[PATCH][v3] dt-bindings: ifc: Update endianness usage

Hi Rob,

-----Original Message-----
From: devicetree-owner at vger.kernel.org [mailto:devicetree-
owner at vger.kernel.org] On Behalf Of Prabhakar Kushwaha
Sent: Tuesday, December 05, 2017 2:45 PM
To: Rob Herring <robh@kernel.org>
Cc: devicetree at vger.kernel.org; mark.rutland at arm.com;
shawnguo at kernel.org; linux-arm-kernel at lists.infradead.org
Subject: RE: [PATCH][v3] dt-bindings: ifc: Update endianness usage

quoted

-----Original Message-----
From: Rob Herring [mailto:robh at kernel.org]
Sent: Tuesday, December 05, 2017 2:17 AM
To: Prabhakar Kushwaha <redacted>
Cc: devicetree at vger.kernel.org; mark.rutland at arm.com;
shawnguo at kernel.org; linux-arm-kernel at lists.infradead.org
Subject: Re: [PATCH][v3] dt-bindings: ifc: Update endianness usage

On Thu, Nov 30, 2017 at 01:36:36PM +0530, Prabhakar Kushwaha wrote:

quoted

IFC controller version < 2.0 support IFC register access as
big endian. These controller version also require IFC NOR signals to
be connected in reverse order with NOR flash.

IFC >= 2.0 is other way around.

So updating IFC binding to take care of both using endianness field.

Signed-off-by: Prabhakar Kushwaha <redacted>
---
Changes for v2: updated subject
Changes for v3: fixed typo for "big-endian"

 Documentation/devicetree/bindings/memory-controllers/fsl/ifc.txt | 6

++++--

quoted

quoted

 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/Documentation/devicetree/bindings/memory-

controllers/fsl/ifc.txt

quoted

b/Documentation/devicetree/bindings/memory-controllers/fsl/ifc.txt

quoted

index 89427b0..824a2ca 100644

--- a/Documentation/devicetree/bindings/memory-controllers/fsl/ifc.txt
+++ b/Documentation/devicetree/bindings/memory-controllers/fsl/ifc.txt

@@ -18,8 +18,10 @@ Properties:
               interrupt (NAND_EVTER_STAT).  If there is only one,
               that interrupt reports both types of event.

-- little-endian : If this property is absent, the big-endian mode will
-                  be in use as default for registers.
+- little-endian or big-endian : It represents how IFC registers to be accessed.
+			It also represents connection between controller and
+			NOR flash. If this property is absent, the big-endian
+			mode will be in use as default.

My question on the prior version remains. I think if you need to handle
more than just register endianness, that should be done with the
compatible string.

I may not able to use compatible string as this information will also be used it
drivers/mtd/maps/physmap_of_core.c other than drivers/memory/fsl_ifc.c.
I am trying to avoid controller specific details in generic file.

This is the reason endianness property is being used.

Please let me know if I am not able to address your review comment

--prabhakar