On Sun, May 14, 2017 at 07:39:33AM -0700, Guenter Roeck wrote:

On 05/11/2017 12:22 AM, Shawn Guo wrote:

quoted

On Fri, May 05, 2017 at 03:32:59PM -0400, Sebastien Bourdelin wrote:

quoted

This watchdog is instantiated in a FPGA and can only be access using a
GPIOs bit-banged bus, called the NBUS by Technologic Systems.
The watchdog is made of only one register, called the feed register.
Writing to this register will re-arm the watchdog for a given time (and
enable it if it was disable). It can be disabled by writing a special
value into it.

Signed-off-by: Sebastien Bourdelin <redacted>
---
Changes v2 -> v3:
 - rebase on master
 - remove the timeout table to simplify the logic (suggested by
 Guenter Roeck)
 - fix the set_timeout function (suggested by Guenter Roeck)
 - hardcode the max_hw_heartbeat_ms value based on the ts4600 max
 timeout value supported (suggested by Guenter Roeck)
 - remove the max_timeout usage and set a default timeout value
(suggested by Guenter Roeck)
 - use the devm_watchdog_register_device function instead of
 watchdog_register_device (suggested by Guenter Roeck)

Changes v1 -> v2:
 - rebase on master
 - retrieve the ts_nbus instantiated by the parent node (suggested by
 Linus Walleij)
 - rename the wdt by watchdog in the device tree and in the
 documentation (suggested by Rob Herring)
 - add a dependency to the TS_NBUS driver in the Kconfig (suggested by
 Guenter Roeck)
 - simplify the set_timeout function (suggested by Guenter Roeck)
 - use the max_hw_heartbeat_ms callback instead of the max_timeout
 callback (suggested by Guenter Roeck)
---
.../devicetree/bindings/watchdog/ts4600-wdt.txt    |  16 ++
arch/arm/boot/dts/imx28-ts4600-common.dtsi         |   5 +

The dts change shouldn't be mixed within driver patch.

Maybe, but Rob Acked the devicetree changes, so I don't see a reason
for splitting it up now.

What Rob Acked is devicetree bindings, and ARM dts changes should
normally go via arm-soc tree, unless there is a good reason they
shouldn't.

Shawn