Re: [PATCH 3/7] mmc: dw_mmc: add device tree support
From: Thomas Abraham <hidden>
Date: 2012-05-12 08:43:14
Also in:
linux-arm-kernel, linux-mmc, linux-samsung-soc, lkml
On 12 May 2012 12:37, Olof Johansson [off-list ref] wrote:
Hi, On Thu, May 10, 2012 at 3:15 AM, Thomas Abraham [off-list ref] wrote:quoted
Hi Olof, On 2 May 2012 23:37, Olof Johansson [off-list ref] wrote:quoted
Hi,[...]quoted
quoted
+# Slots: The slot specific information are contained within child-nodes with + each child-node representing a supported slot. There should be atleast one + child node representing a card slot. The name of the slot child node should + be 'slot{n}' where n is the unique number of the slot connnected to the + controller. The following are optional properties which can be included in + the slot child node.Since we're talking slots / cards on a bus, I think the addressing model would be useful here. So in the main controller node: #address-cells = <1>; #size-cells = <0>; And then each slot would need a reg property and possibly unit address: slot { reg = <0>; ... }; (unit addresses on the slots are only needed if they can't be disambiguated by name, so not needed if you only have one slot).Is the addressing model as described above needed in this case? The address for a slot is not used by the controller driver code and is just a virtual number. It would be sufficient to represent the nodes representing the slots with a unique name.The driver has the concept of slot IDs (slot->id struct member), and the hardware definitely enumerates them. So, I think it makes sense to give a chance to enumerate the slots in the device tree. Otherwise, how do you know which one is which on hardware? It also opens up the flexibility to have the same name for both slots if it makes sense to describe a board that way.
Thanks Olof. Yes, I missed the usage of the id number in the driver. I will add the slot addressing as you have suggested and repost. Thanks, Thomas.
-Olof