Re: [PATCH 0/5 v2] mv643xx_eth: device tree bindings

[Thomas Petazzoni <thomas.petazzoni@free-electrons.com>]

Dear Jason Cooper,

On Thu, 11 Apr 2013 12:53:03 -0400, Jason Cooper wrote:

> With this and Thomas' pci series, we will have Kirkwood fully
> converted to devicetree, can begin removing board files, and finally
> begin migrating everything over to mach-mvebu/.  This will lead to
> the removal of five directories under arch/arm/ (plat-orion,
> mach-kirkwood, mach-orion5x, mach-dove, and mach-mv78xx0).

No, we can't remove mach-orion5x and mach-mv78xx0 for now, they have
not been converted at all in terms of gpio/pinmux support. I've started
writing some pinmux code for orion5x but it's not ready.

I don't think we should rush in doing this merge into mach-mvebu. I'd
prefer to have the following conditions be met before a platform gets
merged into mach-mvebu:

 * The mach-<foo> directory no longer depends on anything in
   plat-orion. For mach-kirkwood, things are OK (for the DT platforms)
   for GPIO, MPP and PCIe, but some work remains for the IRQ controller
   driver and the timer driver.

 * All boards have been converted to the Device Tree. For
   mach-kirkwood, we're getting closer.

Once a given mach-<foo> platform meets these conditions, it will be a
clean platform, and we can merge it into mach-mvebu without any
problem.

So let's continue the cleanup work, conversion to DT, usage of drivers
in drivers/, improvement of pinmux to support orion5x/mv78xx0, and do
the move per-platform, once they are ready.

What do you think?

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com