Is it possible to VLAN over a VLAN interface?

Question

Computer networks

A managed switch is connected to the eth0 interface, one of the ports of which is connected to another managed switch. VLAN 100 is configured on port 1 of the first switch, and VLAN 200 is configured on some ports of the other switch. Can I create a VLAN 100 interface on the physical interface eth0 (eth0.100) in linux, and then create a VLAN 200 interface specifying eth0 as the physical interface .100.
All this is started in order to connect a managed switch to an openwrt router, which itself has a managed switch soldered on the router board.

Answer 1

On the first switch, you need to configure the 200th vlan in the trunk to the port where the first switch is connected and to the port where the network looks, and to the network on eth0 in the form of eth0.200, and not as you want eth0.100.200

Answer 2

Yes, it is possible! The technology is called q-in-q, it is quite actively used, support is needed in the driver of the card, and most modern drivers can do this. www.opennet.ru/tips/info/1381.shtml

Answer 3

May I perform the Indian national rite "nahua"?
IMHO you do not understand the essence of traffic tagging, i.e. 802.1Q standard . On a physical channel, switch-switch, switch-router, router-router, etc. you can skip the required number of VLANs, and on the port where several VLANs are terminated, just hang up subinterfaces. Those. the network architecture, if I could correctly understand the problem statement, will be as follows:
1) the channel "openwrt router with a managed switch onboard <-> managed switch 2" - a trunk through which tagged traffic VLAN 100, VLAN 200, and maybe some other , depending on how the trunk is configured;
2) channel "managed switch 2 <-> managed switch 1" - a trunk through which tagged traffic VLAN 100, VLAN 200, and maybe some other, depending on how you configure the trunk;
3) channel "managed switch 1 <-> interface eth0 Linux host", through which tagged traffic VLAN 100, VLAN 200, and maybe some other, depending on how the trunk is configured;
4) channel termination - subinterfaces ehh0.100, eth0.200, etc. on the eth0 interface of the Linux host.
But if you lack about 4000 VLANs between pieces of equipment, or some other need arises, for example, in a data center or at a provider, push a pack of VLANs of one client inside the VLAN '), in which an additional tag of another VLAN is added to the ethernet frame header, i.e. VLAN is encapsulated within a VLAN. This is usually done on network equipment, which should obviously support this technology - on the one hand, the channel was packed, on the other, it was unpacked, so that the QinQ channel would not come to the host personally, although it exists in the DC of virtual infrastructure providers, but as a host usually the hypervisor acts.