Arun MURTHY wrote: >>> mips-jz4740: pwm: Align with new pwm core driver >>> >>> PWM core driver has been added and has been enabled only for ARM >>> platform. The same can be utilised for mips also. >>> Please align with the pwm core driver(drivers/pwm-core.c). >> >> Is there any reason for artificially limiting it to ARM? > > No not at all, right now I have aligned all existing pwm drivers in ARM to make use of the pwm core driver. > But faced difficulty in aligning the mips-jz4740 pwm driver, without having much knowledge about the device/data sheet. > Hence I have let it to the maintainer of that driver to align and thereafter this limitation will be removed. > Have also comments the same as TODO in the driver. > Ok, I'll take care of adjusting the jz4740 pwm driver once the pwm-core is in proper shape. But I still think it would be better to have a config symbol which would be selected by SoC code and on which PWM_CORE would depend. Then it would be possible for SoC implementation to device whether it wants to provide it's own PWM API implementation or use pwm-core. >>> diff --git a/drivers/pwm/Kconfig b/drivers/pwm/Kconfig >>> index 5d10106..a88640c 100644 >>> --- a/drivers/pwm/Kconfig >>> +++ b/drivers/pwm/Kconfig >>> @@ -4,6 +4,7 @@ >>> >>> menuconfig PWM_DEVICES >>> bool "PWM devices" >>> + depends on ARM >>> default y >>> ---help--- >>> Say Y here to get to see options for device drivers from >> various >>> diff --git a/drivers/pwm/pwm-core.c b/drivers/pwm/pwm-core.c >>> index b84027a..3a0d426 100644 >> >> >> Why can't these changes be in the initial patch which adds pwm-core? >> > Since by default this driver is enabled, and if there is some other pwm driver enabled, both happen to export the same function(pwm_enable/pwm_disable,..) After applying the first patch build may fail. > I would understand that if you were just moving code around, but the pwm_device struct looks completly different now. > Thanks and Regards, > Arun R Murthy > ------------- >
