summaryrefslogtreecommitdiff
path: root/net/ipv4/icmp.c
diff options
context:
space:
mode:
authorPavel Emelyanov <xemul@openvz.org>2008-03-26 12:00:21 +0300
committerDavid S. Miller <davem@davemloft.net>2008-03-26 12:00:21 +0300
commitb34a95ee6e05579faff2672d7d30356576407653 (patch)
tree1587a59fcced1f774193b184301572ee0bc2dda7 /net/ipv4/icmp.c
parent68528f09980a60c8df046d16336333cac4fc0c32 (diff)
downloadlinux-b34a95ee6e05579faff2672d7d30356576407653.tar.xz
[NETNS][ICMP]: Use per-net sysctls in ipv4/icmp.c.
This mostly re-uses the net, used in icmp netnsization patches from Denis. After this ICMP sysctls are completely virtualized. Signed-off-by: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net/ipv4/icmp.c')
-rw-r--r--net/ipv4/icmp.c25
1 files changed, 16 insertions, 9 deletions
diff --git a/net/ipv4/icmp.c b/net/ipv4/icmp.c
index 958a38445120..7049b3f5f3cc 100644
--- a/net/ipv4/icmp.c
+++ b/net/ipv4/icmp.c
@@ -270,7 +270,8 @@ int xrlim_allow(struct dst_entry *dst, int timeout)
return rc;
}
-static inline int icmpv4_xrlim_allow(struct rtable *rt, int type, int code)
+static inline int icmpv4_xrlim_allow(struct net *net, struct rtable *rt,
+ int type, int code)
{
struct dst_entry *dst = &rt->u.dst;
int rc = 1;
@@ -287,8 +288,8 @@ static inline int icmpv4_xrlim_allow(struct rtable *rt, int type, int code)
goto out;
/* Limit if icmp type is enabled in ratemask. */
- if ((1 << type) & init_net.ipv4.sysctl_icmp_ratemask)
- rc = xrlim_allow(dst, init_net.ipv4.sysctl_icmp_ratelimit);
+ if ((1 << type) & net->ipv4.sysctl_icmp_ratemask)
+ rc = xrlim_allow(dst, net->ipv4.sysctl_icmp_ratelimit);
out:
return rc;
}
@@ -390,7 +391,7 @@ static void icmp_reply(struct icmp_bxm *icmp_param, struct sk_buff *skb)
if (ip_route_output_key(net, &rt, &fl))
goto out_unlock;
}
- if (icmpv4_xrlim_allow(rt, icmp_param->data.icmph.type,
+ if (icmpv4_xrlim_allow(net, rt, icmp_param->data.icmph.type,
icmp_param->data.icmph.code))
icmp_push_reply(icmp_param, &ipc, rt);
ip_rt_put(rt);
@@ -501,7 +502,7 @@ void icmp_send(struct sk_buff *skb_in, int type, int code, __be32 info)
struct net_device *dev = NULL;
if (rt->fl.iif &&
- init_net.ipv4.sysctl_icmp_errors_use_inbound_ifaddr)
+ net->ipv4.sysctl_icmp_errors_use_inbound_ifaddr)
dev = dev_get_by_index(net, rt->fl.iif);
if (dev) {
@@ -617,7 +618,7 @@ void icmp_send(struct sk_buff *skb_in, int type, int code, __be32 info)
}
route_done:
- if (!icmpv4_xrlim_allow(rt, type, code))
+ if (!icmpv4_xrlim_allow(net, rt, type, code))
goto ende;
/* RFC says return as much as we can without exceeding 576 bytes. */
@@ -723,7 +724,7 @@ static void icmp_unreach(struct sk_buff *skb)
* get the other vendor to fix their kit.
*/
- if (!init_net.ipv4.sysctl_icmp_ignore_bogus_error_responses &&
+ if (!net->ipv4.sysctl_icmp_ignore_bogus_error_responses &&
inet_addr_type(net, iph->daddr) == RTN_BROADCAST) {
if (net_ratelimit())
printk(KERN_WARNING "%u.%u.%u.%u sent an invalid ICMP "
@@ -818,7 +819,10 @@ out_err:
static void icmp_echo(struct sk_buff *skb)
{
- if (!init_net.ipv4.sysctl_icmp_echo_ignore_all) {
+ struct net *net;
+
+ net = skb->dst->dev->nd_net;
+ if (!net->ipv4.sysctl_icmp_echo_ignore_all) {
struct icmp_bxm icmp_param;
icmp_param.data.icmph = *icmp_hdr(skb);
@@ -1021,6 +1025,9 @@ int icmp_rcv(struct sk_buff *skb)
*/
if (rt->rt_flags & (RTCF_BROADCAST | RTCF_MULTICAST)) {
+ struct net *net;
+
+ net = rt->u.dst.dev->nd_net;
/*
* RFC 1122: 3.2.2.6 An ICMP_ECHO to broadcast MAY be
* silently ignored (we let user decide with a sysctl).
@@ -1029,7 +1036,7 @@ int icmp_rcv(struct sk_buff *skb)
*/
if ((icmph->type == ICMP_ECHO ||
icmph->type == ICMP_TIMESTAMP) &&
- init_net.ipv4.sysctl_icmp_echo_ignore_broadcasts) {
+ net->ipv4.sysctl_icmp_echo_ignore_broadcasts) {
goto error;
}
if (icmph->type != ICMP_ECHO &&