Re: [Ksummit-discuss] Devicetree Workshop at Kernel Summit Prague (26 Oct 2017)

[Rob Herring <robh@kernel.org>]

On Fri, Oct 20, 2017 at 4:55 AM, Alexandre Torgue
<alexandre.torgue@st.com> wrote:
> Hi Frank,
>
>
> On 10/19/2017 08:46 PM, Frank Rowand wrote:
>> On 10/19/17 07:59, Rob Herring wrote:
>>> On Thu, Oct 19, 2017 at 9:00 AM, Alexandre Torgue
>>> <alexandre.torgue@st.com> wrote:
>>>>
>>>> Hi Rob,
>>>>
>>>>
>>>> On 10/19/2017 01:53 AM, Rob Herring wrote:
>>>>> On Wed, Oct 18, 2017 at 6:28 PM, Andrew Turner <andrew@fubar.geek.nz>
>>>>> wrote:
>>>
>>>
>>> [...]

>>>>          -->For example, I want to use the same dtsi files between Linux
>>>> and
>>>> U-boot. If in u-boot dts file I overload several "status" entry by
>>>> "disabled", is it possible that compiler doesn't build it ? And what
>>>> about
>>>> not used phandle ?
>>>
>>>
>>> You certainly could remove disabled nodes in dtc. I'm not sure how
>>> hard it would be to plumb into dtc. I think phandle properties are
>>> already only created if there's a reference to them. If that is
>>
>>
>> Yes, phandles are only created if referenced, unless compiled
>> for loading overlays into:
>>
>
> Are there DTC "extra" options to use to not build those useless phandles ? I
> just tried to revert the dtb to dts (using following command:
>  ./scripts/dtc/dtc -I dtb -O dts -o stm32f469-disco-flat.dts
> arch/arm/boot/dts/stm32f469-disco.dtb)
>
> I see that phandles not used are in the dts output file. It is especially an
> issue for pinmux phandles. All pinmux groups possibilities are written
> inside (in my case) stm32f4-pinctrl.dtsi. This file is included in each
> stm32 board dts files, and in those stm32 board dts files only required node
> are enabled. But I see that all pinmux definitions are embedded inside dtb
> binary (even ones not used in board dts file).

Ah, you mean removing nodes without a phandle reference, not phandles
themselves.

There's no way dtc could do that because no reference doesn't equate
to unused. For example, there's no phandle reference to the /memory
node, but that is for sure needed. We would have to add some
annotation to nodes that could be removed if unused. This could be
some source annotation (/delete-if-unused/), some extension to status
property, or some new property.

Another option would be just mark all those nodes disabled and then
postprocess the dtb to mark them okay if they have a phandle property
and delete the node if not.

Rob