[PATCH v3 4/5] dt-bindings: arm64 ARCH_THUNDER2 platform documentation

On Wed, Feb 8, 2017 at 6:32 AM, Jayachandran C [off-list ref] wrote:

On Tue, Feb 07, 2017 at 04:35:24PM -0600, Rob Herring wrote:

quoted

On Tue, Feb 7, 2017 at 3:19 PM, Jayachandran C [off-list ref] wrote:

quoted

diff --git a/Documentation/devicetree/bindings/arm/cavium-thunder2.txt b/Documentation/devicetree/bindings/arm/cavium-thunder2.txt
new file mode 100644
index 0000000..82276a2
--- /dev/null
+++ b/Documentation/devicetree/bindings/arm/cavium-thunder2.txt

@@ -0,0 +1,5 @@
+Cavium Thunder2 Family device tree bindings
+-------------------------------------------
+
+Boards with Cavium ThunderX2 CN99XX shall have the root property:
+  compatible = "cavium,thunder-99xx";

Is it ThunderX2 or Thunder2? What happened to "CN" in the compatible
string. There's a lot of inconsistencies here.

I have followed the exiting Thunder naming as well as general DT
conventions, Basically:

Platform - ARCH_THUNDER2  (for ThunderX2 CN99XX, ARCH_THUNDER is
           for ThunderX CN8?XX)

The platform name is fine, we don't care too much about that as it is
easily changed
if necessary.

Core - cavium,thunder2 (ThunderX2 cpu core, "cavium,thunder" is ThunderX core)
Board/SoC - cavium,thunder-99xx (for ThunderX2, "cavium,thunder-88xx"
          is existing ThunderX board)

Thunder2 because it is based on the next generation Thunder core, and
processor family is ThunderX2 CN99XX.

I had thought thru other possibilities as well, this seems the best I
could come up with.

We should really avoid having the 'xx' placeholder in there but instead put
actual chip names in the compatible string, preferably the real ID such as
"cn8890" for the older chip. I see that we got it wrong for thunder-88xx.dts,
but that is no reason to duplicate the mistake.

     Arnd