[PATCH V2 3/3] ARM: kirkwood: Define NAND partitions in dts

[Jason Cooper]

On Mon, Mar 26, 2012 at 05:36:10PM +0100, Jamie Lentin wrote:
> On Mon, 26 Mar 2012, Scott Wood wrote:
> 
> >On 03/26/2012 11:20 AM, Jason Cooper wrote:
> >>On Mon, Mar 26, 2012 at 10:53:29AM -0500, Scott Wood wrote:
> >>>On 03/24/2012 08:14 AM, Jamie Lentin wrote:
> >>>>Use devicetree to define NAND partitions. Use D-link partition scheme by
> >>>>default, to be vaguely compatible with their userland.
> >>>>
> >>>>Signed-off-by: Jamie Lentin <jm at lentin.co.uk>
> >>>>---
> >>>> arch/arm/boot/dts/kirkwood-dns320.dts |   35 +++++++++++++++++++++++++++++++++
> >>>> arch/arm/boot/dts/kirkwood-dns325.dts |   35 +++++++++++++++++++++++++++++++++
> >>>> arch/arm/mach-kirkwood/board-dnskw.c  |   31 -----------------------------
> >>>> 3 files changed, 70 insertions(+), 31 deletions(-)
> >>>>
> >>>>diff --git a/arch/arm/boot/dts/kirkwood-dns320.dts b/arch/arm/boot/dts/kirkwood-dns320.dts
> >>>>index 58de7f2..fbf55ff 100644
> >>>>--- a/arch/arm/boot/dts/kirkwood-dns320.dts
> >>>>+++ b/arch/arm/boot/dts/kirkwood-dns320.dts
> >>>>@@ -25,5 +25,40 @@
> >>>> 			clock-frequency = <166666667>;
> >>>> 			status = "ok";
> >>>> 		};
> >>>>+
> >>>>+		nand at 3000000 {
> >>>>+			status = "ok";
> >>>>+
> >>>
> >>>This should be "okay", not "ok" -- see IEEE1275.  Or just leave it out.
> >>
> >>Ack, but it needs to be there.  Most, but not all, kirkwood boards have
> >>nand, so we define it in kirkwood.dtsi and set it as disabled.
> >>Individual boards can then enable it as needed.
> >>
> >>As for 'okay', looks like we may need to patch of_device_is_available()
> >>in drivers/of/base.c (~284) if we want to be consistent with IEEE1275.
> >
> >No need to change of_device_is_available() -- it handles the
> >standards-compliant "okay" as well as "ok" which is non-compliant but
> >probably exists in some broken real OF trees (and even if not, it's bad
> >to break compatibility with older device trees without a good reason).
> >
> >Maybe add a comment indicating which should be used.
> 
> I need to re-jig this part anyway, so will update to "okay"
> too---looks like a fairly harmless change.
> 
> Would it make sense to change drivers/of/base.c to emit a warning if
> "ok" is used? Or are there already too many devicetrees in the wild
> that use "ok"?

There's only two powerpc boards using this.  I have a patch I'm about
to submit that just adds a comment above the test.  Since that's where I
pulled "ok" from, that should prevent others in the future from repeating
my mistake.

thx,

Jason.