On Fri, Jan 16, 2015 at 10:09 PM, Mark Rutland [off-list ref] wrote:

On Fri, Jan 16, 2015 at 12:49:20PM +0000, Orson Zhai wrote:

quoted

Hi, Mark,

On Fri, Jan 16, 2015 at 6:35 PM, Mark Rutland [off-list ref] wrote:

[...]

quoted

 It returns "status busy" after I type the command below.

Is this while other CPUs are active, or after you've hotplugged out all
other CPUs? The latter is expected, but the former is not.

Are you able to hotplug CPU0 out and back in while other CPUs are
active

Yes, I have just tested again and confirm that cpu0 could be
turned on/off when one of the other 3 CPUs was active.

[...]

For this bus you could instead use addresses relative to the bus inside,
rather than absolute addresses,

Do you mean something like  apb at 0 {...} apb at 1 {...} ?

or you could have:

ranges = <0x0 0x70000000 0x0 0x70000000 0x0 0x10000000>;

[...]

quoted

This initial patch is picked up from a very big dt file.
There are several apb buses in this chip.

At the same level us the bus hierarchy?

These buses are led by some bus matrix to different areas.
But I'd like treat them simply at same level.
Is it right to do that?

quoted

So I use apb at starting-address to separate them.
But I remember another rule that the @address needs to equal  first
address in property reg array.
Do I have to delete @7000000 as well if i delete reg line?

Hmm. I'm not too keen on encoding a reg or unit-address here, because
the control interface of the bus isn't at that address. If we want to
add that later then the reg would be different in those cases. Given
there's no control interface here, there shouldn't be a reg or
unit-address.

Do you mean something like

   ap-ahb { ... };
   aon-ahb { ... };

without using @xxxx ?

   Thanks, Orson

Just ensure that the name before the unit-address is unique.

Thanks,
Mark.