[PATCH 2/2] mtd: fsl_ifc_nand: Probe partitions OF node

[Aaron Sierra]

----- Original Message -----
> From: "Scott Wood" <scottwood at freescale.com>
> Sent: Wednesday, August 20, 2014 6:59:56 PM
> 
> On Wed, 2014-08-20 at 18:45 -0500, Aaron Sierra wrote:
> > ----- Original Message -----
> > > From: "Scott Wood" <scottwood at freescale.com>
> > > Sent: Wednesday, August 20, 2014 6:26:25 PM
> > > 
> > > On Fri, 2014-08-15 at 19:47 -0500, Aaron Sierra wrote:
> > > > Previously, the OF node defining the IFC NAND controller was being
> > > > passed to mtd_device_parse_register(), not the node defining the
> > > > partitions. This resulted in no OF-defined partitions being created.
> > > 
> > > This driver probes on "fsl,ifc-nand", not "fsl,ifc".  So how is it
> > > getting the controller node?
> > > 
> > > What does the device tree look like in which you're seeing this happen?
> > > 
> > > -Scott
> > > 
> > 
> > This is the node that is defined in my T1042 device tree:
> > 
> > 		nand0 at 4,0 {
> > 			#address-cells = <0>;
> > 			#size-cells = <0>;
> > 			compatible = "fsl,ifc-nand";
> >                         reg = <4 0x0 0x040000>;
> > 
> > 			nand at 0 {
> > 				#address-cells = <1>;
> > 				#size-cells = <2>;
> > 				compatible = "micron,mt29f32g08";
> > 
> > 				partition at 0 {
> > 					label = "NAND Filesystem";
> > 					reg = <0 0x1 0x00000000>;
> > 				};
> > 			};
> > 		};
> 
> This is wrong.  You shouldn't have a separate nand at 0 node.
> 
> Your patch will break partition detection on all of the existing IFC
> device trees.
> 
> > It is based on a node used previously with the fsl_upm NAND driver on
> > a P5020:
> 
> Why would you base it on a upm node rather than on an ifc node, or on
> the IFC binding document
> (Documentation/devicetree/bindings/memory-controllers/fsl/ifc.txt)?
> 
> -Scott
> 

Ah, I was under the wrong impression that there was a common MTD
partitioning format. I didn't realize it was defined by each
driver. Thanks for clearing that up.

-Aaron