Pinmux bindings proposal V2

Shawn Guo shawn.guo at linaro.org
Sun Jan 29 21:31:50 EST 2012


On Fri, Jan 27, 2012 at 09:29:50AM -0800, Stephen Warren wrote:
> Simon Glass wrote at Thursday, January 26, 2012 10:42 AM:
...
> > 1. It doesn't seem to make full use of the device tree format. For example,
> > 
> >    <TEGRA_PMX_PG_DTD TEGRA_PMX_CONF_DRIVE_STRENGTH 5>
> > 
> > would be better as something like
> > 
> >     drive-strength = <5>;
> > 
> > if we could arrange it. It also reduces the need for these
> > TEGRA_PMX_CONF_DRIVE_STRENGTH defines.
> 
> Yes I can see the argument this is more readable.
> 
> However, it:
> 
> * Requires a lot of string handling when parsing the device tree, since
>   you have to search for lots of individual properties by name.
> 
> * Bloats the device tree quite a bit due to representing each parameter
>   as a separate property, with a longish name, rather than a single u32
>   cell in the config property I proposed.
> 
It bloats device tree more due to the proposal needs to represent every
single muxable entity (pin for imx case) as a node to accommodate the
properties like 'drive-strength' here.

-- 
Regards,
Shawn



More information about the linux-arm-kernel mailing list