On 26-1-2014 16:04, Charles Steinkuehler wrote:
On 1/26/2014 8:41 AM, Bas Laarhoven wrote:
On 26-1-2014 14:32, Charles Steinkuehler wrote:
Without this patch, any pins exported to sysfs by the gpio-of-helper
driver are missing the /sys/class/gpio/gpio<N>/direction file and are
stuck in either input or output mode.
Comments?
Great! It looks so obvious, this makes me wonder why it wasn't
implemented the first time...!
Yes, but it's so simple I fear I'm missing something. There are
That's what I meant to say. But since modern programmers do not write
comments, we'll always keep wondering whether it was a design decision
or an oversight. :-/
definitely times when you might not want an I/O pin to change direction,
and other times (like my use) where it's OK.
I think maybe I should add a new property to allow the direction to be
changed, so DT stubs would look something like:
x_min {
gpio-name = "bebopr:x_min";
gpio = <&gpio3 3 0>;
input;
dir-changeable;
};
I have limited knowledge of what you're trying to do, but you probably
never want to make gpio3 an output on a BeBoPr because that creates two
outputs fighting the signal!
x_ena {
gpio-name = "bebopr:x_ena";
gpio = <&gpio1 4 0>;
output;
init-high;
dir-changeable;
};
Instead of just assuming the direction should _always_ be changeable, so
there's no chance of braking any existing setups.
Agreed, if possible, keep it backwards compatible.
-- Bas
--
For more options, visit http://beagleboard.org/discuss
---
You received this message because you are subscribed to the Google Groups "BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.
