On 27.07.2022 23:31, Frank Wunderlich wrote:

quoted

Gesendet: Dienstag, 26. Juli 2022 um 14:24 Uhr
Von: "Arınç ÜNAL" [off-list ref]

Hi,

quoted

To Frank:
Let me know if MII bindings for port 5 and 6 on MT7531 are okay.

I ack krzysztof, that the change is really huge and it is hard to understand what exactly is changed and why. So please split. I have converted
it to yaml, but not have changed the logic itself. I guess you know the switch better than me.

quoted

Does your recent patch for MT7531 make it possible to set any port for CPU,
including user ports? For now, I put a rule to restrict CPU ports to 5 and
6, as described on the description of dsa port reg property.

i only know that port 5 and 6 are possible, not about the other ports. Afair there was a check if port 5 or 6 (followed by available modes
like rgmii, trgmii or sgmii) and then allow cpu-port-mode else allow only user-port mode. Had not changed this, so currently only these 2
ports can be used as CPU.

quoted

I suppose your patch does not bring support for using MT7530's port 5 as
CPU port. We could try this on a BPI-R2. Device schematics show that
MT7530's GMII pins are wired to GMAC1 of MT7623NI to work as RGMII.

my patches (and the version from Vladimir that was merged) only solves the Problem that CPU-Port was fixed to 6 before. I tested Patches on r2
(mt7530) and r64 too (mt7531) that they do not break anything. But i have not disabled port 6 (maybe i had to do so for a port5-only mode), only
enabled port 5 too and run iperf3 over a vlan-aware bridge between wan-port and port 5.

I've seen this under mt7530_setup():
The bit for MHWTRAP_P6_DIS is cleared to enable port 6 with the comment 
"Enable Port 6 only; P5 as GMAC5 which currently is not supported".

https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/tree/drivers/net/dsa/mt7530.c#n2189

I'm slowly learning C programming so I can't currently manage to do 
something on my own here, just letting you know.

Arınç