That's the plan. I just need slightly more time for this kind of test. Thank you for mentioning that.
The funny thing is that I wouldn't notice unless I connected wrong cables and it worked and It shouldn't 
. I hope your RTL8380 works OK but if isolation is broken, you might not know...
Thank you for explaining the error message. And you're right, on L2 switch, CPU should be pretty much idle.
Good point. I do agree in principle. This is the typical access list problem where removing an acl "works"- failing to recognize that whole point was that it should block traffic.
But I believe I would notice for several reasons. Mainly because I monitor the traffic using mini_snmpd and cacti. Any such leakage would cause very unexpected traffic patterns in my networks, like for example TV streams on the management VLAN. This was quite obvious with the flooding bug I mentioned. Finally, some of my VLANs have enough traffic to bring down the ethernet driver, which is very noticeable
Just to be 103% sure, I just snooped on one of the trunks without much traffic to see if there is anything unexpected there, like a large number of untagged frames or unexpected VLAN IDs or similar. And there wasn't.
Good news, thank you for testing .