linux.git/net/ipv4/ipip.c, branch v2.6.37

tunnels: Fix tunnels change rcu protection

2010-10-27T21:20:08+00:00

After making rcu protection for tunnels (ipip, gre, sit and ip6) a bug
was introduced into the SIOCCHGTUNNEL code.

The tunnel is first unlinked, then addresses change, then it is linked
back probably into another bucket. But while changing the parms, the
hash table is unlocked to readers and they can lookup the improper tunnel.

Respective commits are b7285b79 (ipip: get rid of ipip_lock), 1507850b
(gre: get rid of ipgre_lock), 3a43be3c (sit: get rid of ipip6_lock) and
94767632 (ip6tnl: get rid of ip6_tnl_lock).

The quick fix is to wait for quiescent state to pass after unlinking,
but if it is inappropriate I can invent something better, just let me
know.

Signed-off-by: Pavel Emelyanov 
Acked-by: Eric Dumazet 
Signed-off-by: David S. Miller

net: add a core netdev->rx_dropped counter

2010-10-05T21:47:55+00:00

In various situations, a device provides a packet to our stack and we
drop it before it enters protocol stack :
- softnet backlog full (accounted in /proc/net/softnet_stat)
- bad vlan tag (not accounted)
- unknown/unregistered protocol (not accounted)

We can handle a per-device counter of such dropped frames at core level,
and automatically adds it to the device provided stats (rx_dropped), so
that standard tools can be used (ifconfig, ip link, cat /proc/net/dev)

This is a generalization of commit 8990f468a (net: rx_dropped
accounting), thus reverting it.

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

ipip: enable lockless xmits

2010-09-29T20:25:44+00:00

IPIP tunnels can benefit from lockless xmits, using NETIF_F_LLTX

Bench on a 16 cpus machine (dual E5540 cpus), 16 threads sending
10000000 UDP frames via one ipip tunnel (size:200 bytes per frame)

Before patch :
real	2m53.321s
user	0m10.277s
sys	46m0.597s

After patch:
real	0m32.063s
user	0m9.237s
sys	8m16.255s

Last problem to solve is the contention on dst :

16118.00 28.3% __ip_route_output_key         vmlinux
 6135.00 10.8% dst_release                   vmlinux
 3220.00  5.6% ip_finish_output              vmlinux
 2149.00  3.8% ip_route_output_flow          vmlinux
 1575.00  2.8% ip_append_data                vmlinux
 1481.00  2.6% ip_push_pending_frames        vmlinux
 1349.00  2.4% __xfrm_lookup                 vmlinux
 1216.00  2.1% csum_partial_copy_generic     vmlinux
 1208.00  2.1% udp_sendmsg                   vmlinux

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

ipip: fix percpu stats accounting

2010-09-29T20:23:50+00:00

commit 3c97af99a5aa1 (ipip: percpu stats accounting) forgot the fallback
tunnel case (tunl0), and can crash pretty fast.

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

ipip: percpu stats accounting

2010-09-28T04:30:43+00:00

Maintain per_cpu tx_bytes, tx_packets, rx_bytes, rx_packets.

Other seldom used fields are kept in netdev->stats structure, possibly
unsafe.

This is a preliminary work to support lockless transmit path, and
correct RX stats, that are already unsafe.

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

net: rx_dropped accounting

2010-09-20T17:08:58+00:00

Under load, netif_rx() can drop incoming packets but administrators dont
have a chance to spot which device needs some tuning (RPS activation for
example)

This patch adds rx_dropped accounting in vlans and tunnels.

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

ipip: get rid of ipip_lock

2010-09-16T02:29:46+00:00

As RTNL is held while doing tunnels inserts and deletes, we can remove
ipip_lock spinlock. My initial RCU conversion was conservative and
converted the rwlock to spinlock, with no RTNL requirement.

Use appropriate rcu annotations and modern lockdep checks as well.

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

net: struct xfrm_tunnel in read_mostly section

2010-08-30T20:50:45+00:00

tunnel4_handlers chain being scanned for each incoming packet,
make sure it doesnt share an often dirtied cache line.

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

net-next: remove useless union keyword

2010-06-11T06:31:35+00:00

remove useless union keyword in rtable, rt6_info and dn_route.

Since there is only one member in a union, the union keyword isn't useful.

Signed-off-by: Changli Gao 
Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller

net: Introduce skb_tunnel_rx() helper

2010-05-18T05:36:55+00:00

skb rxhash should be cleared when a skb is handled by a tunnel before
being delivered again, so that correct packet steering can take place.

There are other cleanups and accounting that we can factorize in a new
helper, skb_tunnel_rx()

Signed-off-by: Eric Dumazet 
Signed-off-by: David S. Miller