it exists default namedpace config, and you still create net namespace by yourself.vpp cli reference docs introduces config field.
---------- alexan --------------原始邮件-------------- 发件人："Moritz Beck via lists.fd.io "<[email protected]>; 发送时间：2026年1月29日(星期四) 上午8:46 收件人："[email protected]" <[email protected]>; 主题：[vpp-dev] linux-cp does not work in different namespaces ----------------------------------- /* Font Definitions */ @font-face {font-family:"Cambria Math"; panose-1:2 4 5 3 5 4 6 3 2 4;} @font-face {font-family:Aptos;} /* Style Definitions */ .qmbox p.MsoNormal, .qmbox li.MsoNormal, .qmbox div.MsoNormal {margin:0cm; font-size:12.0pt; font-family:"Aptos",sans-serif;} .qmbox span.E-MailFormatvorlage17 {mso-style-type:personal-compose; font-family:"Aptos",sans-serif; color:windowtext;} .qmbox .MsoChpDefault {mso-style-type:export-only; mso-fareast-language:EN-US;} @page WordSection1 {size:612.0pt 792.0pt; margin:70.85pt 70.85pt 2.0cm 70.85pt;} .qmbox div.WordSection1 {page:WordSection1;} I have two interfaces that are bound to two different tables in VPP. I want to create two kernel interfaces that run in different namespaces and relay the netlink messages from there accordingly. VPP itself is running in the default namespace. linux-cp does not have a default namespace set. # Relevant startup.conf: dpdk { dev 0000:3b:10.1 dev 0000:3b:10.3 } plugins { plugin default { disable } plugin dpdk_plugin.so { enable } plugin linux_cp_plugin.so { enable } plugin linux_nl_plugin.so { enable } } linux-cp { lcp-auto-subint lcp-sync } # Linux config: ip netns add ns1 ip netns add ns2 # VPP config: ip table add 1 ip table add 2 ip6 table add 1 ip6 table add 2 set interface ip table VirtualFunctionEthernet3b/10/1 1 set interface ip table VirtualFunctionEthernet3b/10/3 2 set interface ip6 table VirtualFunctionEthernet3b/10/1 1 set interface ip6 table VirtualFunctionEthernet3b/10/3 2 # working: vpp# show interface Name Idx State MTU (L3/IP4/IP6/MPLS) Counter Count VirtualFunctionEthernet3b/10/1 1 down 9000/0/0/0 VirtualFunctionEthernet3b/10/3 2 down 9000/0/0/0 vpp# lcp create VirtualFunctionEthernet3b/10/1 host-if if-ns1 vpp# lcp create VirtualFunctionEthernet3b/10/3 host-if if-ns2 $ ip link 151: if-ns1: <BROADCAST,MULTICAST> mtu 9000 qdisc mq state DOWN mode DEFAULT group default qlen 1000 link/ether 02:09:c0:4b:41:34 brd ff:ff:ff:ff:ff:ff 152: if-ns2: <BROADCAST,MULTICAST> mtu 9000 qdisc mq state DOWN mode DEFAULT group default qlen 1000 link/ether 02:09:c0:00:3d:84 brd ff:ff:ff:ff:ff:ff $ ip link set dev if-ns1 up $ ip link set dev if-ns2 up $ ip link 151: if-ns1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc mq state UNKNOWN mode DEFAULT group default qlen 1000 link/ether 02:09:c0:4b:41:34 brd ff:ff:ff:ff:ff:ff 152: if-ns2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc mq state UNKNOWN mode DEFAULT group default qlen 1000 link/ether 02:09:c0:00:3d:84 brd ff:ff:ff:ff:ff:ff vpp# show interface Name Idx State MTU (L3/IP4/IP6/MPLS) Counter Count VirtualFunctionEthernet3b/10/1 1 up 9000/0/0/0 VirtualFunctionEthernet3b/10/3 2 up 9000/0/0/0 # Not working: vpp# show interface Name Idx State MTU (L3/IP4/IP6/MPLS) Counter Count VirtualFunctionEthernet3b/10/1 1 down 9000/0/0/0 VirtualFunctionEthernet3b/10/3 2 down 9000/0/0/0 vpp# lcp create VirtualFunctionEthernet3b/10/1 host-if if-ns1 netns ns1 vpp# lcp create VirtualFunctionEthernet3b/10/3 host-if if-ns2 netns ns2 $ ip -all netns exec ip link netns: ns2 1: lo: <LOOPBACK> mtu 65536 qdisc noop state DOWN mode DEFAULT group default qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 3: if-ns2: <BROADCAST,MULTICAST> mtu 9000 qdisc mq state DOWN mode DEFAULT group default qlen 1000 link/ether 02:09:c0:45:49:2d brd ff:ff:ff:ff:ff:ff netns: ns1 1: lo: <LOOPBACK> mtu 65536 qdisc noop state DOWN mode DEFAULT group default qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 3: if-ns1: <BROADCAST,MULTICAST> mtu 9000 qdisc mq state DOWN mode DEFAULT group default qlen 1000 link/ether 02:09:c0:8d:d4:08 brd ff:ff:ff:ff:ff:ff $ ip netns exec ns1 ip link set dev if-ns1 up $ ip netns exec ns2 ip link set dev if-ns2 up $ ip -all netns exec ip link netns: ns2 1: lo: <LOOPBACK> mtu 65536 qdisc noop state DOWN mode DEFAULT group default qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 3: if-ns2: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc mq state UNKNOWN mode DEFAULT group default qlen 1000 link/ether 02:09:c0:45:49:2d brd ff:ff:ff:ff:ff:ff netns: ns1 1: lo: <LOOPBACK> mtu 65536 qdisc noop state DOWN mode DEFAULT group default qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 3: if-ns1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 9000 qdisc mq state UNKNOWN mode DEFAULT group default qlen 1000 link/ether 02:09:c0:8d:d4:08 brd ff:ff:ff:ff:ff:ff vpp# show interface Name Idx State MTU (L3/IP4/IP6/MPLS) Counter Count VirtualFunctionEthernet3b/10/1 1 down 9000/0/0/0 VirtualFunctionEthernet3b/10/3 2 down 9000/0/0/0 # Corresponding log entries: vpp[4403]: vpp[4403]: clib_c11_violation: s2 NULL vpp[4403]: clib_c11_violation: s2 NULL vpp[4403]: vpp[4403]: clib_c11_violation: s2 NULL vpp[4403]: clib_c11_violation: s2 NULL As you can see, the netlink messages only get relayed from the default namespace and not the other ones. Creating the interfaces in the default namespace and then moving them to the other ones also does not work. It results in the interfaces disappearing from the kernel and VPP altogether.
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#26744): https://lists.fd.io/g/vpp-dev/message/26744 Mute This Topic: https://lists.fd.io/mt/117519257/21656 Group Owner: [email protected] Unsubscribe: https://lists.fd.io/g/vpp-dev/leave/14379924/21656/631435203/xyzzy [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
