Rafael J. Wysocki wrote:
On Tue, Mar 14, 2017 at 6:54 PM, Sakari Ailus
<sakari.ailus@xxxxxxxxxxxxxxx> wrote:
Hi Rafael,
Rafael J. Wysocki wrote:
On Tue, Mar 14, 2017 at 9:09 AM, Sakari Ailus
<sakari.ailus@xxxxxxxxxxxxxxx> wrote:
On 03/14/17 10:08, Sakari Ailus wrote:
How about this instead:
All port nodes are located under the device's "_DSD" node in the
hierarchical data extension tree. The property extension related to
each port node must contain the key "port" and an integer value which
is the number of the port.
So with matching strings instead of indices, this will change, too...
It doesn't have to AFAICS, but the number is just redundant IMO. You
only need a boolean property saying "this is a port", so you know that
you should expect a list of endpoints in that object.
No, it's not redundant. It's the number of the physical port in the device
--- this is how the driver gets to know where the connection has been made.
OK, but what exactly do you mean by "physical port"?
The device (or an IP block) has physical interfaces to the world
outside. There could be just one, but there may be more. For an ISP,
there could be e.g. four CSI-2 receivers to each of which you could
connect a camera sensor. So for an ISP device, that number tells which
of the receivers a given sensor is connected to.
The mapping between this number and what the hardware datasheet refers
to needs to be documented per device.
--
Regards,
Sakari Ailus
sakari.ailus@xxxxxxxxxxxxxxx
--
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
