On 11/08/2013 07:24 PM, Jason Cooper wrote:
On Fri, Nov 08, 2013 at 12:06:26PM -0600, Kumar Gala wrote:
On Nov 8, 2013, at 10:57 AM, Jason Cooper wrote:
On Fri, Nov 08, 2013 at 10:13:19AM -0600, Kumar Gala wrote:
On Nov 5, 2013, at 8:28 AM, Sebastian Hesselbarth wrote:
...
.../devicetree/bindings/arm/marvell,berlin.txt | 24 +++
arch/arm/boot/dts/Makefile | 2 +
arch/arm/boot/dts/berlin2-sony-nsz-gs7.dts | 29 +++
arch/arm/boot/dts/berlin2.dtsi | 227 ++++++++++++++++++++
4 files changed, 282 insertions(+)
create mode 100644 Documentation/devicetree/bindings/arm/marvell,berlin.txt
create mode 100644 arch/arm/boot/dts/berlin2-sony-nsz-gs7.dts
create mode 100644 arch/arm/boot/dts/berlin2.dtsi
Haven't we been trying to go away from non-prefixed dts/dtsi?
hmmm, this is the first I've heard of that. Although, your proposal
(in another thread) makes more sense now. :)
So should these be something like marvell-berlin2-...
I don't recall this being brought up at the summit, nor in Grant's
report. I do need to give it a more careful read this weekend, though.
Perhaps I missed something.
This was based on review comments Olof gave when we pushed some .dts
files for MSM/APQ Qualcomm Technologies soc/boards.
As Andrew Lunn mentioned to me earlier, we should consider the fact that
the dts file names are being used by Debian's flash-kernel. Oh no!
Another ABI! ;-)
Personally, I think dtc should be using the board compatible string, and
naming the resultant dtbs the same, eg
$ dtc kirkwood-dreamplug.dts
globalscale,dreamplug-003-ds2001.dtb
This would free us up to rename the dts files as needed. dtb filenames
would be guaranteed unique and consistent. It would also allow us to
catch inadvertent compatible string collisions earlier.
Besides the dtc output scheme, I also like the idea of a more consistent
dts/dtsi file naming. What about the scheme below, which also allows us
to get rid of vendor-soc-vendor-board.dts naming (which already leads to
quite long names):
(a) vendor,soc.dtsi and vendor,board.dts
or if dtc/cpp doesn't like ',' in includes
(b) vendor-soc.dtsi and vendor-board.dts
for this patch, that would give us
marvell,berlin2.dtsi and sony,nsz-gs7.dts.
If you propose a scheme we should follow, I'd be happy to apply that
first on this patch set. If we have a dtc output scheme, too, that
scheme can be more like best practice instead of a hard rule.
Sebastian
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
