ubiswitch:~$ I: Header is empty, no configurations
I: No DB migrations are needed
I: Initializing DSA switch ports
I: Get LAG configuration from database
I: Header is empty, no configurations
I: No LAG configuration in database
I: Get port configuration from database
I: No data found eeprom section
I: No port configuration in database
I: Get VLAN configuration from database
I: Header is empty, no configurations
I: No VLAN configuration in database
It seems to me the firmware is only disabling the MAC of the port via port register 0x1 ForcedLink+LinkValue. However, the datasheet states that for disabling a port, it is better to use register 0x4 PortState. There is also PHY register PowerDown that can be set to save power.
It’s not clear to me why a port with a disabled MAC still transmits data, though…
Yep! This was logged in my backlog for BloxOsLite - thanks for the report.
At present port <int> link ... will enable/disable forwarding on the port (i.e. affects the queue controller policy). It doesn’t touch the MAC or PHY settings and admittedly I suspect this has confused everybody because the word link as an arg has a specific connotation… I’ve updated the docs to reflect this better.
I’ll post here when I implement an update relevant to powering down the internal PHYs.
Hmm, strange. I would expect to see the PHY link light with forwarding disabled, but there should be no egress from the disabled port (i.e. no ping between ports). I see you’ve reported this in a separate email and I’ve already added this to check on BloxOsLite when I get around to the next release.