altera GPIO controller bindings + +Required properties: +- compatible: + - "altr,pio-1.0" +- reg: Physical base address and length of the controller's registers. +- #gpio-cells : Should be two. + - first cell is the pin numberWhat does that mean? The GPIO subsystem is not referring to pins, it's about "gpios" which are just a handler or number.
Quoted 3 lines
It can be a local offset into the numberspace representiong the GPIO lines at this block, I guess that is what is meant here.
Yes that's what I was trying to describe. I'll fix this in the next patch.
Quoted 18 lines
(...)
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 19
Quoted 4 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +4
-0+Altera GPIO specific properties: +- width: Width of the GPIO bank, range from 1-32 +- level_trigger: Specifies whether the GPIO interrupt is level trigger. + This field is required if the Altera GPIO controller used has IRQ enabled.This is a configuration for the
irc_chipportion of the driver I guess, isn't the usual custom that you don't set up things like this from the device tree, but instead use the kernels facilities onrequest_irq()such as:#define IRQF_TRIGGER_NONE 0x00000000 #define IRQF_TRIGGER_RISING 0x00000001 #define IRQF_TRIGGER_FALLING 0x00000002 #define IRQF_TRIGGER_HIGH 0x00000004 #define IRQF_TRIGGER_LOW 0x00000008
Quoted 6 lines
?
I.e. that this is something you do at runtime and not a static config from the device tree?
Or shall this be percieved as some kind of default setting?
The Altera GPIO IP cannot be software configurable. It is determined by the tool that generates the hardware, and thus I've put this as device tree parameter. If I understand correctly, if we implement using irq_set_type, it should be software configurable.
Technically it can still be implemented in irq_set_type, I'm just not sure if it is the correct way, because the trigger type cannot be "set" in Altera GPIO. Please let me know if you think this should still be implemented as irq_set_type.
Quoted 4 lines
(...)
Quoted 1 line
+++
b/drivers/gpio/gpio-altera.c
Quoted 9 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt
Quoted 5 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +5
-0+struct altera_gpio_chip { + struct of_mm_gpio_chip mmchip; + struct irq_domain *irq; /* GPIO controller IRQ number */ + spinlock_t gpio_lock; /* Lock used for synchronization */ + int level_trigger;I doubt this member of the struct. The irq core shall keep track of stuff like this.
We shouldn't be adding anything to the gpio chip struct and has to follow exactly how the irq core defines it?
Quoted 6 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt
Quoted 2 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +2
-0+ int hwirq; +};(then follows some real nice code using irqdomain in a professional way, thanks for doing this!)
Quoted 11 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt
Quoted 7 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +7
-0+static int altera_gpio_irq_set_type(struct irq_data *d, + unsigned int type) +{ + if (type == IRQ_TYPE_NONE) + return 0; + return -EINVAL; +}So if the chip supports setting edge vs level trigging, this is where you implement it, not as device tree properties.
Discussed above
Quoted 16 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt
Quoted 13 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +12
-0+static void altera_gpio_irq_handler(unsigned int irq, struct irq_desc *desc) +{ + struct altera_gpio_chip *altera_gc = irq_desc_get_handler_data(desc); + struct irq_chip *chip = irq_desc_get_chip(desc); + struct of_mm_gpio_chip *mm_gc = &altera_gc->mmchip; + unsigned long status; + + int base; + chip->irq_mask(&desc->irq_data); + + if (altera_gc->level_trigger) + status = readl(mm_gc->regs + ALTERA_GPIO_DATA);So on the level trigged variant you read the
ALTERA_GPIO_DATAand discard the result? Is it some kind of latch register?
No, the status isn't discarded. It will be & with the peripheral's interrupt mask to get the interrupts that needs to be handled.
Quoted 9 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt
Quoted 6 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +4
-0+ else { + status = readl(mm_gc->regs + ALTERA_GPIO_EDGE_CAP); + writel(status, mm_gc->regs + ALTERA_GPIO_EDGE_CAP); + }So you need to get the affected register depending on the type set in the irq descriptor instead.
Yes you are correct.
Quoted 16 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt
Quoted 8 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +7
-0+ status &= readl(mm_gc->regs + ALTERA_GPIO_IRQ_MASK); + for (base = 0; base < mm_gc->gc.ngpio; base++) { + if ((1 << base) & status) { + generic_handle_irq( + irq_linear_revmap(altera_gc->irq, base)); + } + }When we looked at this for some IRQ controllers we realized we had to write the loop like so:
while ((stat =
readl_relaxed(vic->base + VIC_IRQ_STATUS))) { (...)
}
Quoted 6 lines
Only to avoid missing transient IRQs occuring while we were processing another IRQ. (Like, you handle IRQ0, then IRQ4, and while you're handling IRQ4, IRQ0 triggers again and you miss it.
Make sure this does not apply to you...
Noted. Thanks for the info. I'll look into this.
Quoted 10 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt
Quoted 8 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +7
-0+ if (of_property_read_u32(node, "level_trigger", ®)) { + ret = -EINVAL; + pr_err("%s: level_trigger value not set in device tree\n", + node->full_name); + goto teardown; + } + altera_gc->level_trigger = reg;So I'm suspicious about this.
Discussed above.
Quoted 4 lines
Yours,
Linus Walleij
Thanks.
Tien Hock Loh
Quoted 29 lines · Tien Hock Loh
Quoted 17 lines
(...)
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 19
Quoted 4 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +4
-0+Altera GPIO specific properties: +- width: Width of the GPIO bank, range from 1-32 +- level_trigger: Specifies whether the GPIO interrupt is level trigger. + This field is required if the Altera GPIO controller used has IRQ enabled.This is a configuration for the
irc_chipportion of the driver I guess, isn't the usual custom that you don't set up things like this from the device tree, but instead use the kernels facilities onrequest_irq()such as:#define IRQF_TRIGGER_NONE 0x00000000 #define IRQF_TRIGGER_RISING 0x00000001 #define IRQF_TRIGGER_FALLING 0x00000002 #define IRQF_TRIGGER_HIGH 0x00000004 #define IRQF_TRIGGER_LOW 0x00000008Quoted 6 lines
?
I.e. that this is something you do at runtime and not a static config from the device tree?
Or shall this be percieved as some kind of default setting?
The Altera GPIO IP cannot be software configurable. It is determined by the tool that generates the hardware, and thus I've put this as device tree parameter. If I understand correctly, if we implement using
irq_set_type, it should be software configurable.
Quoted 4 lines
Technically it can still be implemented in
irq_set_type, I'm just not sure if it is the correct way, because the trigger type cannot be "set" in Altera GPIO. Please let me know if you think this should still be implemented asirq_set_type.
Hm I must be getting things wrong. It's correct that the driver needs to keep track of if a certain IRQ line has a certain trigger characteristic. So the member in this struct:
+struct altera_gpio_chip {
+ struct of_mm_gpio_chip mmchip; + struct irq_domain *irq; /* GPIO controller IRQ number */ + spinlock_t gpio_lock; /* Lock used for synchronization */ + int level_trigger;
Is OK. But it should definately be a bool instead.
Then the irq_set_type() should check if the requested type is available.
You state that some interrupts are edge triggered, so this is clearly wrong:
+static int altera_gpio_irq_set_type(struct irq_data *d,
+ unsigned int type)
+{
+ if (type == IRQ_TYPE_NONE)
+ return 0;
+ return -EINVAL;
+}
I think it should rather be something like:
static int altera_gpio_irq_set_type(struct irq_data *d,
unsigned int type)
{
struct altera_gpio_chip *a = irq_data_get_irq_chip_data(d);
if (a->level_trigger) {
if (type == IRQ_TYPE_LEVEL_HIGH)
/* Configure the IRQ for high level */
else if (type == IRQ_TYPE_LEVEL_LOW)
/* Configure the IRQ for low level */
else
return -EINVAL;
} else {
/* assume it's edge triggered? */
if (type == IRQ_TYPE_EDGE_RISING)
/* Configure for rising edge */
else if (type == IRQ_TYPE_EDGE_FALLING)
/* Configure for falling edge */
else
return -EINVAL;
}
}As you can see from above a few new questions arise.
When you say an IRQ is "level trigged" does it actually mean it is high or low level triggered? Or is this somehow configurable?
If it is *not* edge triggered, is it then implicitly edge triggered?
If it is then implicitly edge triggered, does it trigger on rising or falling edges or is this configurable?
As you see, the .irq_set_type needs to reflect what the hardware can actually react to, and return errors on the rest.
The current implementation which states that the hardware cannot react to any IRQ is clearly wrong.
Yours,
Linus Walleij