Re: [PATCH (v6) 1/2] mtd: brcmnand: Add brcm,bcm63268-nand device tree binding

[Brian Norris <computersforpeace@gmail.com>]

Hi Simon,

On Wed, Dec 02, 2015 at 08:12:32PM +0000, Simon Arlott wrote:
> On 02/12/15 20:00, Brian Norris wrote:
> > On Wed, Dec 02, 2015 at 07:41:07PM +0000, Simon Arlott wrote:
> >> I've created a bcm963268part driver so there won't need to be any
> >> partitions in DT for bcm63268.
> > 
> > Just curious, do you plan to submit this driver? We're working on
> 
> Yes, it's just the most recent one I've been working on. I still have
> USBH and IUDMA to submit
> 
> > matching up partition parsers to flash devices via device tree
> > of_match_table's, so you could do something like this:
> > 
> > 	nand0: nandcs@0 {
> > 		compatible = "brcm,nandcs";
> > 		...
> > 
> > 		partitions {
> > 			compatible = "brcm,bcm963268-partitions";
> > 			...
> > 		};
> > 	};
> 
> I modified brcmnand to look for a machine matching "brcm,bcm963268", but

Like this?

http://patchwork.ozlabs.org/patch/473180/

I'd like to avoid that (hence the "Rejected" status).

> that way is ok with me. Presumably "ofpart" defers to another matching
> partition parser?

Yes, "ofpart" is for specifying the entire partition table in the device
tree as subnodes of either the flash node or of the flash's "partitions"
subnode. It's not the most flexible, but it does work generically.

> Is there a patch for that method of parser detection available?

I have something working here, but I haven't had time to finish cleaning
it up and submitting it. There's an older patch that works similarly,
though it has some deficiencies:

http://patchwork.ozlabs.org/patch/475988/

The main difference between that and my (yet-to-be-submitted) proposal
is that I'd like parsers to opt in by adding a proper of_match_table
with non-Linux-specific DT bindings, and then we can drop the
"linux,..." naming and make it a reasonably generic property.

Regards,
Brian