still unclear on a couple DT questions from my initial post

still unclear on a couple DT questions from my initial post

[Robert P. J. Day]

  again, more nitpicky pedantry, but from my initial post, i'm still
unclear on a couple things.

  first, a number of powerpc .dts files contain the line:

    dcr-parent = <&{/cpus/cpu@0}>;

and it's not clear to me why there are curly braces there. that is the
only example of that, but i see no mention of that in the spec. is it
meaningful? or are they superfluous?

  also, the spec states that status could be set to "okay", but a pile
of .dts files in the current kernel have the shorter form:

  status = "ok";

and the current kernel code in drivers/of explicitly accepts both
forms. is that a formal spec property, or is the kernel code simply
allowing people to be sloppy?

  i think that's it for now, thanks for your patience.

rday

-- 

========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: still unclear on a couple DT questions from my initial post

[Mark Rutland]

On Fri, Sep 01, 2017 at 01:29:31AM -0400, Robert P. J. Day wrote:
>   again, more nitpicky pedantry, but from my initial post, i'm still
> unclear on a couple things.
> 
>   first, a number of powerpc .dts files contain the line:
> 
>     dcr-parent = <&{/cpus/cpu@0}>;
> 
> and it's not clear to me why there are curly braces there. that is the
> only example of that, but i see no mention of that in the spec. is it
> meaningful? or are they superfluous?

This syntax allows us to refer to a node by path, even if it doesn't
have an existing label. The &foo syntax only allows us to refer to an
existing label called foo.

>   also, the spec states that status could be set to "okay", but a pile
> of .dts files in the current kernel have the shorter form:
> 
>   status = "ok";
> 
> and the current kernel code in drivers/of explicitly accepts both
> forms. is that a formal spec property, or is the kernel code simply
> allowing people to be sloppy?

The kernel is catering for sloppiness. We should probably fix those up
in the kernel tree.

Thanks,
Mark.
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: still unclear on a couple DT questions from my initial post

[Robert P. J. Day]

On Fri, 1 Sep 2017, Mark Rutland wrote:

> On Fri, Sep 01, 2017 at 01:29:31AM -0400, Robert P. J. Day wrote:
> >   again, more nitpicky pedantry, but from my initial post, i'm still
> > unclear on a couple things.
> >
> >   first, a number of powerpc .dts files contain the line:
> >
> >     dcr-parent = <&{/cpus/cpu@0}>;
> >
> > and it's not clear to me why there are curly braces there. that is the
> > only example of that, but i see no mention of that in the spec. is it
> > meaningful? or are they superfluous?
>
> This syntax allows us to refer to a node by path, even if it doesn't
> have an existing label. The &foo syntax only allows us to refer to
> an existing label called foo.

  ah ... is that mentioned in the spec somewhere? i don't recall
seeing that.

> >   also, the spec states that status could be set to "okay", but a pile
> > of .dts files in the current kernel have the shorter form:
> >
> >   status = "ok";
> >
> > and the current kernel code in drivers/of explicitly accepts both
> > forms. is that a formal spec property, or is the kernel code simply
> > allowing people to be sloppy?
>
> The kernel is catering for sloppiness. We should probably fix those up
> in the kernel tree.

  the vast majority of that is under arch/arm*, if you want to take
care of that, and i can see about the rest.

rday

-- 

========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: still unclear on a couple DT questions from my initial post

[Rob Herring]

On Fri, Sep 1, 2017 at 12:14 PM, Robert P. J. Day <rpjday-L09J2beyid0N/H6P543EQg@public.gmane.org> wrote:
> On Fri, 1 Sep 2017, Mark Rutland wrote:
>
>> On Fri, Sep 01, 2017 at 01:29:31AM -0400, Robert P. J. Day wrote:
>> >   again, more nitpicky pedantry, but from my initial post, i'm still
>> > unclear on a couple things.
>> >
>> >   first, a number of powerpc .dts files contain the line:
>> >
>> >     dcr-parent = <&{/cpus/cpu@0}>;
>> >
>> > and it's not clear to me why there are curly braces there. that is the
>> > only example of that, but i see no mention of that in the spec. is it
>> > meaningful? or are they superfluous?
>>
>> This syntax allows us to refer to a node by path, even if it doesn't
>> have an existing label. The &foo syntax only allows us to refer to
>> an existing label called foo.
>
>   ah ... is that mentioned in the spec somewhere? i don't recall
> seeing that.
>
>> >   also, the spec states that status could be set to "okay", but a pile
>> > of .dts files in the current kernel have the shorter form:
>> >
>> >   status = "ok";
>> >
>> > and the current kernel code in drivers/of explicitly accepts both
>> > forms. is that a formal spec property, or is the kernel code simply
>> > allowing people to be sloppy?
>>
>> The kernel is catering for sloppiness. We should probably fix those up
>> in the kernel tree.
>
>   the vast majority of that is under arch/arm*, if you want to take
> care of that, and i can see about the rest.

My preference would be to first add a check into dtc. It has a check
for status currently, but only that it is a string. Then others can
get annoyed with the warning and fix.

Rob
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html