On 10/25/2012 03:00 AM, Daniel Mack wrote:

Hi Jon,

many thanks for your time to look at this.

On 25.10.2012 03:28, Jon Hunter wrote:

quoted

On 10/22/2012 02:55 PM, Daniel Mack wrote:

quoted

diff --git a/Documentation/devicetree/bindings/bus/gpmc.txt b/Documentation/devicetree/bindings/bus/gpmc.txt
new file mode 100644
index 0000000..ef1c6e1
--- /dev/null
+++ b/Documentation/devicetree/bindings/bus/gpmc.txt

@@ -0,0 +1,59 @@
+Device tree bindings for OMAP general purpose memory controllers (GPMC)
+
+The actual devices are instantiated from the child nodes of a GPMC node.
+
+Required properties:
+
+ - compatible: Should be set to "ti,gpmc"

Is this the only required property? I think that "reg" and "ti,hwmods"
are probably also required.

Well yes, but at least "reg" is commonly omitted as it's part of a more
"generic" set of properties. But ok, I can add these.

quoted

Also given that we are describing the hardware, I am wondering if the
number of chip-selects and wait signals should be defined here too. I
recall that different devices had different number of wait pins available.

Hmm, that number is currently hard-coded in GPMC_CS_NUM. It would take
some effort to make that dynamic but I agree that this would be a good
thing to have. Afzal?

I believe that today all OMAP/AM devices have 8 chip-selects so probably
not a big deal. However, given we are moving to DT it would be nice to
move away from having such #defines for hardware related items.

Cheers
Jon