> - [Medium] netlink: AF_PACKET multicast dump misses changes due to > reading sequence after loop
This seems real, I will send a v11 patch, will sample before the walk, and once more after it for the NLMSG_DONE check. > - [Low] __hw_addr_changed() (net/core/dev_addr_lists.c) does > atomic_inc(&dev_net(list->mc_dev)->dev_mc_genid). Same as every other dev->mc writer under the address lock, nothing new.
