On Tuesday 18 June 2013, Chander Kashyap wrote:

quoted

quoted

+       [Core Clocks]
+
+  Clock                      ID
+  ----------------------------
+
+  fin_pll            1
+
+  [Clock Gate for Special Clocks]
+
+  Clock                      ID
+  ----------------------------
+  sclk_uart0         128
+  sclk_uart1         129
+  sclk_uart2         130

quoted

+
+   [Peripheral Clock Gates]
+
+  Clock                      ID
+  ----------------------------
+
+  aclk66_peric               256
+  uart0                      257
+  uart1                      258

It looks like these are actually separate things. Wouldn't it be more sensible
to have separate device nodes for each of the lists and use a local index?

I have listed the parent clock first, then the child clocks, to
maintain  readability.

quoted

What numbers are used in the data sheet?

I didn't get your point?

I would have expected three clock device nodes, one for fin_pll (presumably a
fixed-rate clock?), one for "special clocks" and one for "peripheral clock
gates", and a number space starting at '1' for each of them, rather than
having a shared node and numbers starting at '1', '128' and '256', which looks
a bit clumsy.

Did you take the ID number definitions from a data sheet, or did you make up
the numbers yourself for the purpose of defining a binding?

	Arnd