remove unused header gpio-starfive-vic.h

[pdp7]

The commit that adds the starfive-jh7100 gpio driver:

ad78ee0

still has the header file from the old VIC version before my cleanup and
renaming. No header file is needed anymore so let's get rid of it.

Emil: please fixup the this commit so we jush have a single GPIO driver
commit.

Signed-off-by: Drew Fustini drew@beagleboard.org

[esmil]

@pdp7 I'll gladly fix up the gpio commit next time I rebase, but this header certainly isn't unused:

esmil@stitch~/linux$ git grep gpio-starfive-vic.h
drivers/i2c/busses/i2c-designware-master.c:#include <linux/gpio-starfive-vic.h>
drivers/i2c/busses/i2c-designware-platdrv.c:#include <linux/gpio-starfive-vic.h>
drivers/media/platform/starfive/imx219_mipi.c:#include <linux/gpio-starfive-vic.h>
drivers/media/platform/starfive/ov5640_dvp.c:#include <linux/gpio-starfive-vic.h>
drivers/media/platform/starfive/sc2235.c:#include <linux/gpio-starfive-vic.h>
drivers/video/fbdev/starfive/adv7513.c:#include <linux/gpio-starfive-vic.h>
drivers/video/fbdev/starfive/seeed5inch.c:#include <linux/gpio-starfive-vic.h>

I guess I'll just make it a separate commit.

[pdp7]

@esmil yes, I just noticed that to after running a full build.

It turns out the header is not actually used in the code for:

drivers/i2c/busses/i2c-designware-platdrv.c
drivers/media/platform/starfive/imx219_mipi.c
drivers/media/platform/starfive/sc2235.c
drivers/video/fbdev/starfive/adv7513.c

It is used in drivers/i2c/busses/i2c-designware-master.c code:

        if((dev->scl_gpio > 0) && (dev->sda_gpio > 0)) {
                SET_GPIO_dout_LOW(dev->scl_gpio);
                SET_GPIO_dout_LOW(dev->sda_gpio);
                SET_GPIO_doen_reverse_(dev->scl_gpio,1);
                SET_GPIO_doen_reverse_(dev->sda_gpio,1);             
                switch(dev->adapter.nr) {
                case 0:
                        SET_GPIO_doen_i2c0_pad_sck_oe(dev->scl_gpio);
                        SET_GPIO_doen_i2c0_pad_sda_oe(dev->sda_gpio);
                        SET_GPIO_i2c0_pad_sck_in(dev->scl_gpio);
                        SET_GPIO_i2c0_pad_sda_in(dev->sda_gpio);
                        break;
                case 1:
                        SET_GPIO_doen_i2c1_pad_sck_oe(dev->scl_gpio);
                        SET_GPIO_doen_i2c1_pad_sda_oe(dev->sda_gpio);
                        SET_GPIO_i2c1_pad_sck_in(dev->scl_gpio);
                        SET_GPIO_i2c1_pad_sda_in(dev->sda_gpio);
                        break;                                         
                case 2:
                        SET_GPIO_doen_i2c2_pad_sck_oe(dev->scl_gpio);  
                        SET_GPIO_doen_i2c2_pad_sda_oe(dev->sda_gpio);
                        SET_GPIO_i2c2_pad_sck_in(dev->scl_gpio);
                        SET_GPIO_i2c2_pad_sda_in(dev->sda_gpio);
                        break;
                case 3:
                        SET_GPIO_doen_i2c3_pad_sck_oe(dev->scl_gpio);
                        SET_GPIO_doen_i2c3_pad_sda_oe(dev->sda_gpio);
                        SET_GPIO_i2c3_pad_sck_in(dev->scl_gpio);
                        SET_GPIO_i2c3_pad_sda_in(dev->sda_gpio);
                        break;
                default:

this looks to me like it is just code copied from bootloader. I think it is should be possible to convert it to correct Linux code but for now, yes, let's just have the header file in a separate commit.

[esmil]

Yeah, I think the real solution is to not even have this code in the driver, but have a proper pinctrl driver and specify these settings in the device tree.

[pdp7]

@esmil yes, I going to work on the pinctrl driver after getting initial feedback on the gpio driver. I want to discuss the FMUX with Linus W. to get his thoughts on best way to implement pinctrl.

I think we can just carry gpio-starfive-vic.h header and hack in drivers/i2c/busses/i2c-designware-master.c as downstream patches until I can get pinctrl to the point where it can replace that hacky code in i2c-designware-master.c with the correct pinctrl and gpio function calls.

[pdp7]

I am restructuring this into a gpio cleanup branch that I will push as a new PR so I am closing this one.