Message ID | 1347285112-13542-2-git-send-email-jm@lentin.co.uk (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
On Mon, Sep 10, 2012 at 02:51:51PM +0100, Jamie Lentin wrote: > Allow a gpio-fan to be defined in devicetree, see binding documentation > for details. > > Changes since V1:- > * Don't hide return codes [Guenter Roeck] > * Remove typecast noise [Guenter Roeck] > * Use of_find_property instead of counting u32s [Guenter Roeck] > * Don't count GPIOs twice [Andrew Lunn] > * Use of_prop_next_u32 to get records in a more obvious fashion > * Use CONFIG_OF_GPIO instead of CONFIG_OF > * Apply __devinitdata to of_gpio_fan_match [Andrew Lunn] > > Signed-off-by: Jamie Lentin <jm@lentin.co.uk> > --- > .../devicetree/bindings/gpio/gpio-fan.txt | 25 +++++ > drivers/hwmon/gpio-fan.c | 116 ++++++++++++++++++++ > 2 files changed, 141 insertions(+) > create mode 100644 Documentation/devicetree/bindings/gpio/gpio-fan.txt Hi Jamie Thanks for addressing all the issues raised. However... Anything before the --- ends up in the changelog entry. Anything afterwards does not. The changes you made to address review comments is not something for the changelog, so Changes since V1:- should be after the --- . There is more detail in Documentation/SubmittingPatches Andrew
On Mon, 10 Sep 2012, Andrew Lunn wrote: > On Mon, Sep 10, 2012 at 02:51:51PM +0100, Jamie Lentin wrote: >> Allow a gpio-fan to be defined in devicetree, see binding documentation >> for details. >> >> Changes since V1:- >> * Don't hide return codes [Guenter Roeck] >> * Remove typecast noise [Guenter Roeck] >> * Use of_find_property instead of counting u32s [Guenter Roeck] >> * Don't count GPIOs twice [Andrew Lunn] >> * Use of_prop_next_u32 to get records in a more obvious fashion >> * Use CONFIG_OF_GPIO instead of CONFIG_OF >> * Apply __devinitdata to of_gpio_fan_match [Andrew Lunn] >> >> Signed-off-by: Jamie Lentin <jm@lentin.co.uk> >> --- >> .../devicetree/bindings/gpio/gpio-fan.txt | 25 +++++ >> drivers/hwmon/gpio-fan.c | 116 ++++++++++++++++++++ >> 2 files changed, 141 insertions(+) >> create mode 100644 Documentation/devicetree/bindings/gpio/gpio-fan.txt > > Hi Jamie > > Thanks for addressing all the issues raised. However... > > Anything before the --- ends up in the changelog entry. Anything > afterwards does not. > > The changes you made to address review comments is not something for > the changelog, so Changes since V1:- should be after the --- . Okay, whilst I was submitting the initial DT support it was suggested by Grant Likely they should be above the '---'. I think I prefer this also, but I'll go with whatever is more likely to get patches accepted :) Can resend with the content jiggled if you want, but will put them after the '---' in future. > > There is more detail in Documentation/SubmittingPatches > > Andrew >
On Mon, Sep 10, 2012 at 05:04:26PM +0100, Jamie Lentin wrote: > On Mon, 10 Sep 2012, Andrew Lunn wrote: > > >On Mon, Sep 10, 2012 at 02:51:51PM +0100, Jamie Lentin wrote: > >>Allow a gpio-fan to be defined in devicetree, see binding documentation > >>for details. > >> > >>Changes since V1:- > >>* Don't hide return codes [Guenter Roeck] > >>* Remove typecast noise [Guenter Roeck] > >>* Use of_find_property instead of counting u32s [Guenter Roeck] > >>* Don't count GPIOs twice [Andrew Lunn] > >>* Use of_prop_next_u32 to get records in a more obvious fashion > >>* Use CONFIG_OF_GPIO instead of CONFIG_OF > >>* Apply __devinitdata to of_gpio_fan_match [Andrew Lunn] > >> > >>Signed-off-by: Jamie Lentin <jm@lentin.co.uk> > >>--- > >> .../devicetree/bindings/gpio/gpio-fan.txt | 25 +++++ > >> drivers/hwmon/gpio-fan.c | 116 ++++++++++++++++++++ > >> 2 files changed, 141 insertions(+) > >> create mode 100644 Documentation/devicetree/bindings/gpio/gpio-fan.txt > > > >Hi Jamie > > > >Thanks for addressing all the issues raised. However... > > > >Anything before the --- ends up in the changelog entry. Anything > >afterwards does not. > > > >The changes you made to address review comments is not something for > >the changelog, so Changes since V1:- should be after the --- . > > Okay, whilst I was submitting the initial DT support it was > suggested by Grant Likely they should be above the '---'. I think I > prefer this also, but I'll go with whatever is more likely to get > patches accepted :) > > Can resend with the content jiggled if you want, but will put them > after the '---' in future. There's no need to resend if this is the only issue. I'll just clean it up when I pull it in. In the future, though, please do as Andrew suggested. thx, Jason.
On Mon, Sep 10, 2012 at 02:51:51PM +0100, Jamie Lentin wrote: > Allow a gpio-fan to be defined in devicetree, see binding documentation > for details. > > Changes since V1:- > * Don't hide return codes [Guenter Roeck] > * Remove typecast noise [Guenter Roeck] > * Use of_find_property instead of counting u32s [Guenter Roeck] > * Don't count GPIOs twice [Andrew Lunn] > * Use of_prop_next_u32 to get records in a more obvious fashion > * Use CONFIG_OF_GPIO instead of CONFIG_OF > * Apply __devinitdata to of_gpio_fan_match [Andrew Lunn] > Jamie, As suggested by others, after ---, please > Signed-off-by: Jamie Lentin <jm@lentin.co.uk> > --- > .../devicetree/bindings/gpio/gpio-fan.txt | 25 +++++ > drivers/hwmon/gpio-fan.c | 116 ++++++++++++++++++++ > 2 files changed, 141 insertions(+) > create mode 100644 Documentation/devicetree/bindings/gpio/gpio-fan.txt > > diff --git a/Documentation/devicetree/bindings/gpio/gpio-fan.txt b/Documentation/devicetree/bindings/gpio/gpio-fan.txt > new file mode 100644 > index 0000000..2dd457a > --- /dev/null > +++ b/Documentation/devicetree/bindings/gpio/gpio-fan.txt > @@ -0,0 +1,25 @@ > +Bindings for fan connected to GPIO lines > + > +Required properties: > +- compatible : "gpio-fan" > +- gpios: Specifies the pins that map to bits in the control value, > + ordered MSB-->LSB. > +- gpio-fan,speed-map: A mapping of possible fan RPM speeds and the > + control value that should be set to achieve them. This array > + must have the RPM values in ascending order. > + > +Optional properties: > +- alarm-gpios: This pin going active indicates something is wrong with > + the fan, and a udev event will be fired. > + > +Examples: > + > + gpio_fan { > + compatible = "gpio-fan"; > + gpios = <&gpio1 14 1 > + &gpio1 13 1>; > + gpio-fan,speed-map = <0 0 > + 3000 1 > + 6000 2>; > + alarm-gpios = <&gpio1 15 1>; > + }; > diff --git a/drivers/hwmon/gpio-fan.c b/drivers/hwmon/gpio-fan.c > index 2f4b01b..f0a803b 100644 > --- a/drivers/hwmon/gpio-fan.c > +++ b/drivers/hwmon/gpio-fan.c > @@ -31,6 +31,8 @@ > #include <linux/hwmon.h> > #include <linux/gpio.h> > #include <linux/gpio-fan.h> > +#include <linux/of_platform.h> > +#include <linux/of_gpio.h> > > struct gpio_fan_data { > struct platform_device *pdev; > @@ -400,14 +402,127 @@ static ssize_t show_name(struct device *dev, > > static DEVICE_ATTR(name, S_IRUGO, show_name, NULL); > > + > +#ifdef CONFIG_OF_GPIO > +/* > + * Translate OpenFirmware node properties into platform_data > + */ > +static int gpio_fan_get_of_pdata(struct device *dev, > + struct gpio_fan_platform_data *pdata) > +{ > + struct device_node *node; > + struct gpio_fan_speed *speed; s/ / / > + unsigned *ctrl; > + unsigned i; > + u32 u; > + struct property *prop; > + const __be32 *p; > + > + node = dev->of_node; > + > + /* Fill GPIO pin array */ > + pdata->num_ctrl = of_gpio_count(node); All other callers of this function check if the result is > 0. Thanks, Guenter
diff --git a/Documentation/devicetree/bindings/gpio/gpio-fan.txt b/Documentation/devicetree/bindings/gpio/gpio-fan.txt new file mode 100644 index 0000000..2dd457a --- /dev/null +++ b/Documentation/devicetree/bindings/gpio/gpio-fan.txt @@ -0,0 +1,25 @@ +Bindings for fan connected to GPIO lines + +Required properties: +- compatible : "gpio-fan" +- gpios: Specifies the pins that map to bits in the control value, + ordered MSB-->LSB. +- gpio-fan,speed-map: A mapping of possible fan RPM speeds and the + control value that should be set to achieve them. This array + must have the RPM values in ascending order. + +Optional properties: +- alarm-gpios: This pin going active indicates something is wrong with + the fan, and a udev event will be fired. + +Examples: + + gpio_fan { + compatible = "gpio-fan"; + gpios = <&gpio1 14 1 + &gpio1 13 1>; + gpio-fan,speed-map = <0 0 + 3000 1 + 6000 2>; + alarm-gpios = <&gpio1 15 1>; + }; diff --git a/drivers/hwmon/gpio-fan.c b/drivers/hwmon/gpio-fan.c index 2f4b01b..f0a803b 100644 --- a/drivers/hwmon/gpio-fan.c +++ b/drivers/hwmon/gpio-fan.c @@ -31,6 +31,8 @@ #include <linux/hwmon.h> #include <linux/gpio.h> #include <linux/gpio-fan.h> +#include <linux/of_platform.h> +#include <linux/of_gpio.h> struct gpio_fan_data { struct platform_device *pdev; @@ -400,14 +402,127 @@ static ssize_t show_name(struct device *dev, static DEVICE_ATTR(name, S_IRUGO, show_name, NULL); + +#ifdef CONFIG_OF_GPIO +/* + * Translate OpenFirmware node properties into platform_data + */ +static int gpio_fan_get_of_pdata(struct device *dev, + struct gpio_fan_platform_data *pdata) +{ + struct device_node *node; + struct gpio_fan_speed *speed; + unsigned *ctrl; + unsigned i; + u32 u; + struct property *prop; + const __be32 *p; + + node = dev->of_node; + + /* Fill GPIO pin array */ + pdata->num_ctrl = of_gpio_count(node); + ctrl = devm_kzalloc(dev, pdata->num_ctrl * sizeof(unsigned), + GFP_KERNEL); + if (!ctrl) + return -ENOMEM; + for (i = 0; i < pdata->num_ctrl; i++) { + int val; + + val = of_get_gpio(node, i); + if (val < 0) + return val; + ctrl[i] = val; + } + pdata->ctrl = ctrl; + + /* Get number of RPM/ctrl_val pairs in speed map */ + prop = of_find_property(node, "gpio-fan,speed-map", &i); + if (!prop) { + dev_err(dev, "gpio-fan,speed-map node missing from DT"); + return -ENODEV; + } + i = i / sizeof(u32); + if (i == 0 || i & 1) { + dev_err(dev, "gpio-fan,speed-map contains zero/odd number of entries"); + return -ENODEV; + } + pdata->num_speed = i / 2; + + /* + * Populate speed map + * Speed map is in the form <RPM ctrl_val RPM ctrl_val ...> + * this needs splitting into pairs to create gpio_fan_speed structs + */ + speed = devm_kzalloc(dev, + pdata->num_speed * sizeof(struct gpio_fan_speed), + GFP_KERNEL); + if (!speed) + return -ENOMEM; + p = NULL; + for (i = 0; i < pdata->num_speed; i++) { + p = of_prop_next_u32(prop, p, &u); + if (!p) + return -ENODEV; + speed[i].rpm = u; + p = of_prop_next_u32(prop, p, &u); + if (!p) + return -ENODEV; + speed[i].ctrl_val = u; + } + pdata->speed = speed; + + /* Alarm GPIO if one exists */ + if (of_gpio_named_count(node, "alarm-gpios")) { + struct gpio_fan_alarm *alarm; + int val; + enum of_gpio_flags flags; + + alarm = devm_kzalloc(dev, sizeof(struct gpio_fan_alarm), + GFP_KERNEL); + if (!alarm) + return -ENOMEM; + + val = of_get_named_gpio_flags(node, "alarm-gpios", 0, &flags); + if (val < 0) + return val; + alarm->gpio = val; + alarm->active_low = flags & OF_GPIO_ACTIVE_LOW; + + pdata->alarm = alarm; + } + + return 0; +} + +static struct of_device_id of_gpio_fan_match[] __devinitdata = { + { .compatible = "gpio-fan", }, + {}, +}; +#endif /* CONFIG_OF_GPIO */ + static int __devinit gpio_fan_probe(struct platform_device *pdev) { int err; struct gpio_fan_data *fan_data; struct gpio_fan_platform_data *pdata = pdev->dev.platform_data; +#ifdef CONFIG_OF_GPIO + if (!pdata) { + pdata = devm_kzalloc(&pdev->dev, + sizeof(struct gpio_fan_platform_data), + GFP_KERNEL); + if (!pdata) + return -ENOMEM; + + err = gpio_fan_get_of_pdata(&pdev->dev, pdata); + if (err) + return err; + } +#else /* CONFIG_OF_GPIO */ if (!pdata) return -EINVAL; +#endif /* CONFIG_OF_GPIO */ fan_data = devm_kzalloc(&pdev->dev, sizeof(struct gpio_fan_data), GFP_KERNEL); @@ -511,6 +626,7 @@ static struct platform_driver gpio_fan_driver = { .driver = { .name = "gpio-fan", .pm = GPIO_FAN_PM, + .of_match_table = of_match_ptr(of_gpio_fan_match), }, };
Allow a gpio-fan to be defined in devicetree, see binding documentation for details. Changes since V1:- * Don't hide return codes [Guenter Roeck] * Remove typecast noise [Guenter Roeck] * Use of_find_property instead of counting u32s [Guenter Roeck] * Don't count GPIOs twice [Andrew Lunn] * Use of_prop_next_u32 to get records in a more obvious fashion * Use CONFIG_OF_GPIO instead of CONFIG_OF * Apply __devinitdata to of_gpio_fan_match [Andrew Lunn] Signed-off-by: Jamie Lentin <jm@lentin.co.uk> --- .../devicetree/bindings/gpio/gpio-fan.txt | 25 +++++ drivers/hwmon/gpio-fan.c | 116 ++++++++++++++++++++ 2 files changed, 141 insertions(+) create mode 100644 Documentation/devicetree/bindings/gpio/gpio-fan.txt