1st being routing tables and rules 2nd being network namespacesI'm running into problems while testing some early / small / simple unit tests before getting deeper into a project.
I have a fairly simple configuration with a veth pair connecting the unnamed / root / default NetNS and another NetNS.
I'm using 192.0.2.0/24, .1 on one end and .2 on the other end to talk to each other.
I'm running into issues when I create the l3mdev (VRF) interface and using it as the veth master. (Obviously bring up all links.)
It doesn't matter if the l3mdev (VRF) interface is inside the NetNS or outside (in the unnamed / root / default NetNS).
Sniffing the l3mdev (VRF) interface shows that all the Ethernet frames that it sends are from it's MAC to it's MAC. Even if the kernel has the proper MAC address of the remote system.
I don't know enough about l3mdev's (VRF's) operation to know if this is how it's supposed to be have (as if it's supposed to do some other $magic that I don't know about) or if this is a symptom of a bug.
I almost feel like something in the l3mdev code is conflating the source MAC with the destination MAC on the Ethernet frames that it's sending.
Or, perhaps it's a problem with veth interfaces. I've not had an opportunity to test with more proper (less virtual) interfaces.
Can anyone comment? -- Grant. . . . unix || die
Attachment:
smime.p7s
Description: S/MIME Cryptographic Signature