drivers/gpio: Altera soft IP GPIO driver and devicetree binding
Subsystems (3)
PATCH v7 1/1 drivers/gpio: Altera soft IP GPIO driver and devicetree binding
From: Tien Hock Loh <[email protected]>
Add driver support for Altera GPIO soft IP, including interrupts and I/O. Tested on Altera CV SoC board using dipsw and LED using LED framework.
--- .../devicetree/bindings/gpio/gpio-altera.txt | 44 +++ drivers/gpio/Kconfig | 7 + drivers/gpio/Makefile | 1 + drivers/gpio/gpio-altera.c | 419 +++++++++++++++++++++ 4 files changed, 471 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 +44 -0
@@ -0,0 +1,44 @@ +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 1 + - The first cell is the gpio offset number +- gpio-controller : Marks the device node as a GPIO controller. +- #interrupt-cells : Should be 1. + - The first cell is the GPIO offset number within the GPIO controller. +- interrupts: Specify the interrupt. +- interrupt-controller: Mark the device node as an interrupt controller + +Altera GPIO specific required properties: +- altr,interrupt_trigger: Specifies the interrupt trigger type the GPIO + hardware is synthesized. This field is required if the Altera GPIO controller + used has IRQ enabled as the interrupt type is not software controlled, + but hardware synthesized. Required if GPIO is used as an interrupt + controller. The value is defined in <dt-bindings/interrupt-controller/irq.h> + Only the following flags are supported: + IRQ_TYPE_EDGE_RISING + IRQ_TYPE_EDGE_FALLING + IRQ_TYPE_EDGE_BOTH + IRQ_TYPE_LEVEL_HIGH + +Altera GPIO specific optional properties: +- altr,gpio-bank-width: Width of the GPIO bank. This defines how many pins the + GPIO device has. Ranges between 1-32. Optional and defaults to 32 is not + specified. + +Example: + +gpio_altr: gpio@40000 { + compatible = "altr,pio-1.0"; + reg = <0xff200000 0x10>; + interrupts = <0 45 4>; + altr,gpio-bank-width = <32>; + altr,interrupt_trigger = <IRQ_TYPE_EDGE_RISING>; + #gpio-cells = <1>; + gpio-controller; + #interrupt-cells = <1>; + interrupt-controller; +};
drivers/gpio/Kconfig +7 -0
@@ -128,6 +128,13 @@ 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_BCM_KONA) += gpio-bcm-kona.o
drivers/gpio/gpio-altera.c new file +419 -0
419 lines not on this page · raw message
Signature · 1 line
1.7.11.GIT
Quoted 6 lines · On Mon, Mar 3, 2014 at 6:27 PM, <[email protected]> wrote:
From: Tien Hock Loh <[email protected]>
Add driver support for Altera GPIO soft IP, including interrupts and I/O. Tested on Altera CV SoC board using dipsw and LED using LED framework.
Signed-off-by Tien Hock Loh <[email protected]>
Can I get some ACKs on this? Like from Gerhard, DT folks?
Or is there still more things to do?
Yours,
Linus Walleij
Quoted 5 lines · Linus Walleij
Quoted 2 lines · On Mon, Mar 3, 2014 at 6:27 PM, <[email protected]> wrote:
From: Tien Hock Loh <[email protected]>
Quoted 8 lines
Quoted 4 lines
Add driver support for Altera GPIO soft IP, including interrupts and I/O. Tested on Altera CV SoC board using dipsw and LED using LED framework.
Signed-off-by Tien Hock Loh <[email protected]>Can I get some ACKs on this? Like from Gerhard, DT folks?
Or is there still more things to do?
Sorry, had flagged the patch for review, but did not get around to respond earlier. No ACK from me for v7 as it is.
Am I supposed to provide ACKs at all, given that I'm a reviewer and fellow developer. I always understood that Acked-By is something more formal, where one maintainer signals to another maintainer that it's OK to go though a different tree.
virtually yours
Gerhard Sittig
Signature · 3 lines
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr. 5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: [email protected]
Quoted 9 lines · Gerhard Sittig
Quoted 8 lines · Linus Walleij
Quoted 5 lines · On Mon, Mar 3, 2014 at 6:27 PM, <[email protected]> wrote:
From: Tien Hock Loh <[email protected]>
Add driver support for Altera GPIO soft IP, including interrupts and I/O. Tested on Altera CV SoC board using dipsw and LED using LED framework.
Quoted 8 lines
Quoted 5 lines
Quoted 1 line
Signed-off-by Tien Hock Loh <[email protected]>Can I get some ACKs on this? Like from Gerhard, DT folks?
Or is there still more things to do?
Sorry, had flagged the patch for review, but did not get around to respond earlier. No ACK from me for v7 as it is.
OK I'm holding this back.
Quoted 4 lines
Am I supposed to provide ACKs at all, given that I'm a reviewer and fellow developer. I always understood that Acked-By is something more formal, where one maintainer signals to another maintainer that it's OK to go though a different tree.
In this case I want an Acked-by/Reviewed-by tag (choose!) beacuse I trust your ability to review this specific code, as simple as that. :-)
Yours,
Linus Walleij
Quoted 36 lines · On Mon, Mar 03, 2014 at 06:27:43PM +0800, [email protected] wrote:
From: Tien Hock Loh <[email protected]>
Add driver support for Altera GPIO soft IP, including interrupts and I/O. Tested on Altera CV SoC board using dipsw and LED using LED framework.
Signed-off-by Tien Hock Loh <[email protected]>--- .../devicetree/bindings/gpio/gpio-altera.txt | 44 +++ drivers/gpio/Kconfig | 7 + drivers/gpio/Makefile | 1 + drivers/gpio/gpio-altera.c | 419 +++++++++++++++++++++ 4 files changed, 471 insertions(+) create mode 100644 Documentation/devicetree/bindings/gpio/gpio-altera.txt create mode 100644 drivers/gpio/gpio-altera.cDocumentation/devicetree/bindings/gpio/gpio-altera.txt new file +15
-0@@ -0,0 +1,44 @@ +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 1 + - The first cell is the gpio offset number +- gpio-controller : Marks the device node as a GPIO controller. +- #interrupt-cells : Should be 1. + - The first cell is the GPIO offset number within the GPIO controller. +- interrupts: Specify the interrupt. +- interrupt-controller: Mark the device node as an interrupt controller + +Altera GPIO specific required properties:
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 25
Quoted 10 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +10
-0+- altr,interrupt_trigger: Specifies the interrupt trigger type the GPIO + hardware is synthesized. This field is required if the Altera GPIO controller + used has IRQ enabled as the interrupt type is not software controlled, + but hardware synthesized. Required if GPIO is used as an interrupt + controller. The value is defined in <dt-bindings/interrupt-controller/irq.h> + Only the following flags are supported: + IRQ_TYPE_EDGE_RISING + IRQ_TYPE_EDGE_FALLING + IRQ_TYPE_EDGE_BOTH + IRQ_TYPE_LEVEL_HIGH
I'd suggest "altr,interrupt-trigger" (with a hyphen). So, if I'm understanding properly, this controller doesn't support per-gpio trigger settings?
Low-level triggered interrupts aren't supported?
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 30
Quoted 5 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +5
-0+ +Altera GPIO specific optional properties: +- altr,gpio-bank-width: Width of the GPIO bank. This defines how many pins the + GPIO device has. Ranges between 1-32. Optional and defaults to 32 is not + specified.
Generally, this is called 'ngpio', I think. Might be nice to stay consistent with other bindings.
Quoted 19 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +14
-0+ +Example: + +gpio_altr: gpio@40000 { + compatible = "altr,pio-1.0"; + reg = <0xff200000 0x10>; + interrupts = <0 45 4>; + altr,gpio-bank-width = <32>; + altr,interrupt_trigger = <IRQ_TYPE_EDGE_RISING>; + #gpio-cells = <1>; + gpio-controller; + #interrupt-cells = <1>; + interrupt-controller; +};drivers/gpio/Kconfig +0
-0
drivers/gpio/Kconfig config GPIO_GENERIC_PLATFORM · line 136
Quoted 9 lines
drivers/gpio/Kconfig +6
-0@@ -128,6 +128,13 @@ 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.
nit: you should use tabs for indentation instead of spaces to stay consistent.
Quoted 56 lines
drivers/gpio/Kconfig +1
-0+ 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_BCM_KONA) += gpio-bcm-kona.odrivers/gpio/gpio-altera.c new file +34
-0@@ -0,0 +1,419 @@ +/* +* Copyright (C) 2013 Altera Corporation +* Based on gpio-mpc8xxx.c +* +* This program is free software; you can redistribute it and/or modify +* it under the terms of the GNU General Public License as published by +* the Free Software Foundation; either version 2 of the License, or +* (at your option) any later version. +* +* This program is distributed in the hope that it will be useful, +* but WITHOUT ANY WARRANTY; without even the implied warranty of +* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +* GNU General Public License for more details. +* +* You should have received a copy of the GNU General Public License +* along with this program. If not, see <http://www.gnu.org/licenses/>. +*/ + +#include <linux/gpio.h> +#include <linux/init.h> +#include <linux/irqdomain.h> +#include <linux/interrupt.h> +#include <linux/io.h> +#include <linux/irq.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/of_irq.h> +#include <linux/of_gpio.h> +#include <linux/platform_device.h> +#include <linux/slab.h> + +#include <linux/irqchip/chained_irq.h> + +#define ALTERA_GPIO_DATA 0x0
drivers/gpio/gpio-altera.c line 44
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+#define ALTERA_GPIO_DIR 0x4 +#define ALTERA_GPIO_IRQ_MASK 0x8 +#define ALTERA_GPIO_EDGE_CAP 0xc +#define ALTERA_GPIO_OUTSET 0x10 +#define ALTERA_GPIO_OUTCLEAR 0x14 + +/** +* struct altera_gpio_chip +* @mmchip : memory mapped chip structure. +* @irq : irq domain that this driver is registered to.
s/@irq/@domain/ ?
drivers/gpio/gpio-altera.c line 48
Quoted 4 lines
drivers/gpio/gpio-altera.c new file +4
-0+* @gpio_lock : synchronization lock so that new irq/set/get requests + will be blocked until the current one completes. +* @interrupt_trigger : specifies the hardware configured IRQ trigger type + (rising ,falling ,both ,high )
@edge_type?
drivers/gpio/gpio-altera.c line 52
Quoted 4 lines
drivers/gpio/gpio-altera.c new file +4
-0+* @mapped_irq : kernel mapped irq number. +*/ +struct altera_gpio_chip { +struct of_mm_gpio_chip mmchip;
I had never really looked into of_mm_gpio_chip before but...does this really get you anything? All it does is manage mapping your registers for you; as far as I can tell it's just another layer of indirection with no gain.
I'd suggest lifting 'regs' and 'chip' directly into your altera_gpio_chip:
struct altera_gpio_chip {
struct gpio_chip chip;
struct irq_domain *domain;
void __iomem *regs;
spinlock_t gpio_lock;
int mapped_irq;
};[..]
drivers/gpio/gpio-altera.c line 126
Quoted 11 lines
drivers/gpio/gpio-altera.c new file +11
-0+static unsigned int altera_gpio_irq_startup (struct irq_data *d) +{ +altera_gpio_irq_unmask (d); + +return 0 ; +} + +static void altera_gpio_irq_shutdown (struct irq_data *d) +{ +altera_gpio_irq_unmask (d); +}
Should you really even be touching masks in startup/shutdown?
Quoted 9 lines
drivers/gpio/gpio-altera.c new file +9
-0+static struct irq_chip altera_irq_chip = { + .name ="altera-gpio" , + .irq_mask = altera_gpio_irq_mask, + .irq_unmask = altera_gpio_irq_unmask, + .irq_set_type = altera_gpio_irq_set_type, + .irq_startup = altera_gpio_irq_startup, + .irq_shutdown = altera_gpio_irq_shutdown, +}; +
[..]
drivers/gpio/gpio-altera.c line 214
Quoted 8 lines
drivers/gpio/gpio-altera.c new file +8
-0+static int altera_gpio_to_irq (struct gpio_chip *gc,unsigned offset) +{ +struct of_mm_gpio_chip *mm_gc =to_of_mm_gpio_chip (gc); +struct altera_gpio_chip *altera_gc =container_of (mm_gc, + struct altera_gpio_chip, mmchip); + +if (!altera_gc->domain) +return -ENXIO ;
How could this happen (hint, move your domain creation before gpiochip_add).
Quoted 8 lines
drivers/gpio/gpio-altera.c new file +8
-0+if (offset < altera_gc->mmchip.gc.ngpio) +return irq_find_mapping (altera_gc->domain, offset); +else +return -ENXIO ; +} + +static void altera_gpio_irq_handler (unsigned int irq,struct irq_desc *desc) +{
drivers/gpio/gpio-altera.c line 232
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+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 i; + +chained_irq_enter (chip, desc); +/* Handling for level trigger and edge trigger is different */ +if (altera_gc->interrupt_trigger ==IRQ_TYPE_LEVEL_HIGH ) {
Suggestion: split this into two handling functions, install one-or-the-other in your probe() based on the "altr,trigger-type" property. Bonus points: remove interrupt_trigger member from altera_gpio_chip entirely.
drivers/gpio/gpio-altera.c line 241
Quoted 9 lines
drivers/gpio/gpio-altera.c new file +9
-0+ status =readl_relaxed (mm_gc->regs +ALTERA_GPIO_DATA ); + status &=readl_relaxed (mm_gc->regs +ALTERA_GPIO_IRQ_MASK ); + +for (i =0 ; i < mm_gc->gc.ngpio; i++) { +if (status &BIT (i)) { +generic_handle_irq (irq_find_mapping ( + altera_gc->domain, i)); + } + }
You may find for_each_set_bit() handy.
Quoted 14 lines
drivers/gpio/gpio-altera.c new file +14
-0+ }else { +while ((status = + (readl_relaxed (mm_gc->regs +ALTERA_GPIO_EDGE_CAP ) & +readl_relaxed (mm_gc->regs +ALTERA_GPIO_IRQ_MASK )))) { +writel_relaxed (status, + mm_gc->regs +ALTERA_GPIO_EDGE_CAP ); +for (i =0 ; i < mm_gc->gc.ngpio; i++) { +if (status &BIT (i)) { +generic_handle_irq (irq_find_mapping ( + altera_gc->domain, i)); + } + } + } + }
drivers/gpio/gpio-altera.c line 265
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+ +chained_irq_exit (chip, desc); +} + +static int altera_gpio_irq_map (struct irq_domain *h,unsigned int irq, +irq_hw_number_t hw_irq_num) +{ +irq_set_chip_data (irq, h->host_data); +irq_set_chip_and_handler (irq, &altera_irq_chip, handle_level_irq); +irq_set_irq_type (irq,IRQ_TYPE_NONE );
Does irq_set_irq_type(irq, IRQ_TYPE_NONE) even do anything meaningful here?
drivers/gpio/gpio-altera.c line 270
Quoted 5 lines
drivers/gpio/gpio-altera.c new file +5
-0+ +return 0 ; +} + +static struct irq_domain_ops altera_gpio_irq_ops = {
const?
drivers/gpio/gpio-altera.c line 282
Quoted 12 lines
drivers/gpio/gpio-altera.c new file +12
-0+ .map = altera_gpio_irq_map, + .xlate = irq_domain_xlate_onecell, +}; + +int altera_gpio_probe (struct platform_device *pdev) +{ +struct device_node *node = pdev->dev.of_node; +int i, id, reg, ret; +struct altera_gpio_chip *altera_gc =devm_kzalloc (&pdev->dev, +sizeof (*altera_gc),GFP_KERNEL ); +if (altera_gc ==NULL ) { +pr_err ("%s: out of memory\n" , node->full_name);
This print is not necessary, as the allocator will complain loudly on your behalf. Also, I'd suggest you not define and initialize 'altera_gc' on the same line.
Quoted 27 lines
drivers/gpio/gpio-altera.c new file +26
-0+return -ENOMEM ; + } + altera_gc->domain =0 ; + +spin_lock_init (&altera_gc->gpio_lock); + + id = pdev->id; + +if (of_property_read_u32 (node,"altr,gpio-bank-width" , ®)) +/*By default assume full GPIO controller*/ + altera_gc->mmchip.gc.ngpio =32 ; +else + altera_gc->mmchip.gc.ngpio = reg; + +if (altera_gc->mmchip.gc.ngpio >32 ) { +dev_warn (&pdev->dev, +"ngpio is greater than 32, defaulting to 32\n" ); + altera_gc->mmchip.gc.ngpio =32 ; + } + + altera_gc->mmchip.gc.direction_input = altera_gpio_direction_input; + altera_gc->mmchip.gc.direction_output = altera_gpio_direction_output; + altera_gc->mmchip.gc.get = altera_gpio_get; + altera_gc->mmchip.gc.set = altera_gpio_set; + altera_gc->mmchip.gc.to_irq = altera_gpio_to_irq; + altera_gc->mmchip.gc.owner =THIS_MODULE ;
drivers/gpio/gpio-altera.c line 318
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+ + ret =of_mm_gpiochip_add (node, &altera_gc->mmchip); +if (ret) { +dev_err (&pdev->dev,"Failed adding memory mapped gpiochip\n" ); +return ret; + } + +platform_set_drvdata (pdev, altera_gc); + + altera_gc->mapped_irq =irq_of_parse_and_map (node,0 );
platform_get_irq(pdev, 0);
Quoted 44 lines
drivers/gpio/gpio-altera.c new file +44
-0+ +if (!altera_gc->mapped_irq) + goto skip_irq; + + altera_gc->domain =irq_domain_add_linear (node, + altera_gc->mmchip.gc.ngpio, &altera_gpio_irq_ops, altera_gc); + +if (!altera_gc->domain) { + ret = -ENODEV ; + goto dispose_irq; + } + +for (i =0 ; i < altera_gc->mmchip.gc.ngpio; i++) +irq_create_mapping (altera_gc->domain, i); + +if (of_property_read_u32 (node,"altr,interrupt_type" , ®)) { + ret = -EINVAL ; +dev_err (&pdev->dev, +"altr,interrupt_type value not set in device tree\n" ); + goto teardown; + } + altera_gc->interrupt_trigger = reg; + +irq_set_handler_data (altera_gc->mapped_irq, altera_gc); +irq_set_chained_handler (altera_gc->mapped_irq, altera_gpio_irq_handler); + +return 0 ; + +teardown: +irq_domain_remove (altera_gc->domain); +dispose_irq: +irq_dispose_mapping (altera_gc->mapped_irq); +WARN_ON (gpiochip_remove (&altera_gc->mmchip.gc) <0 ); + +pr_err ("%s: registration failed with status %d\n" , + node->full_name, ret); + +return ret; +skip_irq: +return 0 ; +} + +static int altera_gpio_remove (struct platform_device *pdev) +{
drivers/gpio/gpio-altera.c line 372
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+unsigned int irq, i; +int status; +struct altera_gpio_chip *altera_gc =platform_get_drvdata (pdev); + + status =gpiochip_remove (&altera_gc->mmchip.gc); + +if (status <0 ) +return status; + +if (altera_gc->domain) {
How could this happen?
drivers/gpio/gpio-altera.c
Quoted 1 line
drivers/gpio/gpio-altera.c new file +1
-0+irq_dispose_mapping (altera_gc->mapped_irq);
Does irq_domain_remove() not teardown mappings?
Quoted 6 lines
drivers/gpio/gpio-altera.c new file +6
-0+ +for (i =0 ; i < altera_gc->mmchip.gc.ngpio; i++) { + irq =irq_find_mapping (altera_gc->domain, i); +if (irq >0 ) +irq_dispose_mapping (irq); + }
drivers/gpio/gpio-altera.c line 389
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+ +irq_domain_remove (altera_gc->domain); + } + +irq_set_handler_data (altera_gc->mapped_irq,NULL ); +irq_set_chained_handler (altera_gc->mapped_irq,NULL ); +return -EIO ; +} + +static struct of_device_id altera_gpio_of_match[] = {
const?
Quoted 30 lines
drivers/gpio/gpio-altera.c new file +30
-0+ { .compatible = "altr,pio-1.0", }, + {}, +}; +MODULE_DEVICE_TABLE (of, altera_gpio_of_match); + +static struct platform_driver altera_gpio_driver = { + .driver = { + .name ="altera_gpio" , + .owner =THIS_MODULE , + .of_match_table =of_match_ptr (altera_gpio_of_match), + }, + .probe = altera_gpio_probe, + .remove = altera_gpio_remove, +}; + +static int __init altera_gpio_init (void ) +{ +return platform_driver_register (&altera_gpio_driver); +} +subsys_initcall (altera_gpio_init); + +static void __exit altera_gpio_exit(void) +{ +platform_driver_unregister (&altera_gpio_driver); +} +module_exit (altera_gpio_exit); + +MODULE_AUTHOR ("Tien Hock Loh <[email protected]>" ); +MODULE_DESCRIPTION ("Altera GPIO driver" ); +MODULE_LICENSE ("GPL" );
Signature · 2 lines
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation
Quoted 43 lines · Josh Cartwright
Quoted 42 lines · On Mon, Mar 03, 2014 at 06:27:43PM +0800, [email protected] wrote:
From: Tien Hock Loh <[email protected]>
Add driver support for Altera GPIO soft IP, including interrupts and I/O. Tested on Altera CV SoC board using dipsw and LED using LED framework.
Signed-off-by Tien Hock Loh <[email protected]>--- .../devicetree/bindings/gpio/gpio-altera.txt | 44 +++ drivers/gpio/Kconfig | 7 + drivers/gpio/Makefile | 1 + drivers/gpio/gpio-altera.c | 419 +++++++++++++++++++++ 4 files changed, 471 insertions(+) create mode 100644 Documentation/devicetree/bindings/gpio/gpio-altera.txt create mode 100644 drivers/gpio/gpio-altera.cDocumentation/devicetree/bindings/gpio/gpio-altera.txt new file +21
-0@@ -0,0 +1,44 @@ +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 1 + - The first cell is the gpio offset number +- gpio-controller : Marks the device node as a GPIO controller. +- #interrupt-cells : Should be 1. + - The first cell is the GPIO offset number within the GPIO controller. +- interrupts: Specify the interrupt. +- interrupt-controller: Mark the device node as an interrupt controller + +Altera GPIO specific required properties: +- altr,interrupt_trigger: Specifies the interrupt trigger type the GPIO + hardware is synthesized. This field is required if the Altera GPIO controller + used has IRQ enabled as the interrupt type is not software controlled, + but hardware synthesized. Required if GPIO is used as an interrupt + controller. The value is defined in <dt-bindings/interrupt-controller/irq.h> + Only the following flags are supported:
Quoted 8 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 25
Quoted 4 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +4
-0+ IRQ_TYPE_EDGE_RISING + IRQ_TYPE_EDGE_FALLING + IRQ_TYPE_EDGE_BOTH + IRQ_TYPE_LEVEL_HIGHI'd suggest "altr,interrupt-trigger" (with a hyphen). So, if I'm understanding properly, this controller doesn't support per-gpio trigger settings?
Okay, I'll change it to interrupt-trigger. Yes, the hardware doesn't support per-gpio trigger.
Quoted 2 lines
Low-level triggered interrupts aren't supported?
Yes, the hardware controller doesn't support it.
Quoted 9 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 30
Quoted 6 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +5
-0+ +Altera GPIO specific optional properties: +- altr,gpio-bank-width: Width of the GPIO bank. This defines how many pins the + GPIO device has. Ranges between 1-32. Optional and defaults to 32 is not + specified.Generally, this is called 'ngpio', I think. Might be nice to stay consistent with other bindings.
Noted. Thanks.
Quoted 23 lines
Quoted 23 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +14
-0+ +Example: + +gpio_altr: gpio@40000 { + compatible = "altr,pio-1.0"; + reg = <0xff200000 0x10>; + interrupts = <0 45 4>; + altr,gpio-bank-width = <32>; + altr,interrupt_trigger = <IRQ_TYPE_EDGE_RISING>; + #gpio-cells = <1>; + gpio-controller; + #interrupt-cells = <1>; + interrupt-controller; +};drivers/gpio/Kconfig +0
-0@@ -128,6 +128,13 @@ config GPIO_GENERIC_PLATFORM help Say yes here to support basic platform_device memory-mapped GPIO controllers.
Quoted 9 lines
drivers/gpio/Kconfig config GPIO_GENERIC_PLATFORM · line 136
Quoted 6 lines
drivers/gpio/Kconfig +6
-0+config GPIO_ALTERA + tristate "Altera GPIO" + select IRQ_DOMAIN + depends on OF_GPIO + help + Say yes here to support the Altera PIO device.nit: you should use tabs for indentation instead of spaces to stay consistent.
OK.
Quoted 61 lines
Quoted 61 lines
drivers/gpio/Kconfig +1
-0+ 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_BCM_KONA) += gpio-bcm-kona.odrivers/gpio/gpio-altera.c new file +38
-0@@ -0,0 +1,419 @@ +/* +* Copyright (C) 2013 Altera Corporation +* Based on gpio-mpc8xxx.c +* +* This program is free software; you can redistribute it and/or modify +* it under the terms of the GNU General Public License as published by +* the Free Software Foundation; either version 2 of the License, or +* (at your option) any later version. +* +* This program is distributed in the hope that it will be useful, +* but WITHOUT ANY WARRANTY; without even the implied warranty of +* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +* GNU General Public License for more details. +* +* You should have received a copy of the GNU General Public License +* along with this program. If not, see <http://www.gnu.org/licenses/>. +*/ + +#include <linux/gpio.h> +#include <linux/init.h> +#include <linux/irqdomain.h> +#include <linux/interrupt.h> +#include <linux/io.h> +#include <linux/irq.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/of_irq.h> +#include <linux/of_gpio.h> +#include <linux/platform_device.h> +#include <linux/slab.h> + +#include <linux/irqchip/chained_irq.h> + +#define ALTERA_GPIO_DATA 0x0 +#define ALTERA_GPIO_DIR 0x4 +#define ALTERA_GPIO_IRQ_MASK 0x8 +#define ALTERA_GPIO_EDGE_CAP 0xc +#define ALTERA_GPIO_OUTSET 0x10
Quoted 9 lines
drivers/gpio/gpio-altera.c line 44
Quoted 6 lines
drivers/gpio/gpio-altera.c new file +6
-0+#define ALTERA_GPIO_OUTCLEAR 0x14 + +/** +* struct altera_gpio_chip +* @mmchip : memory mapped chip structure. +* @irq : irq domain that this driver is registered to. s/@irq/@domain/ ?
Yes missed this sorry.
Quoted 7 lines
drivers/gpio/gpio-altera.c line 48
Quoted 4 lines
drivers/gpio/gpio-altera.c new file +4
-0+* @gpio_lock : synchronization lock so that new irq/set/get requests + will be blocked until the current one completes. +* @interrupt_trigger : specifies the hardware configured IRQ trigger type + (rising ,falling ,both ,high )@edge_type?
Oh, a stray. I'll remove this.
Quoted 21 lines
drivers/gpio/gpio-altera.c line 52
Quoted 4 lines
drivers/gpio/gpio-altera.c new file +4
-0+* @mapped_irq : kernel mapped irq number. +*/ +struct altera_gpio_chip { +struct of_mm_gpio_chip mmchip;I had never really looked into
of_mm_gpio_chipbefore but...does this really get you anything? All it does is manage mapping your registers for you; as far as I can tell it's just another layer of indirection with no gain.I'd suggest lifting 'regs' and 'chip' directly into your
altera_gpio_chip:struct altera_gpio_chip { struct gpio_chip chip; struct irq_domain *domain; void __iomem *regs; spinlock_t gpio_lock; int mapped_irq; };
Quoted 2 lines
[..]
The variable is used in the helper function of_mm_gpiochip_add() for general memory mapped gpio to reduce code duplications. A lot of the memory mapped GPIO driver uses this, I just refers to them. Please let me know if you strongly feel otherwise.
Quoted 5 lines
Quoted 5 lines
drivers/gpio/gpio-altera.c new file +5
-0+static unsigned int altera_gpio_irq_startup (struct irq_data *d) +{ +altera_gpio_irq_unmask (d); + +return 0 ;
Quoted 9 lines
drivers/gpio/gpio-altera.c line 126
Quoted 6 lines
drivers/gpio/gpio-altera.c new file +6
-0+} + +static void altera_gpio_irq_shutdown (struct irq_data *d) +{ +altera_gpio_irq_unmask (d); +}Should you really even be touching masks in startup/shutdown?
I'm referring from other GPIO drivers and that's what they do. I do make a mistake by unmasking the PIOs during shutdown. I'll fix this.
Quoted 10 lines
Quoted 9 lines
drivers/gpio/gpio-altera.c new file +9
-0+static struct irq_chip altera_irq_chip = { + .name ="altera-gpio" , + .irq_mask = altera_gpio_irq_mask, + .irq_unmask = altera_gpio_irq_unmask, + .irq_set_type = altera_gpio_irq_set_type, + .irq_startup = altera_gpio_irq_startup, + .irq_shutdown = altera_gpio_irq_shutdown, +}; +[..]
Quoted 11 lines
drivers/gpio/gpio-altera.c line 214
Quoted 8 lines
drivers/gpio/gpio-altera.c new file +8
-0+static int altera_gpio_to_irq (struct gpio_chip *gc,unsigned offset) +{ +struct of_mm_gpio_chip *mm_gc =to_of_mm_gpio_chip (gc); +struct altera_gpio_chip *altera_gc =container_of (mm_gc, + struct altera_gpio_chip, mmchip); + +if (!altera_gc->domain) +return -ENXIO ;How could this happen (hint, move your domain creation before
gpiochip_add).
Noted. Thanks.
Quoted 16 lines
Quoted 16 lines
drivers/gpio/gpio-altera.c new file +15
-0+if (offset < altera_gc->mmchip.gc.ngpio) +return irq_find_mapping (altera_gc->domain, offset); +else +return -ENXIO ; +} + +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 i; +
Quoted 9 lines
drivers/gpio/gpio-altera.c line 232
Quoted 3 lines
drivers/gpio/gpio-altera.c new file +3
-0+chained_irq_enter (chip, desc); +/* Handling for level trigger and edge trigger is different */ +if (altera_gc->interrupt_trigger ==IRQ_TYPE_LEVEL_HIGH ) {Suggestion: split this into two handling functions, install one-or-the-other in your
probe()based on the "altr,trigger-type" property. Bonus points: removeinterrupt_triggermember fromaltera_gpio_chipentirely.
Good point. I'll implement this. However I don't think I can remove interrupt_trigger from altera_gpio_chip since when doing altera_gpio_irq_set_type we still need the variable to determine if we can set the type requested by the user.
Quoted 10 lines
drivers/gpio/gpio-altera.c line 241
Quoted 9 lines
drivers/gpio/gpio-altera.c new file +9
-0+ status =readl_relaxed (mm_gc->regs +ALTERA_GPIO_DATA ); + status &=readl_relaxed (mm_gc->regs +ALTERA_GPIO_IRQ_MASK ); + +for (i =0 ; i < mm_gc->gc.ngpio; i++) { +if (status &BIT (i)) { +generic_handle_irq (irq_find_mapping ( + altera_gc->domain, i)); + } + }
Quoted 2 lines
You may find
for_each_set_bit()handy.
Good point. Will implement. Thanks.
Quoted 18 lines
Quoted 18 lines
drivers/gpio/gpio-altera.c new file +18
-0+ }else { +while ((status = + (readl_relaxed (mm_gc->regs +ALTERA_GPIO_EDGE_CAP ) & +readl_relaxed (mm_gc->regs +ALTERA_GPIO_IRQ_MASK )))) { +writel_relaxed (status, + mm_gc->regs +ALTERA_GPIO_EDGE_CAP ); +for (i =0 ; i < mm_gc->gc.ngpio; i++) { +if (status &BIT (i)) { +generic_handle_irq (irq_find_mapping ( + altera_gc->domain, i)); + } + } + } + } + +chained_irq_exit (chip, desc); +} +
Quoted 9 lines
drivers/gpio/gpio-altera.c line 265
Quoted 6 lines
drivers/gpio/gpio-altera.c new file +6
-0+static int altera_gpio_irq_map (struct irq_domain *h,unsigned int irq, +irq_hw_number_t hw_irq_num) +{ +irq_set_chip_data (irq, h->host_data); +irq_set_chip_and_handler (irq, &altera_irq_chip, handle_level_irq); +irq_set_irq_type (irq,IRQ_TYPE_NONE );Does
irq_set_irq_type(irq, IRQ_TYPE_NONE)even do anything meaningful here?
I'll remove this.
Quoted 8 lines
drivers/gpio/gpio-altera.c line 270
Quoted 5 lines
drivers/gpio/gpio-altera.c new file +5
-0+ +return 0 ; +} + +static struct irq_domain_ops altera_gpio_irq_ops = {const?
I'll add this.
Quoted 8 lines
Quoted 8 lines
drivers/gpio/gpio-altera.c new file +8
-0+ .map = altera_gpio_irq_map, + .xlate = irq_domain_xlate_onecell, +}; + +int altera_gpio_probe (struct platform_device *pdev) +{ +struct device_node *node = pdev->dev.of_node; +int i, id, reg, ret;
Quoted 8 lines
drivers/gpio/gpio-altera.c line 282
Quoted 4 lines
drivers/gpio/gpio-altera.c new file +4
-0+struct altera_gpio_chip *altera_gc =devm_kzalloc (&pdev->dev, +sizeof (*altera_gc),GFP_KERNEL ); +if (altera_gc ==NULL ) { +pr_err ("%s: out of memory\n" , node->full_name);This print is not necessary, as the allocator will complain loudly on your behalf. Also, I'd suggest you not define and initialize '
altera_gc' on the same line.
Noted.
Quoted 32 lines
Quoted 32 lines
drivers/gpio/gpio-altera.c new file +31
-0+return -ENOMEM ; + } + altera_gc->domain =0 ; + +spin_lock_init (&altera_gc->gpio_lock); + + id = pdev->id; + +if (of_property_read_u32 (node,"altr,gpio-bank-width" , ®)) +/*By default assume full GPIO controller*/ + altera_gc->mmchip.gc.ngpio =32 ; +else + altera_gc->mmchip.gc.ngpio = reg; + +if (altera_gc->mmchip.gc.ngpio >32 ) { +dev_warn (&pdev->dev, +"ngpio is greater than 32, defaulting to 32\n" ); + altera_gc->mmchip.gc.ngpio =32 ; + } + + altera_gc->mmchip.gc.direction_input = altera_gpio_direction_input; + altera_gc->mmchip.gc.direction_output = altera_gpio_direction_output; + altera_gc->mmchip.gc.get = altera_gpio_get; + altera_gc->mmchip.gc.set = altera_gpio_set; + altera_gc->mmchip.gc.to_irq = altera_gpio_to_irq; + altera_gc->mmchip.gc.owner =THIS_MODULE ; + + ret =of_mm_gpiochip_add (node, &altera_gc->mmchip); +if (ret) { +dev_err (&pdev->dev,"Failed adding memory mapped gpiochip\n" ); +return ret;
Quoted 9 lines
drivers/gpio/gpio-altera.c line 318
Quoted 6 lines
drivers/gpio/gpio-altera.c new file +5
-0+ } + +platform_set_drvdata (pdev, altera_gc); + + altera_gc->mapped_irq =irq_of_parse_and_map (node,0 );
platform_get_irq(pdev, 0);
OK.
Quoted 48 lines
Quoted 48 lines
drivers/gpio/gpio-altera.c new file +48
-0+ +if (!altera_gc->mapped_irq) + goto skip_irq; + + altera_gc->domain =irq_domain_add_linear (node, + altera_gc->mmchip.gc.ngpio, &altera_gpio_irq_ops, altera_gc); + +if (!altera_gc->domain) { + ret = -ENODEV ; + goto dispose_irq; + } + +for (i =0 ; i < altera_gc->mmchip.gc.ngpio; i++) +irq_create_mapping (altera_gc->domain, i); + +if (of_property_read_u32 (node,"altr,interrupt_type" , ®)) { + ret = -EINVAL ; +dev_err (&pdev->dev, +"altr,interrupt_type value not set in device tree\n" ); + goto teardown; + } + altera_gc->interrupt_trigger = reg; + +irq_set_handler_data (altera_gc->mapped_irq, altera_gc); +irq_set_chained_handler (altera_gc->mapped_irq, altera_gpio_irq_handler); + +return 0 ; + +teardown: +irq_domain_remove (altera_gc->domain); +dispose_irq: +irq_dispose_mapping (altera_gc->mapped_irq); +WARN_ON (gpiochip_remove (&altera_gc->mmchip.gc) <0 ); + +pr_err ("%s: registration failed with status %d\n" , + node->full_name, ret); + +return ret; +skip_irq: +return 0 ; +} + +static int altera_gpio_remove (struct platform_device *pdev) +{ +unsigned int irq, i; +int status; +struct altera_gpio_chip *altera_gc =platform_get_drvdata (pdev); +
Quoted 9 lines
drivers/gpio/gpio-altera.c line 372
Quoted 6 lines
drivers/gpio/gpio-altera.c new file +6
-0+ status =gpiochip_remove (&altera_gc->mmchip.gc); + +if (status <0 ) +return status; + +if (altera_gc->domain) {How could this happen?
Must be me getting overlay paranoid. I'll remove this.
Quoted 3 lines
Quoted 1 line
+
irq_dispose_mapping(altera_gc->mapped_irq);Does
irq_domain_remove()not teardown mappings?
The irq_domain_remove states that caller has to ensure all mapping has been disposed prior to calling the function.
Quoted 10 lines
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+ +for (i =0 ; i < altera_gc->mmchip.gc.ngpio; i++) { + irq =irq_find_mapping (altera_gc->domain, i); +if (irq >0 ) +irq_dispose_mapping (irq); + } + +irq_domain_remove (altera_gc->domain); + } +
Quoted 9 lines
drivers/gpio/gpio-altera.c line 389
Quoted 6 lines
drivers/gpio/gpio-altera.c new file +6
-0+irq_set_handler_data (altera_gc->mapped_irq,NULL ); +irq_set_chained_handler (altera_gc->mapped_irq,NULL ); +return -EIO ; +} + +static struct of_device_id altera_gpio_of_match[] = {const?
OK.
Quoted 34 lines
Quoted 20 lines
drivers/gpio/gpio-altera.c new file +20
-0+ { .compatible = "altr,pio-1.0", }, + {}, +}; +MODULE_DEVICE_TABLE (of, altera_gpio_of_match); + +static struct platform_driver altera_gpio_driver = { + .driver = { + .name ="altera_gpio" , + .owner =THIS_MODULE , + .of_match_table =of_match_ptr (altera_gpio_of_match), + }, + .probe = altera_gpio_probe, + .remove = altera_gpio_remove, +}; + +static int __init altera_gpio_init (void ) +{ +return platform_driver_register (&altera_gpio_driver); +} +subsys_initcall (altera_gpio_init);drivers/gpio/gpio-altera.c line 419
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+ +static void __exit altera_gpio_exit(void) +{ +platform_driver_unregister (&altera_gpio_driver); +} +module_exit (altera_gpio_exit); + +MODULE_AUTHOR ("Tien Hock Loh <[email protected]>" ); +MODULE_DESCRIPTION ("Altera GPIO driver" ); +MODULE_LICENSE ("GPL" );--
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation
Quoted 50 lines · Tien Hock Loh
Quoted 40 lines · Josh Cartwright
Quoted 39 lines · On Mon, Mar 03, 2014 at 06:27:43PM +0800, [email protected] wrote:
From: Tien Hock Loh <[email protected]>
Add driver support for Altera GPIO soft IP, including interrupts and I/O. Tested on Altera CV SoC board using dipsw and LED using LED framework.
Signed-off-by Tien Hock Loh <[email protected]>---
drivers/gpio/gpio-altera.c new file +31
-0+return -ENOMEM ; + } + altera_gc->domain =0 ; + +spin_lock_init (&altera_gc->gpio_lock); + + id = pdev->id; + +if (of_property_read_u32 (node,"altr,gpio-bank-width" , ®)) +/*By default assume full GPIO controller*/ + altera_gc->mmchip.gc.ngpio =32 ; +else + altera_gc->mmchip.gc.ngpio = reg; + +if (altera_gc->mmchip.gc.ngpio >32 ) { +dev_warn (&pdev->dev, +"ngpio is greater than 32, defaulting to 32\n" ); + altera_gc->mmchip.gc.ngpio =32 ; + } + + altera_gc->mmchip.gc.direction_input = altera_gpio_direction_input; + altera_gc->mmchip.gc.direction_output = altera_gpio_direction_output; + altera_gc->mmchip.gc.get = altera_gpio_get; + altera_gc->mmchip.gc.set = altera_gpio_set; + altera_gc->mmchip.gc.to_irq = altera_gpio_to_irq; + altera_gc->mmchip.gc.owner =THIS_MODULE ; + + ret =of_mm_gpiochip_add (node, &altera_gc->mmchip); +if (ret) { +dev_err (&pdev->dev,"Failed adding memory mapped gpiochip\n" ); +return ret;Quoted 9 lines
drivers/gpio/gpio-altera.c line 318
Quoted 6 lines
drivers/gpio/gpio-altera.c new file +5
-0+ } + +platform_set_drvdata (pdev, altera_gc); + + altera_gc->mapped_irq =irq_of_parse_and_map (node,0 );
platform_get_irq(pdev, 0);
Quoted 2 lines
OK.
platform_get_irq doesn't create the irq mapping which is needed by the driver. Since this driver is targeted at using of, should I be using irq_of_parse_and_map or should I still redo the codes with platform_get_irq and irq_create_mapping? I think the latter would be introducing code redundancy. Please advice.
Thanks.
Quoted 4 lines · Tien Hock Loh
Quoted 3 lines · Tien Hock Loh
Quoted 2 lines · Josh Cartwright
Quoted 1 line · On Mon, Mar 03, 2014 at 06:27:43PM +0800, [email protected] wrote:
From: Tien Hock Loh <[email protected]>
[..]
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 2 lines
+
altera_gc->mapped_irq=irq_of_parse_and_map(node, 0);
platform_get_irq(pdev, 0);
Quoted 8 lines
Quoted 2 lines
OK.
platform_get_irqdoesn't create the irq mapping which is needed by the driver. Since this driver is targeted at using of, should I be usingirq_of_parse_and_mapor should I still redo the codes withplatform_get_irqandirq_create_mapping? I think the latter would be introducing code redundancy. Please advice.
Yes, it is technically true that platform_get_irq() doesn't do the mapping directly, but that's because the mapping is setup earlier, when of_device_alloc() (drivers/of/platform.c) allocates resources for your platform device.
Calling irq_of_parse_and_map() should be unnecessary.
Signature · 2 lines
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by The Linux Foundation
Quoted 20 lines · Josh Cartwright
Quoted 4 lines · Tien Hock Loh
Quoted 3 lines · Tien Hock Loh
Quoted 2 lines · Josh Cartwright
Quoted 1 line · On Mon, Mar 03, 2014 at 06:27:43PM +0800, [email protected] wrote:
From: Tien Hock Loh <[email protected]>
[..]
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 2 lines
+
altera_gc->mapped_irq=irq_of_parse_and_map(node, 0);
platform_get_irq(pdev, 0);Quoted 8 lines
Quoted 2 lines
OK.
platform_get_irqdoesn't create the irq mapping which is needed by the driver. Since this driver is targeted at using of, should I be usingirq_of_parse_and_mapor should I still redo the codes withplatform_get_irqandirq_create_mapping? I think the latter would be introducing code redundancy. Please advice.
Quoted 6 lines
Yes, it is technically true that
platform_get_irq()doesn't do the mapping directly, but that's because the mapping is setup earlier, whenof_device_alloc()(drivers/of/platform.c) allocates resources for your platform device.Calling
irq_of_parse_and_map()should be unnecessary.
I checked and tried running the without irq_create_mapping but it seems the mapping is not done. What I've seen other GPIO driver is doing is to create the mapping during the gpio_to_irq call. However Linus suggested we are avoiding that route, thus the use of irq_of_parse_and_map.
Do you agree with my findings?
Quoted 1 line
________________________________
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 20 lines · Josh Cartwright
Quoted 4 lines · Tien Hock Loh
Quoted 3 lines · Tien Hock Loh
Quoted 2 lines · Josh Cartwright
Quoted 1 line · On Mon, Mar 03, 2014 at 06:27:43PM +0800, [email protected] wrote:
From: Tien Hock Loh <[email protected]>
[..]
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines
Quoted 2 lines
+
altera_gc->mapped_irq=irq_of_parse_and_map(node, 0);
platform_get_irq(pdev, 0);Quoted 8 lines
Quoted 2 lines
OK.
platform_get_irqdoesn't create the irq mapping which is needed by the driver. Since this driver is targeted at using of, should I be usingirq_of_parse_and_mapor should I still redo the codes withplatform_get_irqandirq_create_mapping? I think the latter would be introducing code redundancy. Please advice.
Quoted 6 lines
Yes, it is technically true that
platform_get_irq()doesn't do the mapping directly, but that's because the mapping is setup earlier, whenof_device_alloc()(drivers/of/platform.c) allocates resources for your platform device.Calling
irq_of_parse_and_map()should be unnecessary.
In the latest mainline, platform_get_irq() now does the mapping when using an IRQ specified in the device tree.
g.
Quoted 55 lines
Quoted 46 lines
Quoted 46 lines
Quoted 46 lines
Quoted 33 lines
Add driver support for Altera GPIO soft IP, including interrupts and I/O. Tested on Altera CV SoC board using dipsw and LED using LED framework.
Signed-off-by Tien Hock Loh <[email protected]>---
drivers/gpio/gpio-altera.c new file +26
-0+return -ENOMEM ; + } + altera_gc->domain =0 ; + +spin_lock_init (&altera_gc->gpio_lock); + + id = pdev->id; + +if (of_property_read_u32 (node,"altr,gpio-bank-width" , ®)) +/*By default assume full GPIO controller*/ + altera_gc->mmchip.gc.ngpio =32 ; +else + altera_gc->mmchip.gc.ngpio = reg; + +if (altera_gc->mmchip.gc.ngpio >32 ) { +dev_warn (&pdev->dev, +"ngpio is greater than 32, defaulting to 32\n" ); + altera_gc->mmchip.gc.ngpio =32 ; + } + + altera_gc->mmchip.gc.direction_input = altera_gpio_direction_input; + altera_gc->mmchip.gc.direction_output = altera_gpio_direction_output; + altera_gc->mmchip.gc.get = altera_gpio_get; + altera_gc->mmchip.gc.set = altera_gpio_set; + altera_gc->mmchip.gc.to_irq = altera_gpio_to_irq; + altera_gc->mmchip.gc.owner =THIS_MODULE ;drivers/gpio/gpio-altera.c line 318
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+ + ret =of_mm_gpiochip_add (node, &altera_gc->mmchip); +if (ret) { +dev_err (&pdev->dev,"Failed adding memory mapped gpiochip\n" ); +return ret; + } + +platform_set_drvdata (pdev, altera_gc); + + altera_gc->mapped_irq =irq_of_parse_and_map (node,0 );
platform_get_irq(pdev, 0);Quoted 8 lines
Quoted 2 lines
OK.
platform_get_irqdoesn't create the irq mapping which is needed by the driver. Since this driver is targeted at using of, should I be usingirq_of_parse_and_mapor should I still redo the codes withplatform_get_irqandirq_create_mapping? I think the latter would be introducing code redundancy. Please advice.
Quoted 5 lines
Yes, it is technically true that
platform_get_irq()doesn't do the mapping directly, but that's because the mapping is setup earlier, whenof_device_alloc()(drivers/of/platform.c) allocates resources for your platform device.Calling
irq_of_parse_and_map()should be unnecessary.
I checked and tried running the without irq_create_mapping but it seems the mapping is not done.
What I've seen other GPIO driver is doing is to create the mapping during the gpio_to_irq call. However Linus suggested we are avoiding that route, thus the use of irq_of_parse_and_map. Do you agree with my findings?
Quoted 2 lines
Quoted 2 lines
Thanks.
Quoted 7 lines · On Mon, Mar 03, 2014 at 18:27 +0800, [email protected] wrote:
From: Tien Hock Loh <[email protected]>
Add driver support for Altera GPIO soft IP, including interrupts and I/O. Tested on Altera CV SoC board using dipsw and LED using LED framework.
Signed-off-by Tien Hock Loh <[email protected]>
Quoted 8 lines
--- .../devicetree/bindings/gpio/gpio-altera.txt | 44 +++ drivers/gpio/Kconfig | 7 + drivers/gpio/Makefile | 1 + drivers/gpio/gpio-altera.c | 419 +++++++++++++++++++++ 4 files changed, 471 insertions(+) create mode 100644 Documentation/devicetree/bindings/gpio/gpio-altera.txt create mode 100644 drivers/gpio/gpio-altera.c
Let's see. A v7, i.e. quite some iterations done, and still whitespace issues, and not a single line of changelogs. Not exactly an invitation for reviewers to spend their time on it. It's hard to tell which feedback was received before, and what of it has been addressed.
Aren't binding patches to be separate (and first) in series these days, while driver adjustment or introduction then follows to implement what was specified?
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 8
Quoted 11 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +8
-0@@ -0,0 +1,44 @@ +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 1 + - The first cell is the gpio offset number
Will there never be any use of flags, that usually are kept in the second cell? Can we tell for sure that one cell is enough?
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 13
Quoted 5 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +5
-0+- gpio-controller : Marks the device node as a GPIO controller. +- #interrupt-cells : Should be 1. + - The first cell is the GPIO offset number within the GPIO controller. +- interrupts: Specify the interrupt. +- interrupt-controller: Mark the device node as an interrupt controller
Is #interrupt-cells = <1> enough? Are triggers fixed in the hardware? A comment about it may prevent people asking questions. (The motivation might be mentioned below, see the comment on the two "required properties" sections.)
I'd sort the last three items differently. 'interrupts' is where the GPIO block is "a client" to the IRQ controller which it is connected to. '#interrupt-cells' is the "server side" because the GPIO block is an IRQ controller and has the 'interrupt-controller' property.
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 15
Quoted 2 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +2
-0+ +Altera GPIO specific required properties:
This made me stop and wonder. This is the "Altera GPIO controller bindings" document, and it has a "Required properties" section as well as an "Altera GPIO specific required properties" section? Isn't this highly redundant, and confusing at the same time?
Quoted 19 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +19
-0+- altr,interrupt_trigger: Specifies the interrupt trigger type the GPIO + hardware is synthesized. This field is required if the Altera GPIO controller + used has IRQ enabled as the interrupt type is not software controlled, + but hardware synthesized. Required if GPIO is used as an interrupt + controller. The value is defined in <dt-bindings/interrupt-controller/irq.h> + Only the following flags are supported: + IRQ_TYPE_EDGE_RISING + IRQ_TYPE_EDGE_FALLING + IRQ_TYPE_EDGE_BOTH + IRQ_TYPE_LEVEL_HIGH + +Altera GPIO specific optional properties: +- altr,gpio-bank-width: Width of the GPIO bank. This defines how many pins the + GPIO device has. Ranges between 1-32. Optional and defaults to 32 is not + specified. + +Example: + +gpio_altr: gpio@40000 {
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 44
Quoted 10 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +10
-0+ compatible = "altr,pio-1.0"; + reg = <0xff200000 0x10>; + interrupts = <0 45 4>; + altr,gpio-bank-width = <32>; + altr,interrupt_trigger = <IRQ_TYPE_EDGE_RISING>; + #gpio-cells = <1>; + gpio-controller; + #interrupt-cells = <1>; + interrupt-controller; +};
The node's address does not match the 'reg' property. The interrupt trigger uses symbolic names, so could the 'interrupts' spec. Examples should not assume an external 'interrupt-parent' but should list them for completeness.
drivers/gpio/Kconfig config GPIO_GENERIC_PLATFORM · line 137
Quoted 13 lines
drivers/gpio/Kconfig +7
-0@@ -128,6 +128,13 @@ 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. +
I guess checkpatch nags about the rather short help text. Tristate options probably should mention the opportunity to build a module, and by convention should state what the module's name then would be (which in total gets you the minimum number of help text lines as a byproduct).
Quoted 9 lines
drivers/gpio/gpio-altera.c new file +5
-0@@ -0,0 +1,419 @@ + +#include <linux/gpio.h> +#include <linux/init.h> +#include <linux/irqdomain.h> +#include <linux/interrupt.h>
drivers/gpio/gpio-altera.c line 32
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+#include <linux/io.h> +#include <linux/irq.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/of_irq.h> +#include <linux/of_gpio.h> +#include <linux/platform_device.h> +#include <linux/slab.h> + +#include <linux/irqchip/chained_irq.h>
Includes should be alpha-sorted. To detect duplicates, and to avoid or reduce conflicts.
drivers/gpio/gpio-altera.c line 39
Quoted 7 lines
drivers/gpio/gpio-altera.c new file +7
-0+ +#define ALTERA_GPIO_DATA 0x0 +#define ALTERA_GPIO_DIR 0x4 +#define ALTERA_GPIO_IRQ_MASK 0x8 +#define ALTERA_GPIO_EDGE_CAP 0xc +#define ALTERA_GPIO_OUTSET 0x10 +#define ALTERA_GPIO_OUTCLEAR 0x14
OUTSET and OUTCLEAR are not used in the driver source. You might as well drop their declarations (after all you are not declaring the register set layout here, but are just introducing magic offset numbers for some of the registers that the driver may access).
drivers/gpio/gpio-altera.c line 65
Quoted 6 lines
drivers/gpio/gpio-altera.c new file +6
-0+static void altera_gpio_irq_unmask (struct irq_data *d) +{ +struct altera_gpio_chip *altera_gc =irq_data_get_irq_chip_data (d); +struct of_mm_gpio_chip *mm_gc = &altera_gc->mmchip; +unsigned long flags; +unsigned int intmask;
I'd suggest <stdint.h> style fixed width data types (uint32_t) or their kernel equivalents (u32) for operations on fixed width peripheral registers.
Quoted 24 lines
drivers/gpio/gpio-altera.c new file +23
-0+ +spin_lock_irqsave (&altera_gc->gpio_lock, flags); + intmask =readl (mm_gc->regs +ALTERA_GPIO_IRQ_MASK ); +/* Set ALTERA_GPIO_IRQ_MASK bit to unmask */ + intmask |=BIT (irqd_to_hwirq (d)); +writel (intmask, mm_gc->regs +ALTERA_GPIO_IRQ_MASK ); +spin_unlock_irqrestore (&altera_gc->gpio_lock, flags); +} +static int altera_gpio_irq_set_type (struct irq_data *d, +unsigned int type) +{ +struct altera_gpio_chip *altera_gc =irq_data_get_irq_chip_data (d); + +if (type ==IRQ_TYPE_NONE ) +return 0 ; + +if (type ==IRQ_TYPE_LEVEL_HIGH && + altera_gc->interrupt_trigger ==IRQ_TYPE_LEVEL_HIGH ) { +return 0 ; + }else { +if (type ==IRQ_TYPE_EDGE_RISING && + altera_gc->interrupt_trigger ==IRQ_TYPE_EDGE_RISING ) +return 0 ;
drivers/gpio/gpio-altera.c line 114
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+else if (type ==IRQ_TYPE_EDGE_FALLING && + altera_gc->interrupt_trigger ==IRQ_TYPE_EDGE_FALLING ) +return 0 ; +else if (type ==IRQ_TYPE_EDGE_BOTH && + altera_gc->interrupt_trigger ==IRQ_TYPE_EDGE_BOTH ) +return 0 ; + } + +return -EINVAL ; +}
Whitespace issues here, obfuscating what's a condition and what what the instructions are. Given that the bodies do 'return', the 'elses' may be unnecessary and could save indentation levels.
drivers/gpio/gpio-altera.c line 228
Quoted 8 lines
drivers/gpio/gpio-altera.c new file +8
-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 i;
Several blocks of declarations? Remove the empty line.
And I'm really not a fan of mixing assignment instructions "this complex, beyond trivial initialization" with declarations. But this style is rather popular. :( Since this is a new driver, there is not reason to "continue" what others did before.
Quoted 17 lines
drivers/gpio/gpio-altera.c new file +17
-0+ +chained_irq_enter (chip, desc); +/* Handling for level trigger and edge trigger is different */ +if (altera_gc->interrupt_trigger ==IRQ_TYPE_LEVEL_HIGH ) { + status =readl_relaxed (mm_gc->regs +ALTERA_GPIO_DATA ); + status &=readl_relaxed (mm_gc->regs +ALTERA_GPIO_IRQ_MASK ); + +for (i =0 ; i < mm_gc->gc.ngpio; i++) { +if (status &BIT (i)) { +generic_handle_irq (irq_find_mapping ( + altera_gc->domain, i)); + } + } + }else { +while ((status = + (readl_relaxed (mm_gc->regs +ALTERA_GPIO_EDGE_CAP ) & +readl_relaxed (mm_gc->regs +ALTERA_GPIO_IRQ_MASK )))) {
drivers/gpio/gpio-altera.c line 255
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+writel_relaxed (status, + mm_gc->regs +ALTERA_GPIO_EDGE_CAP ); +for (i =0 ; i < mm_gc->gc.ngpio; i++) { +if (status &BIT (i)) { +generic_handle_irq (irq_find_mapping ( + altera_gc->domain, i)); + } + } + } + }
for_each_set_bit() might be more descriptive, save indentation levels, and could use availalbe CPU instructions.
There are more whitespace issues here.
And I'm afraid that use of _relaxed() accessors makes the driver depend on specific architectures, which should then reflect in the Kconfig dependencies.
Given that we are not talking about a GPIO block that is contained in SoCs which implement a specific CPU, but instead talk about an IP block that is supposed to get implemented in FPGAs connected to arbitrary CPUs, I'd suggest to use more portable instructions.
Quoted 6 lines
drivers/gpio/gpio-altera.c new file +4
-0+ +chained_irq_exit (chip, desc); +} +int altera_gpio_probe (struct platform_device *pdev)
drivers/gpio/gpio-altera.c line 285
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+{ +struct device_node *node = pdev->dev.of_node; +int i, id, reg, ret; +struct altera_gpio_chip *altera_gc =devm_kzalloc (&pdev->dev, +sizeof (*altera_gc),GFP_KERNEL ); +if (altera_gc ==NULL ) { +pr_err ("%s: out of memory\n" , node->full_name); +return -ENOMEM ; + } + altera_gc->domain =0 ;
This really needs a whitespace cleanup. I do suggest to clearly separate declarations and instructions. This reduces diffs upon further maintenance, and really isn't that expensive in terms of typing characters (which should not be a criterion during development anyway).
Quoted 6 lines
drivers/gpio/gpio-altera.c new file +6
-0+ +spin_lock_init (&altera_gc->gpio_lock); + + id = pdev->id; + +if (of_property_read_u32 (node,"altr,gpio-bank-width" , ®))
drivers/gpio/gpio-altera.c line 301
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+/*By default assume full GPIO controller*/ + altera_gc->mmchip.gc.ngpio =32 ; + else + altera_gc->mmchip.gc.ngpio = reg; + +if (altera_gc->mmchip.gc.ngpio >32 ) { +dev_warn (&pdev->dev, +"ngpio is greater than 32, defaulting to 32\n" ); + altera_gc->mmchip.gc.ngpio =32 ; + }
Does this "maximum number of supported pins per bank" deserve a symbolic name? The "32' is repeated here several times within a few lines.
Quoted 34 lines
drivers/gpio/gpio-altera.c new file +34
-0+ + altera_gc->mmchip.gc.direction_input = altera_gpio_direction_input; + altera_gc->mmchip.gc.direction_output = altera_gpio_direction_output; + altera_gc->mmchip.gc.get = altera_gpio_get; + altera_gc->mmchip.gc.set = altera_gpio_set; + altera_gc->mmchip.gc.to_irq = altera_gpio_to_irq; + altera_gc->mmchip.gc.owner =THIS_MODULE ; + + ret =of_mm_gpiochip_add (node, &altera_gc->mmchip); +if (ret) { +dev_err (&pdev->dev,"Failed adding memory mapped gpiochip\n" ); +return ret; + } + +platform_set_drvdata (pdev, altera_gc); + + altera_gc->mapped_irq =irq_of_parse_and_map (node,0 ); + +if (!altera_gc->mapped_irq) + goto skip_irq; + + altera_gc->domain =irq_domain_add_linear (node, + altera_gc->mmchip.gc.ngpio, &altera_gpio_irq_ops, altera_gc); + +if (!altera_gc->domain) { + ret = -ENODEV ; + goto dispose_irq; + } + +for (i =0 ; i < altera_gc->mmchip.gc.ngpio; i++) +irq_create_mapping (altera_gc->domain, i); + +if (of_property_read_u32 (node,"altr,interrupt_type" , ®)) { + ret = -EINVAL ;
drivers/gpio/gpio-altera.c line 345
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+dev_err (&pdev->dev, +"altr,interrupt_type value not set in device tree\n" ); + goto teardown; + } + altera_gc->interrupt_trigger = reg; + +irq_set_handler_data (altera_gc->mapped_irq, altera_gc); +irq_set_chained_handler (altera_gc->mapped_irq, altera_gpio_irq_handler); + +return 0 ;
The 'skip_irq' label should be here. It's really not an exceptional case or error path, it's just a shortcut for the absence of an optional feature.
Quoted 4 lines
drivers/gpio/gpio-altera.c new file +4
-0+ +teardown: +irq_domain_remove (altera_gc->domain); +dispose_irq:
drivers/gpio/gpio-altera.c line 359
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+irq_dispose_mapping (altera_gc->mapped_irq); +WARN_ON (gpiochip_remove (&altera_gc->mmchip.gc) <0 ); + +pr_err ("%s: registration failed with status %d\n" , + node->full_name, ret); + +return ret; +skip_irq: +return 0 ; +}
virtually yours
Gerhard Sittig
Signature · 3 lines
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr. 5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: [email protected]
Quoted 17 lines · Gerhard Sittig
Quoted 7 lines · On Mon, Mar 03, 2014 at 18:27 +0800, [email protected] wrote:
From: Tien Hock Loh <[email protected]>
Add driver support for Altera GPIO soft IP, including interrupts and I/O. Tested on Altera CV SoC board using dipsw and LED using LED framework.
Signed-off-by Tien Hock Loh <[email protected]>Quoted 8 lines
--- .../devicetree/bindings/gpio/gpio-altera.txt | 44 +++ drivers/gpio/Kconfig | 7 + drivers/gpio/Makefile | 1 + drivers/gpio/gpio-altera.c | 419 +++++++++++++++++++++ 4 files changed, 471 insertions(+) create mode 100644 Documentation/devicetree/bindings/gpio/gpio-altera.txt create mode 100644 drivers/gpio/gpio-altera.c
Quoted 5 lines
Let's see. A v7, i.e. quite some iterations done, and still whitespace issues, and not a single line of changelogs. Not exactly an invitation for reviewers to spend their time on it. It's hard to tell which feedback was received before, and what of it has been addressed.
Noted. I'll be sure to put changelogs and fix all the whitespace issues.
Quoted 4 lines
Aren't binding patches to be separate (and first) in series these days, while driver adjustment or introduction then follows to implement what was specified?
OK. I'll split the patch to two with DT bindings as the first one.
Quoted 15 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 8
Quoted 12 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +8
-0@@ -0,0 +1,44 @@ +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 1 + - The first cell is the gpio offset numberWill there never be any use of flags, that usually are kept in the second cell? Can we tell for sure that one cell is enough?
OK. I'll add a second cell and mark it as currently unused.
Quoted 7 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 13
Quoted 6 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +5
-0+- gpio-controller : Marks the device node as a GPIO controller. +- #interrupt-cells : Should be 1. + - The first cell is the GPIO offset number within the GPIO controller. +- interrupts: Specify the interrupt. +- interrupt-controller: Mark the device node as an interrupt controller
Quoted 5 lines
Is #interrupt-cells = <1> enough? Are triggers fixed in the hardware? A comment about it may prevent people asking questions. (The motivation might be mentioned below, see the comment on the two "required properties" sections.)
Yes the triggers are fixed in the hardware. I'll add a comment on this.
Quoted 6 lines
I'd sort the last three items differently. 'interrupts' is where the GPIO block is "a client" to the IRQ controller which it is connected to. '#interrupt-cells' is the "server side" because the GPIO block is an IRQ controller and has the 'interrupt-controller' property.
I'm assuming the order should be:
"interrupt-controller", "interrupt-cells", "interrupts". Am I correct here?
Quoted 9 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 15
Quoted 2 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +2
-0+ +Altera GPIO specific required properties:This made me stop and wonder. This is the "Altera GPIO controller bindings" document, and it has a "Required properties" section as well as an "Altera GPIO specific required properties" section? Isn't this highly redundant, and confusing at the same time?
Noted. I were referring to how other binding docs works and they have these "specific properties" header.
It seems I should just specify required properties and optional properties. I'll lump them together.
Quoted 26 lines
Quoted 26 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +26
-0+- altr,interrupt_trigger: Specifies the interrupt trigger type the GPIO + hardware is synthesized. This field is required if the Altera GPIO controller + used has IRQ enabled as the interrupt type is not software controlled, + but hardware synthesized. Required if GPIO is used as an interrupt + controller. The value is defined in <dt-bindings/interrupt-controller/irq.h> + Only the following flags are supported: + IRQ_TYPE_EDGE_RISING + IRQ_TYPE_EDGE_FALLING + IRQ_TYPE_EDGE_BOTH + IRQ_TYPE_LEVEL_HIGH + +Altera GPIO specific optional properties: +- altr,gpio-bank-width: Width of the GPIO bank. This defines how many pins the + GPIO device has. Ranges between 1-32. Optional and defaults to 32 is not + specified. + +Example: + +gpio_altr: gpio@40000 { + compatible = "altr,pio-1.0"; + reg = <0xff200000 0x10>; + interrupts = <0 45 4>; + altr,gpio-bank-width = <32>; + altr,interrupt_trigger = <IRQ_TYPE_EDGE_RISING>; + #gpio-cells = <1>; + gpio-controller;
Quoted 9 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt line 44
Quoted 3 lines
Documentation/devicetree/bindings/gpio/gpio-altera.txt new file +3
-0+ #interrupt-cells = <1>; + interrupt-controller; +};The node's address does not match the 'reg' property. The interrupt trigger uses symbolic names, so could the 'interrupts' spec. Examples should not assume an external 'interrupt-parent' but should list them for completeness.
Noted. I'll change the address.
Quoted 6 lines
Quoted 6 lines
drivers/gpio/Kconfig +0
-0@@ -128,6 +128,13 @@ config GPIO_GENERIC_PLATFORM help Say yes here to support basic platform_device memory-mapped GPIO controllers.
Quoted 14 lines
drivers/gpio/Kconfig config GPIO_GENERIC_PLATFORM · line 137
Quoted 7 lines
drivers/gpio/Kconfig +7
-0+config GPIO_ALTERA + tristate "Altera GPIO" + select IRQ_DOMAIN + depends on OF_GPIO + help + Say yes here to support the Altera PIO device. +I guess checkpatch nags about the rather short help text. Tristate options probably should mention the opportunity to build a module, and by convention should state what the module's name then would be (which in total gets you the minimum number of help text lines as a byproduct).
Noted. I'll update this.
Quoted 14 lines
Quoted 14 lines
drivers/gpio/gpio-altera.c new file +10
-0@@ -0,0 +1,419 @@ + +#include <linux/gpio.h> +#include <linux/init.h> +#include <linux/irqdomain.h> +#include <linux/interrupt.h> +#include <linux/io.h> +#include <linux/irq.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/of_irq.h>
Quoted 9 lines
drivers/gpio/gpio-altera.c line 32
Quoted 5 lines
drivers/gpio/gpio-altera.c new file +5
-0+#include <linux/of_gpio.h> +#include <linux/platform_device.h> +#include <linux/slab.h> + +#include <linux/irqchip/chained_irq.h> Includes should be alpha-sorted. To detect duplicates, and to avoid or reduce conflicts.
Noted.
Quoted 5 lines
Quoted 5 lines
drivers/gpio/gpio-altera.c new file +5
-0+ +#define ALTERA_GPIO_DATA 0x0 +#define ALTERA_GPIO_DIR 0x4 +#define ALTERA_GPIO_IRQ_MASK 0x8 +#define ALTERA_GPIO_EDGE_CAP 0xc
Quoted 9 lines
drivers/gpio/gpio-altera.c line 39
Quoted 2 lines
drivers/gpio/gpio-altera.c new file +2
-0+#define ALTERA_GPIO_OUTSET 0x10 +#define ALTERA_GPIO_OUTCLEAR 0x14OUTSET and OUTCLEAR are not used in the driver source. You might as well drop their declarations (after all you are not declaring the register set layout here, but are just introducing magic offset numbers for some of the registers that the driver may access).
OK noted.
Quoted 11 lines
drivers/gpio/gpio-altera.c line 65
Quoted 6 lines
drivers/gpio/gpio-altera.c new file +6
-0+static void altera_gpio_irq_unmask (struct irq_data *d) +{ +struct altera_gpio_chip *altera_gc =irq_data_get_irq_chip_data (d); +struct of_mm_gpio_chip *mm_gc = &altera_gc->mmchip; +unsigned long flags; +unsigned int intmask;I'd suggest <
stdint.h> style fixed width data types (uint32_t) or their kernel equivalents (u32) for operations on fixed width peripheral registers.
OK. Noted.
Quoted 30 lines
Quoted 30 lines
drivers/gpio/gpio-altera.c new file +29
-0+ +spin_lock_irqsave (&altera_gc->gpio_lock, flags); + intmask =readl (mm_gc->regs +ALTERA_GPIO_IRQ_MASK ); +/* Set ALTERA_GPIO_IRQ_MASK bit to unmask */ + intmask |=BIT (irqd_to_hwirq (d)); +writel (intmask, mm_gc->regs +ALTERA_GPIO_IRQ_MASK ); +spin_unlock_irqrestore (&altera_gc->gpio_lock, flags); +} +static int altera_gpio_irq_set_type (struct irq_data *d, +unsigned int type) +{ +struct altera_gpio_chip *altera_gc =irq_data_get_irq_chip_data (d); + +if (type ==IRQ_TYPE_NONE ) +return 0 ; + +if (type ==IRQ_TYPE_LEVEL_HIGH && + altera_gc->interrupt_trigger ==IRQ_TYPE_LEVEL_HIGH ) { +return 0 ; + }else { +if (type ==IRQ_TYPE_EDGE_RISING && + altera_gc->interrupt_trigger ==IRQ_TYPE_EDGE_RISING ) +return 0 ; +else if (type ==IRQ_TYPE_EDGE_FALLING && + altera_gc->interrupt_trigger ==IRQ_TYPE_EDGE_FALLING ) +return 0 ; +else if (type ==IRQ_TYPE_EDGE_BOTH && + altera_gc->interrupt_trigger ==IRQ_TYPE_EDGE_BOTH ) +return 0 ;
Quoted 9 lines
drivers/gpio/gpio-altera.c line 114
Quoted 4 lines
drivers/gpio/gpio-altera.c new file +4
-0+ } + +return -EINVAL ; +}Whitespace issues here, obfuscating what's a condition and what what the instructions are. Given that the bodies do 'return', the 'elses' may be unnecessary and could save indentation levels.
Noted. I'll fix this.
Quoted 11 lines
drivers/gpio/gpio-altera.c line 228
Quoted 8 lines
drivers/gpio/gpio-altera.c new file +8
-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 i;Several blocks of declarations? Remove the empty line.
Quoted 5 lines
And I'm really not a fan of mixing assignment instructions "this complex, beyond trivial initialization" with declarations. But this style is rather popular. :( Since this is a new driver, there is not reason to "continue" what others did before.
OK noted. I'll implement this correctly.
Quoted 37 lines
Quoted 17 lines
drivers/gpio/gpio-altera.c new file +17
-0+ +chained_irq_enter (chip, desc); +/* Handling for level trigger and edge trigger is different */ +if (altera_gc->interrupt_trigger ==IRQ_TYPE_LEVEL_HIGH ) { + status =readl_relaxed (mm_gc->regs +ALTERA_GPIO_DATA ); + status &=readl_relaxed (mm_gc->regs +ALTERA_GPIO_IRQ_MASK ); + +for (i =0 ; i < mm_gc->gc.ngpio; i++) { +if (status &BIT (i)) { +generic_handle_irq (irq_find_mapping ( + altera_gc->domain, i)); + } + } + }else { +while ((status = + (readl_relaxed (mm_gc->regs +ALTERA_GPIO_EDGE_CAP ) & +readl_relaxed (mm_gc->regs +ALTERA_GPIO_IRQ_MASK )))) {drivers/gpio/gpio-altera.c line 255
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+writel_relaxed (status, + mm_gc->regs +ALTERA_GPIO_EDGE_CAP ); +for (i =0 ; i < mm_gc->gc.ngpio; i++) { +if (status &BIT (i)) { +generic_handle_irq (irq_find_mapping ( + altera_gc->domain, i)); + } + } + } + }
for_each_set_bit()might be more descriptive, save indentation levels, and could use availalbe CPU instructions.There are more whitespace issues here.
And I'm afraid that use of
_relaxed()accessors makes the driver depend on specific architectures, which should then reflect in theKconfigdependencies.
Quoted 6 lines
Given that we are not talking about a GPIO block that is contained in SoCs which implement a specific CPU, but instead talk about an IP block that is supposed to get implemented in FPGAs connected to arbitrary CPUs, I'd suggest to use more portable instructions.
I'll use for_each_set_bit() function as the iterator.
I'll fix the whitespace issue and run the check_patch before submitting again. Yes. There are complains about using _relaxed(). I'll remove the _relaxed() function and use just writel and readl.
Quoted 14 lines
Quoted 14 lines
drivers/gpio/gpio-altera.c new file +12
-0+ +chained_irq_exit (chip, desc); +} +int altera_gpio_probe (struct platform_device *pdev) +{ +struct device_node *node = pdev->dev.of_node; +int i, id, reg, ret; +struct altera_gpio_chip *altera_gc =devm_kzalloc (&pdev->dev, +sizeof (*altera_gc),GFP_KERNEL ); +if (altera_gc ==NULL ) { +pr_err ("%s: out of memory\n" , node->full_name); +return -ENOMEM ;
Quoted 9 lines
drivers/gpio/gpio-altera.c
Quoted 2 lines
drivers/gpio/gpio-altera.c new file +2
-0+ } + altera_gc->domain =0 ;This really needs a whitespace cleanup. I do suggest to clearly separate declarations and instructions. This reduces diffs upon further maintenance, and really isn't that expensive in terms of typing characters (which should not be a criterion during development anyway).
Noted.
Quoted 12 lines
Quoted 12 lines
drivers/gpio/gpio-altera.c new file +12
-0+ +spin_lock_init (&altera_gc->gpio_lock); + + id = pdev->id; + +if (of_property_read_u32 (node,"altr,gpio-bank-width" , ®)) +/*By default assume full GPIO controller*/ + altera_gc->mmchip.gc.ngpio =32 ; +else + altera_gc->mmchip.gc.ngpio = reg; + +if (altera_gc->mmchip.gc.ngpio >32 ) {
Quoted 9 lines
drivers/gpio/gpio-altera.c line 301
Quoted 4 lines
drivers/gpio/gpio-altera.c new file +4
-0+dev_warn (&pdev->dev, +"ngpio is greater than 32, defaulting to 32\n" ); + altera_gc->mmchip.gc.ngpio =32 ; + }Does this "maximum number of supported pins per bank" deserve a symbolic name? The "32' is repeated here several times within a few lines.
OK I'll add a macro to the number.
Quoted 40 lines
Quoted 40 lines
drivers/gpio/gpio-altera.c new file +40
-0+ + altera_gc->mmchip.gc.direction_input = altera_gpio_direction_input; + altera_gc->mmchip.gc.direction_output = altera_gpio_direction_output; + altera_gc->mmchip.gc.get = altera_gpio_get; + altera_gc->mmchip.gc.set = altera_gpio_set; + altera_gc->mmchip.gc.to_irq = altera_gpio_to_irq; + altera_gc->mmchip.gc.owner =THIS_MODULE ; + + ret =of_mm_gpiochip_add (node, &altera_gc->mmchip); +if (ret) { +dev_err (&pdev->dev,"Failed adding memory mapped gpiochip\n" ); +return ret; + } + +platform_set_drvdata (pdev, altera_gc); + + altera_gc->mapped_irq =irq_of_parse_and_map (node,0 ); + +if (!altera_gc->mapped_irq) + goto skip_irq; + + altera_gc->domain =irq_domain_add_linear (node, + altera_gc->mmchip.gc.ngpio, &altera_gpio_irq_ops, altera_gc); + +if (!altera_gc->domain) { + ret = -ENODEV ; + goto dispose_irq; + } + +for (i =0 ; i < altera_gc->mmchip.gc.ngpio; i++) +irq_create_mapping (altera_gc->domain, i); + +if (of_property_read_u32 (node,"altr,interrupt_type" , ®)) { + ret = -EINVAL ; +dev_err (&pdev->dev, +"altr,interrupt_type value not set in device tree\n" ); + goto teardown; + } + altera_gc->interrupt_trigger = reg; +
Quoted 9 lines
drivers/gpio/gpio-altera.c line 345
Quoted 4 lines
drivers/gpio/gpio-altera.c new file +4
-0+irq_set_handler_data (altera_gc->mapped_irq, altera_gc); +irq_set_chained_handler (altera_gc->mapped_irq, altera_gpio_irq_handler); + +return 0 ;The '
skip_irq' label should be here. It's really not an exceptional case or error path, it's just a shortcut for the absence of an optional feature.
OK noted.
Quoted 22 lines
Quoted 4 lines
drivers/gpio/gpio-altera.c new file +4
-0+ +teardown: +irq_domain_remove (altera_gc->domain); +dispose_irq:drivers/gpio/gpio-altera.c line 359
Quoted 10 lines
drivers/gpio/gpio-altera.c new file +10
-0+irq_dispose_mapping (altera_gc->mapped_irq); +WARN_ON (gpiochip_remove (&altera_gc->mmchip.gc) <0 ); + +pr_err ("%s: registration failed with status %d\n" , + node->full_name, ret); + +return ret; +skip_irq: +return 0 ; +}virtually yours
Gerhard Sittig
--
DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr. 5, D-82194 Groebenzell, Germany
Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: [email protected]
Quoted 6 lines · On Mon, Mar 3, 2014 at 11:27 AM, <[email protected]> wrote:
From: Tien Hock Loh <[email protected]>
Add driver support for Altera GPIO soft IP, including interrupts and I/O. Tested on Altera CV SoC board using dipsw and LED using LED framework.
Signed-off-by Tien Hock Loh <[email protected]>
Some time has passed since this was posted and we now have GPIO irqchip helpers in the gpiolib core.
If you repost this, please reform this driver to use the gpiochip irqchip helpers in the manner of e.g. drivers/gpio/gpio-pl061.c.
More examples and documentation is available on the "devel" branch in the gpio git tree (also in linux-next).
Yours,
Linus Walleij
Quoted 9 lines · Linus Walleij
Quoted 6 lines · On Mon, Mar 3, 2014 at 11:27 AM, <[email protected]> wrote:
From: Tien Hock Loh <[email protected]>
Add driver support for Altera GPIO soft IP, including interrupts and I/O. Tested on Altera CV SoC board using dipsw and LED using LED framework.
Signed-off-by Tien Hock Loh <[email protected]>
Quoted 8 lines
Some time has passed since this was posted and we now have GPIO irqchip helpers in the gpiolib core.
If you repost this, please reform this driver to use the gpiochip irqchip helpers in the manner of e.g.
drivers/gpio/gpio-pl061.c.More examples and documentation is available on the "devel" branch in the gpio git tree (also in linux-next).
OK noted. I'll do that.
Thanks for the info.
Quoted 3 lines
Yours,
Linus Walleij