| Message ID | 20240114152759.1040563-2-andriy.shevchenko@linux.intel.com (mailing list archive) |
|---|---|
| State | Handled Elsewhere |
| Headers | show |
| Series | backlight: hx8357: Clean up and make OF-independent | expand |
Andy Shevchenko <andriy.shevchenko@linux.intel.com> writes: Hello Andy, > Convert the module to be property provider agnostic and allow > it to be used on non-OF platforms. > > Include mod_devicetable.h explicitly to replace the dropped of.h > which included mod_devicetable.h indirectly. > > Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> > --- > drivers/video/backlight/hx8357.c | 14 ++++++++------ > 1 file changed, 8 insertions(+), 6 deletions(-) > > diff --git a/drivers/video/backlight/hx8357.c b/drivers/video/backlight/hx8357.c > index bf18337ff0c2..c7fd10d55c5d 100644 > --- a/drivers/video/backlight/hx8357.c > +++ b/drivers/video/backlight/hx8357.c > @@ -8,9 +8,9 @@ > #include <linux/delay.h> > #include <linux/gpio/consumer.h> > #include <linux/lcd.h> > +#include <linux/mod_devicetable.h> > #include <linux/module.h> > -#include <linux/of.h> > -#include <linux/of_device.h> > +#include <linux/property.h> > #include <linux/spi/spi.h> > > #define HX8357_NUM_IM_PINS 3 > @@ -564,6 +564,8 @@ static struct lcd_ops hx8357_ops = { > .get_power = hx8357_get_power, > }; > > +typedef int (*hx8357_init)(struct lcd_device *); > + This kind of typedef usage is frowned upon in the Linux coding style [0] (per my understanding at least) and indeed in my opinion it makes harder to grep. [0] https://www.kernel.org/doc/Documentation/process/coding-style.rst > static const struct of_device_id hx8357_dt_ids[] = { > { > .compatible = "himax,hx8357", > @@ -582,7 +584,7 @@ static int hx8357_probe(struct spi_device *spi) > struct device *dev = &spi->dev; > struct lcd_device *lcdev; > struct hx8357_data *lcd; > - const struct of_device_id *match; > + hx8357_init init; > int i, ret; > > lcd = devm_kzalloc(&spi->dev, sizeof(*lcd), GFP_KERNEL); > @@ -597,8 +599,8 @@ static int hx8357_probe(struct spi_device *spi) > > lcd->spi = spi; > > - match = of_match_device(hx8357_dt_ids, &spi->dev); > - if (!match || !match->data) > + init = device_get_match_data(dev); > + if (!init) > return -EINVAL; > > lcd->reset = devm_gpiod_get(dev, "reset", GPIOD_OUT_LOW); > @@ -627,7 +629,7 @@ static int hx8357_probe(struct spi_device *spi) > > hx8357_lcd_reset(lcdev); > > - ret = ((int (*)(struct lcd_device *))match->data)(lcdev); This is what I mean, before it was clear what was stored in match->data. But after you changes, what is returned by the device_get_match_data() function is opaque and you need to look at the typedef hx8357_init to figure that out. No strong opinion though and I see other drivers doing the same (but no other driver in drivers/video/backlight). Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
On Mon, Jan 15, 2024 at 09:20:46AM +0100, Javier Martinez Canillas wrote: > Andy Shevchenko <andriy.shevchenko@linux.intel.com> writes: ... > > +typedef int (*hx8357_init)(struct lcd_device *); > > This kind of typedef usage is frowned upon in the Linux coding style [0] > (per my understanding at least) and indeed in my opinion it makes harder > to grep. > > [0] https://www.kernel.org/doc/Documentation/process/coding-style.rst Thanks for pointing this out. However, this piece does _not_ clarify typedef:s for function pointers which I personally find a good to have. ... > > - ret = ((int (*)(struct lcd_device *))match->data)(lcdev); > > This is what I mean, before it was clear what was stored in match->data. > But after you changes, what is returned by the device_get_match_data() > function is opaque and you need to look at the typedef hx8357_init to > figure that out. The above is so ugly in my opinion, that justifies using typedef:s even if you are quite skeptical about them. ... > No strong opinion though and I see other drivers doing the same (but no > other driver in drivers/video/backlight). > > Reviewed-by: Javier Martinez Canillas <javierm@redhat.com> Thank you for the review!
On Sun, Jan 21, 2024 at 03:48:05PM +0200, Andy Shevchenko wrote: > On Mon, Jan 15, 2024 at 09:20:46AM +0100, Javier Martinez Canillas wrote: > > Andy Shevchenko <andriy.shevchenko@linux.intel.com> writes: > > ... > > > > +typedef int (*hx8357_init)(struct lcd_device *); > > > > This kind of typedef usage is frowned upon in the Linux coding style [0] > > (per my understanding at least) and indeed in my opinion it makes harder > > to grep. > > > > [0] https://www.kernel.org/doc/Documentation/process/coding-style.rst > > Thanks for pointing this out. However, this piece does _not_ clarify typedef:s > for function pointers which I personally find a good to have. > > ... > > > > - ret = ((int (*)(struct lcd_device *))match->data)(lcdev); > > > > This is what I mean, before it was clear what was stored in match->data. > > But after you changes, what is returned by the device_get_match_data() > > function is opaque and you need to look at the typedef hx8357_init to > > figure that out. > > The above is so ugly in my opinion, that justifies using typedef:s even > if you are quite skeptical about them. FWIW I was pretty skeptical about it to. Largely because the three touchs (typedef, variable initialization, use) spread things around a bit too much. Can we at least name the type to make it obvious that it is a function pointer? Something like hx8357_init_fn . Daniel.
On Sun, Jan 14, 2024 at 05:25:08PM +0200, Andy Shevchenko wrote: > Convert the module to be property provider agnostic and allow > it to be used on non-OF platforms. > > Include mod_devicetable.h explicitly to replace the dropped of.h > which included mod_devicetable.h indirectly. > > Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> > --- > drivers/video/backlight/hx8357.c | 14 ++++++++------ > 1 file changed, 8 insertions(+), 6 deletions(-) > > diff --git a/drivers/video/backlight/hx8357.c b/drivers/video/backlight/hx8357.c > index bf18337ff0c2..c7fd10d55c5d 100644 > --- a/drivers/video/backlight/hx8357.c > +++ b/drivers/video/backlight/hx8357.c > @@ -8,9 +8,9 @@ > #include <linux/delay.h> > #include <linux/gpio/consumer.h> > #include <linux/lcd.h> > +#include <linux/mod_devicetable.h> > #include <linux/module.h> > -#include <linux/of.h> > -#include <linux/of_device.h> > +#include <linux/property.h> > #include <linux/spi/spi.h> > > #define HX8357_NUM_IM_PINS 3 > @@ -564,6 +564,8 @@ static struct lcd_ops hx8357_ops = { > .get_power = hx8357_get_power, > }; > > +typedef int (*hx8357_init)(struct lcd_device *); > + > static const struct of_device_id hx8357_dt_ids[] = { > { > .compatible = "himax,hx8357", > @@ -582,7 +584,7 @@ static int hx8357_probe(struct spi_device *spi) > struct device *dev = &spi->dev; > struct lcd_device *lcdev; > struct hx8357_data *lcd; > - const struct of_device_id *match; > + hx8357_init init; As somewhere else in this thread, I'd find this a lot more readable as: hx8357_init_fn init_fn; Other than that, LGTM. Daniel.
On Wed, Jan 24, 2024 at 05:19:53PM +0000, Daniel Thompson wrote: > On Sun, Jan 14, 2024 at 05:25:08PM +0200, Andy Shevchenko wrote: ... > > +typedef int (*hx8357_init)(struct lcd_device *); > > + hx8357_init init; > > As somewhere else in this thread, I'd find this a lot more readable > as: > hx8357_init_fn init_fn; I agree with you, dunno why I haven't added _fn in the first place...
diff --git a/drivers/video/backlight/hx8357.c b/drivers/video/backlight/hx8357.c index bf18337ff0c2..c7fd10d55c5d 100644 --- a/drivers/video/backlight/hx8357.c +++ b/drivers/video/backlight/hx8357.c @@ -8,9 +8,9 @@ #include <linux/delay.h> #include <linux/gpio/consumer.h> #include <linux/lcd.h> +#include <linux/mod_devicetable.h> #include <linux/module.h> -#include <linux/of.h> -#include <linux/of_device.h> +#include <linux/property.h> #include <linux/spi/spi.h> #define HX8357_NUM_IM_PINS 3 @@ -564,6 +564,8 @@ static struct lcd_ops hx8357_ops = { .get_power = hx8357_get_power, }; +typedef int (*hx8357_init)(struct lcd_device *); + static const struct of_device_id hx8357_dt_ids[] = { { .compatible = "himax,hx8357", @@ -582,7 +584,7 @@ static int hx8357_probe(struct spi_device *spi) struct device *dev = &spi->dev; struct lcd_device *lcdev; struct hx8357_data *lcd; - const struct of_device_id *match; + hx8357_init init; int i, ret; lcd = devm_kzalloc(&spi->dev, sizeof(*lcd), GFP_KERNEL); @@ -597,8 +599,8 @@ static int hx8357_probe(struct spi_device *spi) lcd->spi = spi; - match = of_match_device(hx8357_dt_ids, &spi->dev); - if (!match || !match->data) + init = device_get_match_data(dev); + if (!init) return -EINVAL; lcd->reset = devm_gpiod_get(dev, "reset", GPIOD_OUT_LOW); @@ -627,7 +629,7 @@ static int hx8357_probe(struct spi_device *spi) hx8357_lcd_reset(lcdev); - ret = ((int (*)(struct lcd_device *))match->data)(lcdev); + ret = init(lcdev); if (ret) { dev_err(&spi->dev, "Couldn't initialize panel\n"); return ret;
Convert the module to be property provider agnostic and allow it to be used on non-OF platforms. Include mod_devicetable.h explicitly to replace the dropped of.h which included mod_devicetable.h indirectly. Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> --- drivers/video/backlight/hx8357.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-)