Thread (10 messages) 10 messages, 4 authors, 2015-01-16

Re: [Patch v5 2/2] gpio: Document GPIO hogging mechanism

From: Alexandre Courbot <hidden>
Date: 2015-01-12 21:43:56
Also in: linux-gpio, lkml

On Tue, Jan 13, 2015 at 1:39 AM, Benoit Parrot [off-list ref] wrote:
Linus Walleij [off-list ref] wrote on Mon [2015-Jan-12 11:20:14 +0100]:
quoted
On Fri, Dec 19, 2014 at 9:07 PM, Benoit Parrot [off-list ref] wrote:
quoted
Add GPIO hogging documentation to gpio.txt

Signed-off-by: Benoit Parrot <redacted>
Reviewed-by: Alexandre Courbot <acourbot@nvidia.com>
This is starting to look good ...
quoted
+               line_b {
+                       gpio-hog;
+                       gpios = <6 0>;
+                       state = "output-low";
I don't like the state string.

Instead have boolean properties for all states.

line_b {
    gpio-hog;
    gpios = <6 0>;
    output-low;
    line-name = "foo-bar-gpio";
}

Then use of_property_read_bool() in the code to check which
state is to be selected intially. You can check that no mutually
exclusive state are selected, I don't like that an arbitrary string
select the state like that, if we do it that way an enumerator would
be better, I prefer bools.
I am sorry but that is how it was originally in the first patch.
Alexandre's review comment suggested this method in [1] and [2] (below).

Alexandre, any comments?

[1] http://marc.info/?l=linux-gpio&m=141456662426151&w=2

[2] http://marc.info/?l=linux-gpio&m=141715982424744&w=2
When Linus and I are in conflict, follow Linus. Arnd's suggestion of
having enums defined in (IIUC) include/dt-bindings/gpio and using them
sounds good to me too and might make everyone happy (no possibility of
conflicting definitions + no strings). Linus, could you comment on it?
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help