Skip to content

icmp_socket namespacing

Denis V. Lunev containers netdev 1 version In review

Get series with b4
b4 am [email protected]

Versions newest first

  1. v1 Latest12 patches 7 repliesIn review

Patches in v1 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/12: Remove struct net_proto_family* from _init calls.include/linux/icmpv6.h include/net/icmp.h include/net/ndisc.h +8 more+15 -150 replies
2/12: Add return code to icmp_init.include/net/icmp.h net/ipv4/af_inet.c net/ipv4/icmp.c+25 -60 replies
3/12: Optimize icmp_socket usage.net/ipv4/icmp.c+17 -120 replies
4/12: Store sock rather than socket for ICMP flow control.net/ipv4/icmp.c net/ipv6/icmp.c+26 -260 replies
5/12: Pass proper ICMP socket into icmp(v6)_xmit_(un)lock.net/ipv4/icmp.c net/ipv6/icmp.c+22 -210 replies
6/12: Allocate data for __icmp(v6)_sk dynamically.net/ipv4/icmp.c net/ipv6/icmp.c+19 -100 replies
7/12No need for a separate __netlink_release call.net/netlink/af_netlink.c+12 -180 replies
8/12Make netlink_kernel_release publically available as sk_release_kernel.include/net/sock.h net/core/sock.c net/netlink/af_netlink.c+33 -160 replies
9/12: icmp(v6)_sk should not pin a namespace.net/ipv4/icmp.c net/ipv6/icmp.c+9 -141 reply
10/12: Make icmp_sk per namespace.include/net/netns/ipv4.h net/ipv4/icmp.c+34 -170 replies
11/12: Make icmpv6_sk per namespace.include/net/netns/ipv6.h net/ipv6/icmp.c+48 -210 replies
12/12: Consolidate fail path icmpv6_sk_init with icmpv6_sk_exit.net/ipv6/icmp.c+14 -155 replies

Reviews trailers per patch

All 12 patches have a Reviewed-by or Acked-by.

PatchDaniel Lezcano
1/12A-b
2/12A-b
3/12A-b
4/12A-b
5/12A-b
6/12A-b
7/12A-b
8/12A-b
9/12A-b
10/12A-b
11/12A-b
12/12A-b

Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.

Read the v1 thread

Cover letter [PATCH 0/12 net-2.6.26] icmp_socket namespacing

This set of patches is targeted to create separate icmp_socket inside each namespace. Both IPv4 and IPv6 codepaths are affected.

Though, in order to do this smoothly, a bit of optimisations are performed. The kernel from now on will use sock rather than socket on the ICMP send path.