| Message ID | 20210307220549.354263-1-andy.shevchenko@gmail.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | [v1,1/2] Input: tsc2007 - convert to GPIO descriptors | expand |
Hi Andy, On Mon, Mar 08, 2021 at 12:05:48AM +0200, Andy Shevchenko wrote: > @@ -226,11 +226,12 @@ static int tsc2007_get_pendown_state_gpio(struct device *dev) > struct i2c_client *client = to_i2c_client(dev); > struct tsc2007 *ts = i2c_get_clientdata(client); > > - return !gpio_get_value(ts->gpio); > + return !gpiod_get_value(ts->gpiod); This is not correct. gpio_get_value() is raw polarity vs gpiod_get_value() using logical active/inactive, and tsc2007 GPIO lines are active low. The negation must be dropped after switching to GPIOD API. > } > > static int tsc2007_probe_dt(struct i2c_client *client, struct tsc2007 *ts) > { > + struct device *dev = &client->dev; > struct device_node *np = client->dev.of_node; > u32 val32; > u64 val64; > @@ -266,13 +267,11 @@ static int tsc2007_probe_dt(struct i2c_client *client, struct tsc2007 *ts) > return -EINVAL; > } > > - ts->gpio = of_get_gpio(np, 0); > - if (gpio_is_valid(ts->gpio)) > - ts->get_pendown_state = tsc2007_get_pendown_state_gpio; > - else > - dev_warn(&client->dev, > - "GPIO not specified in DT (of_get_gpio returned %d)\n", > - ts->gpio); > + ts->gpiod = devm_gpiod_get_optional(dev, NULL, GPIOD_IN); GPIO is definitely not optional in DT case, at least in the way the driver written right now. > + if (IS_ERR(ts->gpiod)) > + return PTR_ERR(ts->gpiod); > + > + ts->get_pendown_state = tsc2007_get_pendown_state_gpio; > > return 0; > } > -- > 2.30.1 > Thanks.
On Mon, Mar 8, 2021 at 12:57 AM Dmitry Torokhov <dmitry.torokhov@gmail.com> wrote: > On Mon, Mar 08, 2021 at 12:05:48AM +0200, Andy Shevchenko wrote: ... > > - return !gpio_get_value(ts->gpio); > > + return !gpiod_get_value(ts->gpiod); > > This is not correct. gpio_get_value() is raw polarity vs > gpiod_get_value() using logical active/inactive, and tsc2007 GPIO lines > are active low. The negation must be dropped after switching to GPIOD > API. Ah, indeed, I missed that, thanks! ... > > - ts->gpio = of_get_gpio(np, 0); > > - if (gpio_is_valid(ts->gpio)) > > - ts->get_pendown_state = tsc2007_get_pendown_state_gpio; > > - else > > - dev_warn(&client->dev, > > - "GPIO not specified in DT (of_get_gpio returned %d)\n", > > - ts->gpio); > > + ts->gpiod = devm_gpiod_get_optional(dev, NULL, GPIOD_IN); > > GPIO is definitely not optional in DT case, at least in the way the > driver written right now. Can you elaborate this, please? I don't see from the dev_warn() w/o any error code returned that it's mandatory. In the bindings one may read: Optional properties: - gpios: the interrupt gpio the chip is connected to (trough the penirq pin). The penirq pin goes to low when the panel is touched. (see GPIO binding[1] for more details). Nothing suggested it's mandatory. What have I missed? > > + if (IS_ERR(ts->gpiod)) > > + return PTR_ERR(ts->gpiod);
On Mon, Mar 08, 2021 at 11:10:38AM +0200, Andy Shevchenko wrote: > On Mon, Mar 8, 2021 at 12:57 AM Dmitry Torokhov > <dmitry.torokhov@gmail.com> wrote: > > On Mon, Mar 08, 2021 at 12:05:48AM +0200, Andy Shevchenko wrote: > > ... > > > > - return !gpio_get_value(ts->gpio); > > > + return !gpiod_get_value(ts->gpiod); > > > > This is not correct. gpio_get_value() is raw polarity vs > > gpiod_get_value() using logical active/inactive, and tsc2007 GPIO lines > > are active low. The negation must be dropped after switching to GPIOD > > API. > > Ah, indeed, I missed that, thanks! > > ... > > > > - ts->gpio = of_get_gpio(np, 0); > > > - if (gpio_is_valid(ts->gpio)) > > > - ts->get_pendown_state = tsc2007_get_pendown_state_gpio; > > > - else > > > - dev_warn(&client->dev, > > > - "GPIO not specified in DT (of_get_gpio returned %d)\n", > > > - ts->gpio); > > > + ts->gpiod = devm_gpiod_get_optional(dev, NULL, GPIOD_IN); > > > > GPIO is definitely not optional in DT case, at least in the way the > > driver written right now. > > Can you elaborate this, please? I don't see from the dev_warn() w/o > any error code returned that it's mandatory. > In the bindings one may read: > > Optional properties: > - gpios: the interrupt gpio the chip is connected to (trough the penirq pin). > The penirq pin goes to low when the panel is touched. > (see GPIO binding[1] for more details). > > Nothing suggested it's mandatory. What have I missed? Ah, indeed, I misread the code and thought we'd abort if there is no pendown GPIO. I wonder if we should remove the warning since we seem to support operations without it. Thanks.
On Mon, Mar 8, 2021 at 9:29 PM Dmitry Torokhov <dmitry.torokhov@gmail.com> wrote: > > On Mon, Mar 08, 2021 at 11:10:38AM +0200, Andy Shevchenko wrote: > > On Mon, Mar 8, 2021 at 12:57 AM Dmitry Torokhov > > <dmitry.torokhov@gmail.com> wrote: > > > On Mon, Mar 08, 2021 at 12:05:48AM +0200, Andy Shevchenko wrote: > > > > ... > > > > > > - return !gpio_get_value(ts->gpio); > > > > + return !gpiod_get_value(ts->gpiod); > > > > > > This is not correct. gpio_get_value() is raw polarity vs > > > gpiod_get_value() using logical active/inactive, and tsc2007 GPIO lines > > > are active low. The negation must be dropped after switching to GPIOD > > > API. > > > > Ah, indeed, I missed that, thanks! > > > > ... > > > > > > - ts->gpio = of_get_gpio(np, 0); > > > > - if (gpio_is_valid(ts->gpio)) > > > > - ts->get_pendown_state = tsc2007_get_pendown_state_gpio; > > > > - else > > > > - dev_warn(&client->dev, > > > > - "GPIO not specified in DT (of_get_gpio returned %d)\n", > > > > - ts->gpio); > > > > + ts->gpiod = devm_gpiod_get_optional(dev, NULL, GPIOD_IN); > > > > > > GPIO is definitely not optional in DT case, at least in the way the > > > driver written right now. > > > > Can you elaborate this, please? I don't see from the dev_warn() w/o > > any error code returned that it's mandatory. > > In the bindings one may read: > > > > Optional properties: > > - gpios: the interrupt gpio the chip is connected to (trough the penirq pin). > > The penirq pin goes to low when the panel is touched. > > (see GPIO binding[1] for more details). > > > > Nothing suggested it's mandatory. What have I missed? > > Ah, indeed, I misread the code and thought we'd abort if there is no > pendown GPIO. I wonder if we should remove the warning since we seem to > support operations without it. But that's what I have done, i.e. removed the warning as well. So, if there are no other concerns than inverted value, I'll send v2.
On Mon, Mar 08, 2021 at 11:04:59PM +0200, Andy Shevchenko wrote: > On Mon, Mar 8, 2021 at 9:29 PM Dmitry Torokhov > <dmitry.torokhov@gmail.com> wrote: > > > > On Mon, Mar 08, 2021 at 11:10:38AM +0200, Andy Shevchenko wrote: > > > On Mon, Mar 8, 2021 at 12:57 AM Dmitry Torokhov > > > <dmitry.torokhov@gmail.com> wrote: > > > > On Mon, Mar 08, 2021 at 12:05:48AM +0200, Andy Shevchenko wrote: > > > > > > ... > > > > > > > > - return !gpio_get_value(ts->gpio); > > > > > + return !gpiod_get_value(ts->gpiod); > > > > > > > > This is not correct. gpio_get_value() is raw polarity vs > > > > gpiod_get_value() using logical active/inactive, and tsc2007 GPIO lines > > > > are active low. The negation must be dropped after switching to GPIOD > > > > API. > > > > > > Ah, indeed, I missed that, thanks! > > > > > > ... > > > > > > > > - ts->gpio = of_get_gpio(np, 0); > > > > > - if (gpio_is_valid(ts->gpio)) > > > > > - ts->get_pendown_state = tsc2007_get_pendown_state_gpio; > > > > > - else > > > > > - dev_warn(&client->dev, > > > > > - "GPIO not specified in DT (of_get_gpio returned %d)\n", > > > > > - ts->gpio); > > > > > + ts->gpiod = devm_gpiod_get_optional(dev, NULL, GPIOD_IN); > > > > > > > > GPIO is definitely not optional in DT case, at least in the way the > > > > driver written right now. > > > > > > Can you elaborate this, please? I don't see from the dev_warn() w/o > > > any error code returned that it's mandatory. > > > In the bindings one may read: > > > > > > Optional properties: > > > - gpios: the interrupt gpio the chip is connected to (trough the penirq pin). > > > The penirq pin goes to low when the panel is touched. > > > (see GPIO binding[1] for more details). > > > > > > Nothing suggested it's mandatory. What have I missed? > > > > Ah, indeed, I misread the code and thought we'd abort if there is no > > pendown GPIO. I wonder if we should remove the warning since we seem to > > support operations without it. > > But that's what I have done, i.e. removed the warning as well. Well, what can I say, -ENOCOFFEE. > > So, if there are no other concerns than inverted value, I'll send v2. Yes, please.
diff --git a/drivers/input/touchscreen/tsc2007.h b/drivers/input/touchscreen/tsc2007.h index 91c60bf6dcaf..69b08dd6c8df 100644 --- a/drivers/input/touchscreen/tsc2007.h +++ b/drivers/input/touchscreen/tsc2007.h @@ -19,6 +19,8 @@ #ifndef _TSC2007_H #define _TSC2007_H +struct gpio_desc; + #define TSC2007_MEASURE_TEMP0 (0x0 << 4) #define TSC2007_MEASURE_AUX (0x2 << 4) #define TSC2007_MEASURE_TEMP1 (0x4 << 4) @@ -69,7 +71,7 @@ struct tsc2007 { int fuzzy; int fuzzz; - unsigned int gpio; + struct gpio_desc *gpiod; int irq; wait_queue_head_t wait; diff --git a/drivers/input/touchscreen/tsc2007_core.c b/drivers/input/touchscreen/tsc2007_core.c index 3b80abfc1eca..9b747c7d11f1 100644 --- a/drivers/input/touchscreen/tsc2007_core.c +++ b/drivers/input/touchscreen/tsc2007_core.c @@ -19,11 +19,11 @@ #include <linux/module.h> #include <linux/slab.h> +#include <linux/gpio/consumer.h> #include <linux/input.h> #include <linux/interrupt.h> #include <linux/i2c.h> #include <linux/of_device.h> -#include <linux/of_gpio.h> #include <linux/platform_data/tsc2007.h> #include "tsc2007.h" @@ -226,11 +226,12 @@ static int tsc2007_get_pendown_state_gpio(struct device *dev) struct i2c_client *client = to_i2c_client(dev); struct tsc2007 *ts = i2c_get_clientdata(client); - return !gpio_get_value(ts->gpio); + return !gpiod_get_value(ts->gpiod); } static int tsc2007_probe_dt(struct i2c_client *client, struct tsc2007 *ts) { + struct device *dev = &client->dev; struct device_node *np = client->dev.of_node; u32 val32; u64 val64; @@ -266,13 +267,11 @@ static int tsc2007_probe_dt(struct i2c_client *client, struct tsc2007 *ts) return -EINVAL; } - ts->gpio = of_get_gpio(np, 0); - if (gpio_is_valid(ts->gpio)) - ts->get_pendown_state = tsc2007_get_pendown_state_gpio; - else - dev_warn(&client->dev, - "GPIO not specified in DT (of_get_gpio returned %d)\n", - ts->gpio); + ts->gpiod = devm_gpiod_get_optional(dev, NULL, GPIOD_IN); + if (IS_ERR(ts->gpiod)) + return PTR_ERR(ts->gpiod); + + ts->get_pendown_state = tsc2007_get_pendown_state_gpio; return 0; }
This converts the driver to use GPIO descriptors. Signed-off-by: Andy Shevchenko <andy.shevchenko@gmail.com> --- drivers/input/touchscreen/tsc2007.h | 4 +++- drivers/input/touchscreen/tsc2007_core.c | 17 ++++++++--------- 2 files changed, 11 insertions(+), 10 deletions(-)