puzzled by online explanation as to how to "overwrite" DT node

puzzled by online explanation as to how to "overwrite" DT node

[Robert P. J. Day]

  (WARNING: sort of newbie alert, but you know that, right? :-)

  was perusing online tutorial on customizing device trees here:

    http://developer.toradex.com/device-tree-customization

and got a bit confused about the part describing how to (as i read it)
"overwrite" entire nodes.

  first, part on "overwriting properties" makes perfect sense --
here's an example in that piece about overwriting default USB mode:

  &usb0 {
      dr_mode = "host";
  };

so, as i've always understood it, the above will do one of two things:

  1) if that property is not set in the node referenced by that label,
     it will be *added* to what is already there, or

  2) if that property is already set (to whatever), it will be
     overwritten

so far, so good?

  as a special example of that right below that, there's a short
paragraph called "Activating/Deactivating Devices" that uses this
example:

  &uart4 {
      status = "okay";
  };

but, really, that's just a special case of the first example so
there's nothing new here. but it's the next section called
"Overwriting nodes" that seems confusing:

  "Whole nodes can be overwritten by simply redefining them. Like
  overwriting properties, latter definitions overwrite earlier
  definitions."

wait, what? entire nodes can be overwritten by redefining them? the
example given looks like this:

  "E.g. to overwrite the pin configuration of Vybrids UART2 (UART_B)
  overwrite the uart2grp node by simply redefining it in your device
  tree ..."

with the device tree content

  &iomuxc {
      vf610-colibri {
          pinctrl_uart2: uart2grp {
              fsl,pins = <
                  VF610_PAD_PTD0__UART2_TX        0x21a2
                  VF610_PAD_PTD1__UART2_RX        0x21a1
              >;
          };
          ...
      };
  };

hang on ... isn't "vf610-colibri" also a "node" in this device tree?
does the above mean that the *entire* vf610-colibri node is being
overwritten? and that i would have to supply the entire new content of
that node, including simply duplicating the rest of the content that
is already there that i want to keep? (is that what the "..."
represents?)

  that seems bizarre, but it's what that wording seems to suggest. can
someone clarify what would happen in the above, and give perhaps a
link to the part of the DTspec that covers this? i'm sure i must be
misreading this.

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: puzzled by online explanation as to how to "overwrite" DT node

[Rob Herring]

On Wed, Oct 18, 2017 at 11:40 AM, Robert P. J. Day
<rpjday-L09J2beyid0N/H6P543EQg@public.gmane.org> wrote:
>
>   (WARNING: sort of newbie alert, but you know that, right? :-)
>
>   was perusing online tutorial on customizing device trees here:
>
>     http://developer.toradex.com/device-tree-customization

Never seen that one.

> and got a bit confused about the part describing how to (as i read it)
> "overwrite" entire nodes.
>
>   first, part on "overwriting properties" makes perfect sense --
> here's an example in that piece about overwriting default USB mode:
>
>   &usb0 {
>       dr_mode = "host";
>   };
>
> so, as i've always understood it, the above will do one of two things:
>
>   1) if that property is not set in the node referenced by that label,
>      it will be *added* to what is already there, or
>
>   2) if that property is already set (to whatever), it will be
>      overwritten
>
> so far, so good?

Yes.

>   as a special example of that right below that, there's a short
> paragraph called "Activating/Deactivating Devices" that uses this
> example:
>
>   &uart4 {
>       status = "okay";
>   };
>
> but, really, that's just a special case of the first example so
> there's nothing new here. but it's the next section called
> "Overwriting nodes" that seems confusing:
>
>   "Whole nodes can be overwritten by simply redefining them. Like
>   overwriting properties, latter definitions overwrite earlier
>   definitions."

I don't think that's right. I think you have to use /delete-node/ first.

> wait, what? entire nodes can be overwritten by redefining them? the
> example given looks like this:
>
>   "E.g. to overwrite the pin configuration of Vybrids UART2 (UART_B)
>   overwrite the uart2grp node by simply redefining it in your device
>   tree ..."
>
> with the device tree content
>
>   &iomuxc {
>       vf610-colibri {
>           pinctrl_uart2: uart2grp {
>               fsl,pins = <
>                   VF610_PAD_PTD0__UART2_TX        0x21a2
>                   VF610_PAD_PTD1__UART2_RX        0x21a1
>               >;
>           };
>           ...
>       };
>   };
>
> hang on ... isn't "vf610-colibri" also a "node" in this device tree?
> does the above mean that the *entire* vf610-colibri node is being
> overwritten? and that i would have to supply the entire new content of
> that node, including simply duplicating the rest of the content that
> is already there that i want to keep? (is that what the "..."
> represents?)

No, you should get a merged vf610-colibri node with the original
contents and any new or overwritten properties from this chunk.

IOW, only properties are overwritten.

>   that seems bizarre, but it's what that wording seems to suggest. can
> someone clarify what would happen in the above, and give perhaps a
> link to the part of the DTspec that covers this? i'm sure i must be
> misreading this.

It's all source syntax which is mostly outside the scope of the spec.

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

Re: puzzled by online explanation as to how to "overwrite" DT node

[Robert P. J. Day]

On Wed, 18 Oct 2017, Rob Herring wrote:

> On Wed, Oct 18, 2017 at 11:40 AM, Robert P. J. Day
> <rpjday-L09J2beyid0N/H6P543EQg@public.gmane.org> wrote:
> >
> >   (WARNING: sort of newbie alert, but you know that, right? :-)
> >
> >   was perusing online tutorial on customizing device trees here:
> >
> >     http://developer.toradex.com/device-tree-customization
>
> Never seen that one.

  apparently, i picked a bad example to ask about, then. :-)

> > and got a bit confused about the part describing how to (as i read
> > it) "overwrite" entire nodes.
> >
> >   first, part on "overwriting properties" makes perfect sense --
> > here's an example in that piece about overwriting default USB
> > mode:
> >
> >   &usb0 {
> >       dr_mode = "host";
> >   };
> >
> > so, as i've always understood it, the above will do one of two
> > things:
> >
> >   1) if that property is not set in the node referenced by that label,
> >      it will be *added* to what is already there, or
> >
> >   2) if that property is already set (to whatever), it will be
> >      overwritten
> >
> > so far, so good?
>
> Yes.

  right, that was the easy part.

> >   as a special example of that right below that, there's a short
> > paragraph called "Activating/Deactivating Devices" that uses this
> > example:
> >
> >   &uart4 {
> >       status = "okay";
> >   };
> >
> > but, really, that's just a special case of the first example so
> > there's nothing new here. but it's the next section called
> > "Overwriting nodes" that seems confusing:
> >
> >   "Whole nodes can be overwritten by simply redefining them. Like
> >   overwriting properties, latter definitions overwrite earlier
> >   definitions."
>
> I don't think that's right. I think you have to use /delete-node/
> first.

  right, that was also my limited understanding, so you can
understand my confusion based on what could be interpreted as a claim
that entire nodes could be overwritten.

> > wait, what? entire nodes can be overwritten by redefining them?
> > the example given looks like this:
> >
> >   "E.g. to overwrite the pin configuration of Vybrids UART2 (UART_B)
> >   overwrite the uart2grp node by simply redefining it in your device
> >   tree ..."
> >
> > with the device tree content
> >
> >   &iomuxc {
> >       vf610-colibri {
> >           pinctrl_uart2: uart2grp {
> >               fsl,pins = <
> >                   VF610_PAD_PTD0__UART2_TX        0x21a2
> >                   VF610_PAD_PTD1__UART2_RX        0x21a1
> >               >;
> >           };
> >           ...
> >       };
> >   };
> >
> > hang on ... isn't "vf610-colibri" also a "node" in this device
> > tree? does the above mean that the *entire* vf610-colibri node is
> > being overwritten? and that i would have to supply the entire new
> > content of that node, including simply duplicating the rest of the
> > content that is already there that i want to keep? (is that what
> > the "..." represents?)
>
> No, you should get a merged vf610-colibri node with the original
> contents and any new or overwritten properties from this chunk.
>
> IOW, only properties are overwritten.

  good, that makes *way* more sense, thanks for clarifying that.
followup question coming shortly ...

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