5 Mahesh Bandewar <maheshb AT google.com>
8 This is conceptually very similar to the macvlan driver with one major
9 exception of using L3 for mux-ing /demux-ing among slaves. This property makes
10 the master device share the L2 with it's slave devices. I have developed this
11 driver in conjunction with network namespaces and not sure if there is use case
15 2. Building and Installation:
16 In order to build the driver, please select the config item CONFIG_IPVLAN.
17 The driver can be built into the kernel (CONFIG_IPVLAN=y) or as a module
22 There are no module parameters for this driver and it can be configured
23 using IProute2/ip utility.
25 ip link add link <master-dev> <slave-dev> type ipvlan mode { l2 | l3 | l3s }
27 e.g. ip link add link ipvl0 eth0 type ipvlan mode l2
31 IPvlan has two modes of operation - L2 and L3. For a given master device,
32 you can select one of these two modes and all slaves on that master will
33 operate in the same (selected) mode. The RX mode is almost identical except
34 that in L3 mode the slaves wont receive any multicast / broadcast traffic.
35 L3 mode is more restrictive since routing is controlled from the other (mostly)
39 In this mode TX processing happens on the stack instance attached to the
40 slave device and packets are switched and queued to the master device to send
41 out. In this mode the slaves will RX/TX multicast and broadcast (if applicable)
45 In this mode TX processing up to L3 happens on the stack instance attached
46 to the slave device and packets are switched to the stack instance of the
47 master device for the L2 processing and routing from that instance will be
48 used before packets are queued on the outbound device. In this mode the slaves
49 will not receive nor can send multicast / broadcast traffic.
52 This is very similar to the L3 mode except that iptables (conn-tracking)
53 works in this mode and hence it is L3-symmetric (L3s). This will have slightly less
54 performance but that shouldn't matter since you are choosing this mode over plain-L3
55 mode to make conn-tracking work.
57 5. What to choose (macvlan vs. ipvlan)?
58 These two devices are very similar in many regards and the specific use
59 case could very well define which device to choose. if one of the following
60 situations defines your use case then you can choose to use ipvlan -
61 (a) The Linux host that is connected to the external switch / router has
62 policy configured that allows only one mac per port.
63 (b) No of virtual devices created on a master exceed the mac capacity and
64 puts the NIC in promiscuous mode and degraded performance is a concern.
65 (c) If the slave device is to be put into the hostile / untrusted network
66 namespace where L2 on the slave could be changed / misused.
69 6. Example configuration:
71 +=============================================================+
74 | +----------------------+ +----------------------+ |
75 | | NS:ns0 | | NS:ns1 | |
78 | | ipvl0 | | ipvl1 | |
79 | +----------#-----------+ +-----------#----------+ |
81 | ################################ |
83 +==============================#==============================+
86 (a) Create two network namespaces - ns0, ns1
90 (b) Create two ipvlan slaves on eth0 (master device)
91 ip link add link eth0 ipvl0 type ipvlan mode l2
92 ip link add link eth0 ipvl1 type ipvlan mode l2
94 (c) Assign slaves to the respective network namespaces
95 ip link set dev ipvl0 netns ns0
96 ip link set dev ipvl1 netns ns1
98 (d) Now switch to the namespace (ns0 or ns1) to configure the slave devices
100 (1) ip netns exec ns0 bash
101 (2) ip link set dev ipvl0 up
102 (3) ip link set dev lo up
103 (4) ip -4 addr add 127.0.0.1 dev lo
104 (5) ip -4 addr add $IPADDR dev ipvl0
105 (6) ip -4 route add default via $ROUTER dev ipvl0
107 (1) ip netns exec ns1 bash
108 (2) ip link set dev ipvl1 up
109 (3) ip link set dev lo up
110 (4) ip -4 addr add 127.0.0.1 dev lo
111 (5) ip -4 addr add $IPADDR dev ipvl1
112 (6) ip -4 route add default via $ROUTER dev ipvl1