vf610: Add GPIO support
Subsystems (4)
cover vf610: Add GPIO support
This patchset adds GPIO support for Vybrid. The first patch is a preparation patch which makes sure we can detect whether a pin is initialized by the pinmux subsystem or not. This is required since the gpio_request_enable/gpio_set_direction function need to know if the pins mux register offsets are valid or not.
To get valid mux register offset for a pin, we need to have a pinmux group configuration in the device tree. For the SD card detection pin this is within the esdhc pingroup. Since the pingroup is used by the driver, the mux and configuration register gets applied by the pinmux subsystem.
However, the group does not need to be used anywhere, just the configuration itself makes sure the mux register offset is known to the pinmux driver. If a GPIO is requested then, the SoC's default pin config is still in place. Is this the expected behaviour of a pinmux driver? How can I make sure that a GPIO only pin (used by user-space for instance) get a valid pin configuration applied?
To get the mux register offset, I also thought about calculating the value in those two functions instead, however this seems not to be consistent with the rest of the driver.
Other then that, the GPIO/IRQ chip driver is quite straight forward and makes use of CONFIG_GPIOLIB_IRQCHIP. The only thing which is a bit weird: There is no actual mask register for the GPIO interrupts. There is only a configuration register, which, if configured, enables the interrupt. Since the mask/unmask functions are requierd, the driver writes/clears the configuration register in the mask/umask functions.
Stefan Agner (4):
pinctrl: imx: detect uninitialized pins
pinctrl: imx: add gpio pinmux support for vf610
gpio: vf610: add gpiolib/IRQ chip driver for Vybird
ARM: dts: vf610: Use new GPIO support
arch/arm/boot/dts/vf610-colibri.dtsi | 8 + arch/arm/boot/dts/vf610-twr.dts | 1 + arch/arm/boot/dts/vf610.dtsi | 1 + drivers/gpio/Kconfig | 7 + drivers/gpio/Makefile | 1 + drivers/gpio/gpio-vf610.c | 285 +++++++++++++++++++++++++++++++++++ drivers/pinctrl/pinctrl-imx.c | 63 +++++++- drivers/pinctrl/pinctrl-imx.h | 8 +- drivers/pinctrl/pinctrl-vf610.c | 4 +- 9 files changed, 368 insertions(+), 10 deletions(-) create mode 100644 drivers/gpio/gpio-vf610.c
Signature · 1 line
2.1.0
PATCH 1/4 pinctrl: imx: detect uninitialized pins
The pinctrl driver initialized the register offsets for the pins with 0. On Vybrid an offset of 0 is a valid offset for the pinctrl mux register. So far, this was solved using the ZERO_OFFSET_VALID flag which allowed offsets of 0. However, this does not allow to verify whether a pins struct imx_pmx_func was initialized or not.
Use signed offset values for register offsets and initialize those with -1 in order to detect uninitialized offset values reliable.
--- drivers/pinctrl/pinctrl-imx.c | 9 +++++---- drivers/pinctrl/pinctrl-imx.h | 7 +++---- drivers/pinctrl/pinctrl-vf610.c | 2 +- 3 files changed, 9 insertions(+), 9 deletions(-)
drivers/pinctrl/pinctrl-imx.c +5 -4
@@ -204,7 +204,7 @@ static int imx_pmx_enable(struct pinctrl_dev *pctldev, unsigned selector, pin_id = pin->pin; pin_reg = &info->pin_regs[pin_id]; -if (!(info->flags &ZERO_OFFSET_VALID ) && !pin_reg->mux_reg) { +if (pin_reg->mux_reg ==-1 ) {dev_err (ipctl->dev,"Pin(%s) does not support mux function\n" , info->pins[pin_id].name);return -EINVAL ;
@@ -308,7 +308,7 @@ static int imx_pinconf_get(struct pinctrl_dev *pctldev,const struct imx_pinctrl_soc_info *info = ipctl->info;const struct imx_pin_reg *pin_reg = &info->pin_regs[pin_id]; -if (!(info->flags &ZERO_OFFSET_VALID ) && !pin_reg->conf_reg) { +if (pin_reg->conf_reg ==-1 ) {dev_err (info->dev,"Pin(%s) does not support config function\n" , info->pins[pin_id].name);return -EINVAL ;
@@ -331,7 +331,7 @@ static int imx_pinconf_set(struct pinctrl_dev *pctldev,const struct imx_pin_reg *pin_reg = &info->pin_regs[pin_id];int i; -if (!(info->flags &ZERO_OFFSET_VALID ) && !pin_reg->conf_reg) { +if (pin_reg->conf_reg ==-1 ) {dev_err (info->dev,"Pin(%s) does not support config function\n" , info->pins[pin_id].name);return -EINVAL ;
@@ -586,10 +586,11 @@ int imx_pinctrl_probe(struct platform_device *pdev,if (!ipctl)return -ENOMEM ; - info->pin_regs =devm_kzalloc (&pdev->dev,sizeof (*info->pin_regs) * + info->pin_regs =devm_kmalloc (&pdev->dev,sizeof (*info->pin_regs) * info->npins,GFP_KERNEL );if (!info->pin_regs)return -ENOMEM ; +memset (info->pin_regs,0xff ,sizeof (*info->pin_regs) * info->npins); res =platform_get_resource (pdev,IORESOURCE_MEM ,0 ); ipctl->base =devm_ioremap_resource (&pdev->dev, res);
drivers/pinctrl/pinctrl-imx.h +3 -4
@@ -67,8 +67,8 @@ struct imx_pmx_func {* @conf_reg: config register offset */ struct imx_pin_reg { -u16 mux_reg; -u16 conf_reg; +s16 mux_reg; +s16 conf_reg; };struct imx_pinctrl_soc_info {
@@ -83,8 +83,7 @@ struct imx_pinctrl_soc_info {unsigned int flags; }; -#define ZERO_OFFSET_VALID 0x1 -#define SHARE_MUX_CONF_REG 0x2 +#define SHARE_MUX_CONF_REG 0x1#define NO_MUX 0x0#define NO_PAD 0x0
drivers/pinctrl/pinctrl-vf610.c +1 -1
@@ -299,7 +299,7 @@ static const struct pinctrl_pin_desc vf610_pinctrl_pads[] = {static struct imx_pinctrl_soc_info vf610_pinctrl_info = { .pins = vf610_pinctrl_pads, .npins =ARRAY_SIZE (vf610_pinctrl_pads), - .flags =ZERO_OFFSET_VALID |SHARE_MUX_CONF_REG , + .flags =SHARE_MUX_CONF_REG , };static struct of_device_id vf610_pinctrl_of_match[] = {
Signature · 1 line
2.1.0
PATCH 2/4 pinctrl: imx: add gpio pinmux support for vf610
In reply to Stefan AgnerAdd pinmux support for GPIO for Vybrid (vf610) IOMUX controller. This is needed since direction configuration is not part of the GPIO module in Vybrid.
--- drivers/pinctrl/pinctrl-imx.c | 54 +++++++++++++++++++++++++++++++++++++++++ drivers/pinctrl/pinctrl-imx.h | 1 + drivers/pinctrl/pinctrl-vf610.c | 4 +-- 3 files changed, 57 insertions(+), 2 deletions(-)
drivers/pinctrl/pinctrl-imx.c +54 -0
@@ -294,10 +294,59 @@ static int imx_pmx_get_groups(struct pinctrl_dev *pctldev, unsigned selector,return 0 ; } +static int imx_pmx_gpio_request_enable (struct pinctrl_dev *pctldev, +struct pinctrl_gpio_range *range,unsigned offset) +{ +struct imx_pinctrl *ipctl =pinctrl_dev_get_drvdata (pctldev); +const struct imx_pinctrl_soc_info *info = ipctl->info; +const struct imx_pin_reg *pin_reg; +u32 reg; + +if (!(info->flags &GPIO_CONTROL )) +return -EINVAL ; + + pin_reg = &info->pin_regs[offset]; +if (pin_reg->mux_reg ==-1 ) +return -EINVAL ; + + reg =readl (ipctl->base + pin_reg->mux_reg); + reg &= ~(0x7 <<20 ); +writel (reg, ipctl->base + pin_reg->mux_reg); + +return 0 ; +} + +static int imx_pmx_gpio_set_direction (struct pinctrl_dev *pctldev, +struct pinctrl_gpio_range *range,unsigned offset,bool input) +{ +struct imx_pinctrl *ipctl =pinctrl_dev_get_drvdata (pctldev); +const struct imx_pinctrl_soc_info *info = ipctl->info; +const struct imx_pin_reg *pin_reg; +u32 reg; + +if (!(info->flags &GPIO_CONTROL )) +return -EINVAL ; + + pin_reg = &info->pin_regs[offset]; +if (pin_reg->mux_reg ==-1 ) +return -EINVAL ; + + reg =readl (ipctl->base + pin_reg->mux_reg); +if (input) + reg &= ~0x2 ; +else + reg |=0x2 ; +writel (reg, ipctl->base + pin_reg->mux_reg); + +return 0 ; +} +static const struct pinmux_ops imx_pmx_ops = { .get_functions_count = imx_pmx_get_funcs_count, .get_function_name = imx_pmx_get_func_name, .get_function_groups = imx_pmx_get_groups, + .gpio_request_enable = imx_pmx_gpio_request_enable, + .gpio_set_direction = imx_pmx_gpio_set_direction, .enable = imx_pmx_enable, };
@@ -579,6 +628,11 @@ int imx_pinctrl_probe(struct platform_device *pdev,dev_err (&pdev->dev,"wrong pinctrl info\n" );return -EINVAL ; } + +/* GPIO control functions only intended for shared mux/conf register */ +if (info->flags &GPIO_CONTROL ) +BUG_ON (!(info->flags &SHARE_MUX_CONF_REG )); + info->dev = &pdev->dev;/* Create state holders etc for this driver */
drivers/pinctrl/pinctrl-imx.h +1 -0
@@ -84,6 +84,7 @@ struct imx_pinctrl_soc_info { };#define SHARE_MUX_CONF_REG 0x1 +#define GPIO_CONTROL 0x2#define NO_MUX 0x0#define NO_PAD 0x0
drivers/pinctrl/pinctrl-vf610.c +2 -2
@@ -299,7 +299,7 @@ static const struct pinctrl_pin_desc vf610_pinctrl_pads[] = {static struct imx_pinctrl_soc_info vf610_pinctrl_info = { .pins = vf610_pinctrl_pads, .npins =ARRAY_SIZE (vf610_pinctrl_pads), - .flags =SHARE_MUX_CONF_REG , + .flags =SHARE_MUX_CONF_REG |GPIO_CONTROL , };static struct of_device_id vf610_pinctrl_of_match[] = {
@@ -326,7 +326,7 @@ static int __init vf610_pinctrl_init(void) {return platform_driver_register (&vf610_pinctrl_driver); } -arch_initcall (vf610_pinctrl_init); +postcore_initcall (vf610_pinctrl_init); static void __exit vf610_pinctrl_exit(void) {
Signature · 1 line
2.1.0
PATCH 3/4 gpio: vf610: add gpiolib/IRQ chip driver for Vybird
In reply to Stefan AgnerAdd a gpiolib and IRQ chip driver for Vybrid ARM SoC using the Vybrid's GPIO and PORT module. The driver is instanced once per each GPIO/PORT module pair and handles 32 GPIO's.
--- drivers/gpio/Kconfig | 7 ++ drivers/gpio/Makefile | 1 + drivers/gpio/gpio-vf610.c | 285 ++++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 293 insertions(+) create mode 100644 drivers/gpio/gpio-vf610.c
drivers/gpio/Kconfig +7 -0
@@ -334,6 +334,13 @@ config GPIO_TZ1090_PDC help Say yes here to support Toumaz Xenif TZ1090 PDC GPIOs. +config GPIO_VF610 + def_bool y + depends on ARCH_MXC && SOC_VF610 + select GPIOLIB_IRQCHIP + help + Say yes here to support Vybrid vf610 GPIOs. + config GPIO_XILINX bool "Xilinx GPIO support" depends on PPC_OF || MICROBLAZE || ARCH_ZYNQ
drivers/gpio/Makefile +1 -0
@@ -95,6 +95,7 @@ obj-$(CONFIG_GPIO_TWL6040) += gpio-twl6040.o obj-$(CONFIG_GPIO_TZ1090) += gpio-tz1090.o obj-$(CONFIG_GPIO_TZ1090_PDC) += gpio-tz1090-pdc.o obj-$(CONFIG_GPIO_UCB1400) += gpio-ucb1400.o +obj-$(CONFIG_GPIO_VF610) += gpio-vf610.o obj-$(CONFIG_GPIO_VIPERBOARD) += gpio-viperboard.o obj-$(CONFIG_GPIO_VR41XX) += gpio-vr41xx.o obj-$(CONFIG_GPIO_VX855) += gpio-vx855.o
drivers/gpio/gpio-vf610.c new file +285 -0
285 lines not on this page · raw message
Signature · 1 line
2.1.0
PATCH 4/4 ARM: dts: vf610: Use new GPIO support
In reply to Stefan AgnerUse GPIO support by adding SD card detection configuration and GPIO pinmux for Colibri's standard GPIO pins.
--- arch/arm/boot/dts/vf610-colibri.dtsi | 8 ++++++++ arch/arm/boot/dts/vf610-twr.dts | 1 + arch/arm/boot/dts/vf610.dtsi | 1 + 3 files changed, 10 insertions(+)
arch/arm/boot/dts/vf610-colibri.dtsi +8 -0
@@ -31,6 +31,7 @@ pinctrl-names = "default"; pinctrl-0 = <&pinctrl_esdhc1>; bus-width = <4>; + cd-gpios = <&gpio2 10 GPIO_ACTIVE_LOW>; }; &fec1 {
@@ -71,6 +72,13 @@ &iomuxc { vf610-colibri { + pinctrl_gpio: gpios { + fsl,pins = < + VF610_PAD_PTD9__GPIO_88 0x219d + VF610_PAD_PTD10__GPIO_89 0x219d + >; + }; + pinctrl_esdhc1: esdhc1grp { fsl,pins = < VF610_PAD_PTA24__ESDHC1_CLK 0x31ef
arch/arm/boot/dts/vf610-twr.dts +1 -0
@@ -116,6 +116,7 @@ pinctrl-names = "default"; pinctrl-0 = <&pinctrl_esdhc1>; bus-width = <4>; + cd-gpios = <&gpio5 6 GPIO_ACTIVE_LOW>; status = "okay"; };
arch/arm/boot/dts/vf610.dtsi +1 -0
@@ -11,6 +11,7 @@ #include "vf610-pinfunc.h" #include <dt-bindings/clock/vf610-clock.h> #include <dt-bindings/interrupt-controller/irq.h> +#include <dt-bindings/gpio/gpio.h> / { aliases {
Signature · 1 line
2.1.0
Quoted 1 line · Stefan Agner
This patchset adds GPIO support for Vybrid.
Can I get some review from the i.MX driver maintainers for this patchset?
Yours,
Linus Walleij
Quoted 6 lines · Linus Walleij
Quoted 1 line · Stefan Agner
This patchset adds GPIO support for Vybrid.
Can I get some review from the i.MX driver maintainers for this patchset?
I will now just apply the patches.
Yours,
Linus Walleij
Re: [PATCH 1/4] pinctrl: imx: detect uninitialized pins
In reply to Stefan AgnerQuoted 10 lines · Stefan Agner
The pinctrl driver initialized the register offsets for the pins with 0. On Vybrid an offset of 0 is a valid offset for the pinctrl mux register. So far, this was solved using the
ZERO_OFFSET_VALIDflag which allowed offsets of 0. However, this does not allow to verify whether a pins structimx_pmx_funcwas initialized or not.Use signed offset values for register offsets and initialize those with -1 in order to detect uninitialized offset values reliable.
Signed-off-by Stefan Agner <[email protected]>
Patch applied.
Yours,
Linus Walleij
Re: [PATCH 2/4] pinctrl: imx: add gpio pinmux support for vf610
In reply to Stefan AgnerQuoted 5 lines · Stefan Agner
Add pinmux support for GPIO for Vybrid (vf610) IOMUX controller. This is needed since direction configuration is not part of the GPIO module in Vybrid.
Signed-off-by Stefan Agner <[email protected]>
(...)
drivers/pinctrl/pinctrl-vf610.c static int __init vf610_pinctrl_init(void) · line 329
Quoted 2 lines
drivers/pinctrl/pinctrl-vf610.c +1
-1-arch_initcall (vf610_pinctrl_init); +postcore_initcall (vf610_pinctrl_init);
Why is this necessary? You should be able to rely on deferred probing to do its work here. I think this should be module_init() or driver_initcall() really.
Yours,
Linus Walleij
Re: [PATCH 3/4] gpio: vf610: add gpiolib/IRQ chip driver for Vybird
In reply to Stefan AgnerQuoted 5 lines · Stefan Agner
Add a gpiolib and IRQ chip driver for Vybrid ARM SoC using the Vybrid's GPIO and PORT module. The driver is instanced once per each GPIO/PORT module pair and handles 32 GPIO's.
Signed-off-by Stefan Agner <[email protected]>
(...)
Quoted 2 lines
obj-$(CONFIG_GPIO_UCB1400) += gpio-ucb1400.o
+obj-$(CONFIG_GPIO_VF610) += gpio-vf610.o
Some like to keep GPIOs tightly associated with a pin controller in a file next to the pin controller.
I.e. in drivers/pinctrl/freescale/gpio-vf610.c
But this works too. Any preference?
Quoted 1 line
+#define
GPIO_PER_PORT32
Very generic define. VF610_GPIOS_PER_PORT?
drivers/gpio/gpio-vf610.c line 40
Quoted 6 lines
drivers/gpio/gpio-vf610.c new file +6
-0+struct vf610_gpio_port { +struct gpio_chip gc; +void __iomem *base; +void __iomem *gpio_base; +u8 irqc[GPIO_PER_PORT ]; +int irq;
irq? Why do you need to keep this around?
drivers/gpio/gpio-vf610.c line 72
Quoted 8 lines
drivers/gpio/gpio-vf610.c new file +8
-0+static const struct of_device_id vf610_gpio_dt_ids[] = { + { .compatible ="fsl,vf610-gpio" }, + {/* sentinel */ } +}; + +static inline void vf610_gpio_writel (u32 val,void __iomem *reg) +{ +__raw_writel (val, reg);
Use writel_relaxed() instead unless you can explain why you want this.
(Same for all occurences.)
drivers/gpio/gpio-vf610.c line 95
Quoted 6 lines
drivers/gpio/gpio-vf610.c new file +6
-0+static int vf610_gpio_get (struct gpio_chip *gc,unsigned int gpio) +{ +struct vf610_gpio_port *port = + container_of(gc, struct vf610_gpio_port, gc); + +return !!(vf610_gpio_readl (port->gpio_base +GPIO_PDIR ) &1 << gpio);
#include <linux/bitops.h>
... & BIT(gpio)
drivers/gpio/gpio-vf610.c line 102
Quoted 5 lines
drivers/gpio/gpio-vf610.c new file +5
-0+static void vf610_gpio_set (struct gpio_chip *gc,unsigned int gpio,int val) +{ +struct vf610_gpio_port *port = + container_of(gc, struct vf610_gpio_port, gc); +unsigned long mask =1 << gpio;
= BIT(gpio);
drivers/gpio/gpio-vf610.c line 135
Quoted 13 lines
drivers/gpio/gpio-vf610.c new file +13
-0+static void vf610_gpio_irq_handler (u32 irq,struct irq_desc *desc) +{ +struct vf610_gpio_port *port =irq_get_handler_data (irq); +struct irq_chip *chip =irq_desc_get_chip (desc); +int pin; +unsigned long irq_isfr; + +chained_irq_enter (chip, desc); + + irq_isfr =vf610_gpio_readl (port->base +PORT_ISFR ); + +for_each_set_bit (pin, &irq_isfr,GPIO_PER_PORT ) { +vf610_gpio_writel (1 << pin, port->base +PORT_ISFR );
BIT(pin)
(etc)
drivers/gpio/gpio-vf610.c line 232
Quoted 3 lines
drivers/gpio/gpio-vf610.c new file +3
-0+ port->irq =irq_of_parse_and_map (pdev->dev.of_node,0 ); +if (port->irq ==NO_IRQ ) +return -ENODEV ;
Can't you just use a local int irq; variable for this?
drivers/gpio/gpio-vf610.c line 281
Quoted 5 lines
drivers/gpio/gpio-vf610.c new file +5
-0+static int __init gpio_vf610_init (void ) +{ +return platform_driver_register (&vf610_gpio_driver); +} +postcore_initcall (gpio_vf610_init);
postcore again. I don't like this, can you get rid of it?
Overall the driver looks very nice except for these nitty gritty details.
Yours,
Linus Walleij
Re: [PATCH 4/4] ARM: dts: vf610: Use new GPIO support
In reply to Stefan AgnerQuoted 4 lines · Stefan Agner
Use GPIO support by adding SD card detection configuration and GPIO pinmux for Colibri's standard GPIO pins.
Signed-off-by Stefan Agner <[email protected]>
I guess you need the rest to be merged first though?
Yours,
Linus Walleij
Re: [PATCH 2/4] pinctrl: imx: add gpio pinmux support for vf610
In reply to Linus WalleijAm 2014-09-23 11:48, schrieb Linus Walleij:
Quoted 7 lines · Stefan Agner
Quoted 6 lines
Add pinmux support for GPIO for Vybrid (vf610) IOMUX controller. This is needed since direction configuration is not part of the GPIO module in Vybrid.
Signed-off-by Stefan Agner <[email protected]>
Quoted 8 lines
(...)
drivers/pinctrl/pinctrl-vf610.c static int __init vf610_pinctrl_init(void) · line 329
Quoted 2 lines
drivers/pinctrl/pinctrl-vf610.c +1
-1-arch_initcall (vf610_pinctrl_init); +postcore_initcall (vf610_pinctrl_init);Why is this necessary? You should be able to rely on deferred probing to do its work here. I think this should be
module_init()ordriver_initcall()really.
Currently deferred probe doesn't work for gpiolib drivers which try to add gpio-ranges from device tree:gpiochip_add calls of_gpiochip_add_pin_range (through of_gpiochip_add). This function tries to get the pinctrl driver, which is not registred at that time. Currently the driver does not defer probing but fails silently...
We would need to alter the return values of those two functions (of_gpiochip_add_pin_range/of_gpiochip_add) and honor the return value in gpiochip_add.
Currently, it seems that we quite often use an early initcall to get the pinctrl loaded early:
$ grep -h -o ".*_initcall" drivers/pinctrl/*.c | sort | uniq -c
18 arch_initcall
3 core_initcall
3 postcore_initcall
1 subsys_initcallIMHO for those core services, it also makes sense to have them initialized early. I'd rather prefer this hard coded than having dozens of "requests probe deferral" messages...
--
Stefan
Re: [PATCH 3/4] gpio: vf610: add gpiolib/IRQ chip driver for Vybird
In reply to Linus WalleijHi Linus,
Thanks for your review!
Am 2014-09-23 11:58, schrieb Linus Walleij:
Quoted 11 lines · Stefan Agner
Quoted 6 lines
Add a gpiolib and IRQ chip driver for Vybrid ARM SoC using the Vybrid's GPIO and PORT module. The driver is instanced once per each GPIO/PORT module pair and handles 32 GPIO's.
Signed-off-by Stefan Agner <[email protected]>(...)
Quoted 2 lines
obj-$(CONFIG_GPIO_UCB1400) += gpio-ucb1400.o
+obj-$(CONFIG_GPIO_VF610) += gpio-vf610.o
Quoted 6 lines
Some like to keep GPIOs tightly associated with a pin controller in a file next to the pin controller.
I.e. in
drivers/pinctrl/freescale/gpio-vf610.cBut this works too. Any preference?
The other Freescale GPIO drivers (e.g. gpio-mxs.c/gpio-mxc.c) are located under drivers/gpio/ hence I would prefer to leave them there, even we use pinctrl here. Unless someone at Freescale has another opinion on this?
Quoted 4 lines
Quoted 2 lines
+#define
GPIO_PER_PORT32Very generic define.
VF610_GPIOS_PER_PORT?
Agreed
Quoted 21 lines
drivers/gpio/gpio-vf610.c line 40
Quoted 7 lines
drivers/gpio/gpio-vf610.c new file +6
-0+struct vf610_gpio_port { +struct gpio_chip gc; +void __iomem *base; +void __iomem *gpio_base; +u8 irqc[GPIO_PER_PORT ]; +int irq;irq? Why do you need to keep this around?
drivers/gpio/gpio-vf610.c line 72
Quoted 8 lines
drivers/gpio/gpio-vf610.c new file +8
-0+static const struct of_device_id vf610_gpio_dt_ids[] = { + { .compatible ="fsl,vf610-gpio" }, + {/* sentinel */ } +}; + +static inline void vf610_gpio_writel (u32 val,void __iomem *reg) +{ +__raw_writel (val, reg);Use
writel_relaxed()instead unless you can explain why you want this.
Quoted 1 line
(Same for all occurences.)
Agreed, I have don't know why I used the __raw variant here. I think its because copied this two stubs from gpio-tegra.c.
Quoted 36 lines
drivers/gpio/gpio-vf610.c line 95
Quoted 7 lines
drivers/gpio/gpio-vf610.c new file +6
-0+static int vf610_gpio_get (struct gpio_chip *gc,unsigned int gpio) +{ +struct vf610_gpio_port *port = + container_of(gc, struct vf610_gpio_port, gc); + +return !!(vf610_gpio_readl (port->gpio_base +GPIO_PDIR ) &1 << gpio);#include <
linux/bitops.h>... & BIT(gpio)
drivers/gpio/gpio-vf610.c line 102
Quoted 5 lines
drivers/gpio/gpio-vf610.c new file +5
-0+static void vf610_gpio_set (struct gpio_chip *gc,unsigned int gpio,int val) +{ +struct vf610_gpio_port *port = + container_of(gc, struct vf610_gpio_port, gc); +unsigned long mask =1 << gpio;= BIT(gpio);
drivers/gpio/gpio-vf610.c line 135
Quoted 13 lines
drivers/gpio/gpio-vf610.c new file +13
-0+static void vf610_gpio_irq_handler (u32 irq,struct irq_desc *desc) +{ +struct vf610_gpio_port *port =irq_get_handler_data (irq); +struct irq_chip *chip =irq_desc_get_chip (desc); +int pin; +unsigned long irq_isfr; + +chained_irq_enter (chip, desc); + + irq_isfr =vf610_gpio_readl (port->base +PORT_ISFR ); + +for_each_set_bit (pin, &irq_isfr,GPIO_PER_PORT ) { +vf610_gpio_writel (1 << pin, port->base +PORT_ISFR );BIT(pin)
Quoted 1 line
(etc)
Ok, will replace those bit operations.
Quoted 7 lines
drivers/gpio/gpio-vf610.c line 232
Quoted 4 lines
drivers/gpio/gpio-vf610.c new file +3
-0+ port->irq =irq_of_parse_and_map (pdev->dev.of_node,0 ); +if (port->irq ==NO_IRQ ) +return -ENODEV ;Can't you just use a local int irq; variable for this?
Quoted 7 lines
drivers/gpio/gpio-vf610.c line 281
Quoted 5 lines
drivers/gpio/gpio-vf610.c new file +5
-0+static int __init gpio_vf610_init (void ) +{ +return platform_driver_register (&vf610_gpio_driver); +} +postcore_initcall (gpio_vf610_init);postcore again. I don't like this, can you get rid of it?
I guess we could load this driver easily a bit later. IMHO, since lots of other driver use GPIO's, we should it load before all the drivers gets loaded (before device_initcall).
Most GPIO driver do this, some statistic again:
$ grep -h -o ".*_initcall" drivers/gpio/*.c | sort | uniq -c | sort -n
-r
33 subsys_initcall
14 postcore_initcall
2 device_initcall
2 arch_initcall
1 late_initcall
1 core_initcallMy proposal:
Use subsys_initcall (which is called after arch_initcall) in this GPIO driver and leave the pinctrl driver as arch_initcall. This way we are absolutely sure that the GPIO driver gets loaded after the pinctrl and also leave the pinctrl at its current value.
--
Stefan
Quoted 5 lines
Overall the driver looks very nice except for these nitty gritty details.
Yours,
Linus Walleij
Re: [PATCH 3/4] gpio: vf610: add gpiolib/IRQ chip driver for Vybird
Quoted 6 lines · Stefan Agner
[Me]
Quoted 1 line
postcore again. I don't like this, can you get rid of it?
I guess we could load this driver easily a bit later. IMHO, since lots of other driver use GPIO's, we should it load before all the drivers gets loaded (before
device_initcall).
Nope. We use deferred probing to control that today. Ideally all drivers should be device_initcall() and deferred probe be used to order things, not by playing around with initcalls.
Quoted 9 lines
Most GPIO driver do this, some statistic again:
$ grep -h -o ".*_initcall" drivers/gpio/*.c | sort | uniq -c | sort -n -r 33 subsys_initcall 14 postcore_initcall 2 device_initcall 2 arch_initcall 1 late_initcall 1 core_initcall
Yeah old legacy. There are patch attacks to get rid of this.
The reason we can't just change them is because sometimes dependent drivers do not handle the errorpath very well can can't defer cleanly.
With a new driver I expect deferred probe to be used.
Yours,
Linus Walleij
Re: [PATCH 3/4] gpio: vf610: add gpiolib/IRQ chip driver for Vybird
Am 2014-09-24 13:10, schrieb Linus Walleij:
Quoted 12 lines · Stefan Agner
Quoted 6 lines
[Me]
Quoted 1 line
postcore again. I don't like this, can you get rid of it?
I guess we could load this driver easily a bit later. IMHO, since lots of other driver use GPIO's, we should it load before all the drivers gets loaded (before
device_initcall).Nope. We use deferred probing to control that today. Ideally all drivers should be
device_initcall()and deferred probe be used to order things, not by playing around with initcalls.
You can not "nope" my humble opinion, this is not possible, its write protected! :-)
I fully understand the deferred probe approach, and I also think its really a good approach for almost all drivers. The system by itself makes sure the drivers are loaded in correct order. But if all drivers use pinctrl, and the pinctrl happens to be the last initialized driver, I first get 30 messages with probe deferred before the pinctrl driver finally gets initialized. IMHO, this is a waste of resources... Giving the system some hint what we need early (e.g. pinctrl or even GPIO drivers) is just sensible. But maybe I miss something here...
Fact is, currently, without touching GPIO infrastructure code, the GPIO driver can not be deferred when the pinctrl driver is missing. The of_gpiochip_add_pin_range function does not handle the missing pinctrl driver. Hence as of now, the pinctrl still needs an earlier initcall. But anyway, currently the pinctrl driver is already using arch_initcall, this patch set is no longer touching that.
Quoted 10 lines
Quoted 9 lines
Most GPIO driver do this, some statistic again:
$ grep -h -o ".*_initcall" drivers/gpio/*.c | sort | uniq -c | sort -n -r 33 subsys_initcall 14 postcore_initcall 2 device_initcall 2 arch_initcall 1 late_initcall 1 core_initcall
Quoted 8 lines
Yeah old legacy. There are patch attacks to get rid of this.
The reason we can't just change them is because sometimes dependent drivers do not handle the errorpath very well can can't defer cleanly.
With a new driver I expect deferred probe to be used.
Actually, the esdhc driver gets the cd-gpio directly and does not handle EPROBE_DEFER properly, hence Vybrid would be affected too. But I understand that we need to start somewhere. I will change the GPIO driver to device_initcall and going to fix esdhc driver.
--
Stefan