[PATCH] OMAP2+: mux: Fixed gpio mux mode analysis.

[Tony Lindgren]

* oleg.matcovschi at ti.com <oleg.matcovschi at ti.com> [121115 13:42]:
> From: Oleg Matcovschi <oleg.matcovschi at ti.com>
> 
> OMAP_MODE_GPIO() macro verified only OMAP_MUX_MODE4.

Indeed, that needs to be fixed.

> It is not correct for following platforms:
>     2430 - gpio mux mode 3
>     44xx - gpio mux mode 3

Looks like these are set properly to GPIO_IN_MODE3 with
omap_mux_init().

>     54xx - gpio mux mode 6

This will be only using pinctrl-single.c, so we don't
really have to worrry about this one. But I guess we might
as well fix that too while at it if somebody backports omap5
support to some older kernel..
 
> Patch reserves first 3 bits in partition flags for storing gpio mux
> mode in same format as stored in control pad register.
> Modified OMAP_MODE_GPIO() macro handles all possible cases of gpio mux mode.
> Modified omap_mux_init() flags of omap34xx to include OMAP_MUX_GPIO_IN_MODE4.

Why don't you just add int gpio to struct omap_mux_partition?

You're not saving many bytes as at most we have two partitions
so far per SoC.
 
> --- a/arch/arm/mach-omap2/mux34xx.c
> +++ b/arch/arm/mach-omap2/mux34xx.c
> @@ -2053,7 +2053,7 @@ int __init omap3_mux_init(struct omap_board_mux *board_subset, int flags)
>  		return -EINVAL;
>  	}
>  
> -	return omap_mux_init("core", 0,
> +	return omap_mux_init("core", OMAP_MUX_GPIO_IN_MODE4,
>  			     OMAP3_CONTROL_PADCONF_MUX_PBASE,
>  			     OMAP3_CONTROL_PADCONF_MUX_SIZE,
>  			     omap3_muxmodes, package_subset, board_subset,

The default is GPIO_IN_MODE4, but that seems to be a bad
default choice as only omap3 uses it. So yeah, might as well
initialize that too and not assume any defaults GPIO mode.

Regards,

Tony