On Thu, Mar 17, 2016 at 1:06 PM, Robin Murphy [off-list ref] wrote:

Hi Rob,

On 17/03/16 17:09, Rob Herring wrote:

quoted

On Wed, Mar 09, 2016 at 11:24:04AM +0100, Neil Armstrong wrote:

quoted

Add timer-width optional property to specify a different vendor
specific timer counter bit-width.

Signed-off-by: Neil Armstrong <redacted>
---
  Documentation/devicetree/bindings/timer/arm,sp804.txt | 2 ++
  1 file changed, 2 insertions(+)

diff --git a/Documentation/devicetree/bindings/timer/arm,sp804.txt

b/Documentation/devicetree/bindings/timer/arm,sp804.txt
index 5cd8eee7..141e143 100644

--- a/Documentation/devicetree/bindings/timer/arm,sp804.txt
+++ b/Documentation/devicetree/bindings/timer/arm,sp804.txt

@@ -17,6 +17,8 @@ Optional properties:
  - arm,sp804-has-irq = <#>: In the case of only 1 timer irq line

connected, this
        specifies if the irq connection is for timer 1 or timer 2. A
value of 1
        or 2 should be used.
+- arm,timer-width: Should contain the width in number of bits of the
counter,
+       is considered by default 32 but can be changed for vendor
variants.

That would not be an SP804 nor would the vendor be ARM in that case. So
add a new compatible string for the vendor that decided to hack up ARM's
IP block.

By all accounts this is some ancient reference design[1] which later evolved
_into_ the SP804, so that vendor would probably still be ARM ;)

Right.

A separate compatible string would indeed make more sense, though. Both
semantically and in terms of letting the driver account for the differences
automatically.

Robin.

[1]:http://infocenter.arm.com/help/topic/com.arm.doc.ddi0170a/I350250.html

Humm, same as integrator timers perhaps?

Rob

quoted

Rob

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel