diff options
| author | Hans Schillstrom | 2010-11-19 14:25:13 +0100 |
|---|---|---|
| committer | Simon Horman | 2010-11-25 02:42:59 +0100 |
| commit | 986a075795339c5ea1122ce9290dfd5504252eb0 (patch) | |
| tree | dc908479b0259f311892a9e46d865207ce0dbedc /net/netfilter/ipvs/ip_vs_core.c | |
| parent | IPVS: Backup, Adding Version 1 receive capability (diff) | |
| download | kernel-qcow2-linux-986a075795339c5ea1122ce9290dfd5504252eb0.tar.gz kernel-qcow2-linux-986a075795339c5ea1122ce9290dfd5504252eb0.tar.xz kernel-qcow2-linux-986a075795339c5ea1122ce9290dfd5504252eb0.zip | |
IPVS: Backup, Change sending to Version 1 format
Enable sending and removal of version 0 sending
Affected functions,
ip_vs_sync_buff_create()
ip_vs_sync_conn()
ip_vs_core.c removal of IPv4 check.
*v5
Just check cp->pe_data_len in ip_vs_sync_conn
Check if padding needed before adding a new sync_conn
to the buffer, i.e. avoid sending padding at the end.
*v4
moved sanity check and pe_name_len after sloop.
use cp->pe instead of cp->dest->svc->pe
real length in each sync_conn, not padded length
however total size of a sync_msg includes padding.
*v3
Sending ip_vs_sync_conn_options in network order.
Sending Templates for ONE_PACKET conn.
Renaming of ip_vs_sync_mesg to ip_vs_sync_mesg_v0
Signed-off-by: Hans Schillstrom <hans.schillstrom@ericsson.com>
Acked-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Simon Horman <horms@verge.net.au>
Diffstat (limited to 'net/netfilter/ipvs/ip_vs_core.c')
| -rw-r--r-- | net/netfilter/ipvs/ip_vs_core.c | 13 |
1 files changed, 9 insertions, 4 deletions
diff --git a/net/netfilter/ipvs/ip_vs_core.c b/net/netfilter/ipvs/ip_vs_core.c index 3445da6e8c95..5287771d0647 100644 --- a/net/netfilter/ipvs/ip_vs_core.c +++ b/net/netfilter/ipvs/ip_vs_core.c @@ -1560,9 +1560,15 @@ ip_vs_in(unsigned int hooknum, struct sk_buff *skb, int af) * * Sync connection if it is about to close to * encorage the standby servers to update the connections timeout + * + * For ONE_PKT let ip_vs_sync_conn() do the filter work. */ - pkts = atomic_add_return(1, &cp->in_pkts); - if (af == AF_INET && (ip_vs_sync_state & IP_VS_STATE_MASTER) && + if (cp->flags & IP_VS_CONN_F_ONE_PACKET) + pkts = sysctl_ip_vs_sync_threshold[0]; + else + pkts = atomic_add_return(1, &cp->in_pkts); + + if ((ip_vs_sync_state & IP_VS_STATE_MASTER) && cp->protocol == IPPROTO_SCTP) { if ((cp->state == IP_VS_SCTP_S_ESTABLISHED && (pkts % sysctl_ip_vs_sync_threshold[1] @@ -1577,8 +1583,7 @@ ip_vs_in(unsigned int hooknum, struct sk_buff *skb, int af) } /* Keep this block last: TCP and others with pp->num_states <= 1 */ - else if (af == AF_INET && - (ip_vs_sync_state & IP_VS_STATE_MASTER) && + else if ((ip_vs_sync_state & IP_VS_STATE_MASTER) && (((cp->protocol != IPPROTO_TCP || cp->state == IP_VS_TCP_S_ESTABLISHED) && (pkts % sysctl_ip_vs_sync_threshold[1] |