"Serge E. Hallyn" <serue@xxxxxxxxxx> writes: > Quoting Krishna Vamsi-B22174 (avamsi@xxxxxxxxxxxxx): >> >> >> Hi, >> >> I am a newbie to this list. Here is my use case , we have Loadable >> Kernel Module which applies security to >> the packets arriving from 1000 networks with overlap addresses. There >> are 3 different user space process which handles >> control traffic from these 1000 networks . >> >> Please let me know >> >> 1)How to create a Network Namespace Object ? > > clone(CLONE_NEWNET) > >> 2)How to delete a Network Namespace Object ? > > exit > >> 3)Can these 3 user space process see all the Network Namespace objects >> created in the kernel ? > > No, network namespaces are fully isolated. A virtual nic can only exist > in one network namespace, and physical nics can only exist in the > initial network namespace. Sockets can be passed between network namespaces if you set things up correctly. At which point you can have 3 user space processes doing all of the work. It can be a bit of a pain to have processes lying around just so you can create a socket in another network namespace but the code works today and isn't too bad. >> If so, how can they access these objects? >> 4)How to group 2-3 interfaces under a particular Network Namespace ? > > I don't understand the question, but you pass a veth endpoint into a > network namespace using > > /sbin/ip link set veth1 netns $pid_in_other_netns yep. Eric _______________________________________________ Containers mailing list Containers@xxxxxxxxxxxxxxxxxxxxxxxxxx https://lists.linux-foundation.org/mailman/listinfo/containers