On Thu, Aug 19, 2021 at 08:04:15PM +0300, Vladimir Oltean wrote:

The ocelot switch driver used to ignore ports which do not have a
phy-handle property and not probe those, but this is not quite ok since
it is valid to not have a phy-handle property if there is a fixed-link.

It seems that checking for a phy-handle was a proxy for the proper check
which is for the status, but that doesn't make a lot of sense, since the
ocelot driver already iterates using for_each_available_child_of_node
which skips the disabled ports, so I have no idea.

Anyway, a widespread pattern in device trees is for a SoC dtsi to
disable by default all hardware, and let board dts files enable what is
used. So let's do that and enable only the ports with a phy-handle in
the pcb120 and pcb123 device tree files.

Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
---
 arch/mips/boot/dts/mscc/ocelot.dtsi       | 11 +++++++++++
 arch/mips/boot/dts/mscc/ocelot_pcb120.dts |  8 ++++++++
 arch/mips/boot/dts/mscc/ocelot_pcb123.dts |  4 ++++
 3 files changed, 23 insertions(+)

applied to mips-next.

Thomas.

-- 
Crap can work. Given enough thrust pigs will fly, but it's not necessarily a
good idea.                                                [ RFC1925, 2.3 ]