Hi Joe,
On 7 December 2017 at 11:45, Joe Hershberger joe.hershberger@gmail.com wrote:
On Wed, Dec 6, 2017 at 8:35 PM, Chen-Yu Tsai wens@csie.org wrote:
On Thu, Dec 7, 2017 at 4:20 AM, Joe Hershberger joe.hershberger@gmail.com wrote:
On Tue, Dec 5, 2017 at 8:34 PM, Chen-Yu Tsai wens@csie.org wrote:
On Wed, Dec 6, 2017 at 4:50 AM, Joe Hershberger joe.hershberger@ni.com wrote:
On Fri, Nov 24, 2017 at 11:08 PM, Chen-Yu Tsai wens@csie.org wrote:
The EMAC syscon has configurable RX/TX delay chains for use with RGMII PHYs.
This adds support for configuring them via device tree properties. The property names and format were defined in Linux's dwmac-sun8i binding that was merged at one point.
I'm not seeing this in doc/device-tree-bindings/net/
See https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Docu...
The bindings have been restored as of v4.15-rc1.
We are following DT bindings as defined in the Linux kernel. Deviation is kept to a minimum, and eliminated if possible. We still need to migrate the driver to the new bindings for the internal PHY bits. But that bit might still be changed during the 4.15 release cycle.
That's good, but we want to have the currently supported bindings copied into the U-Boot tree under doc/device-tree-bindings/net/. Please include a patch that adds the bindings that your driver is using.
Looks like this is a new requirement. Or it wasn't really enforced before. Doesn't this make U-boot prone to having diverging device tree bindings? It has already happened with the regulator bindings, specifically the "regulator-name" property.
And by "currently supported", are you referring to what the driver expects, and not what the end result, i.e. the accepted bindings in Linux, should be? This driver is still in a state of catchup. The driver supports a previously merged then reverted set of bindings. These bindings were then brought back and updated for Linux v4.15 (unreleased yet, so may still change, again). So which set of bindings should I submit here?
And the goal _is_ to migrate the driver to what Linux is using, so we can share the device tree files.
I believe the goal is to have them match the current state of the driver as supported in U-Boot, and a higher-level goal of keeping the bindings + driver in sync with Linux for sharing.
Yes, they should match, and we should have the same binding file.
Also, please use dev_read_...() instead of fdtdec...() for all new code since that supports live tree.
Correct, Simon?
Thanks, -Joe
Regards, Simon