I have tested this change over the weekend on one of my systems configured for LACP, and under light use nothing seems amiss so far. I don't see any Kyua tests for net code, so some guidance with testing will be appreciated!
I hope someone familiar with the locking can sanity-check that particular aspect.
I have tested changing MTU several times, and I have tested the following error paths:
* Attempting to set an invalid MTU (15000) on the lagg to ensure no change is attempted
* Attempting to set a valid MTU on the lagg (16384) that is unsupported by a port (em) to ensure all ports are reverted and the error message is correctly displayed.