lk/linux-kernel-test
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

ipmr: RCU conversion of mroute_sk

Browse Source 
Use RCU and RTNL to protect (struct mr_table)->mroute_sk

Readers use RCU, writers use RTNL.

ip_ra_control() already use an RCU grace period before
ip_ra_destroy_rcu(), so we dont need synchronize_rcu() in
mrtsock_destruct()

Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
Eric Dumazet
2010-10-01 16:15:01 +00:00
committed by David S. Miller
parent 55747a0a73
commit 4c9687098f
1 changed files with 49 additions and 42 deletions
Download Patch File Download Diff File Expand all files Collapse all files

91
net/ipv4/ipmr.c
View File

@@ -75,7 +75,7 @@ struct mr_table {
struct net *net; struct net *net;
#endif #endif
u32 id; u32 id;
struct sock *mroute_sk; struct sock __rcu *mroute_sk;
struct timer_list ipmr_expire_timer; struct timer_list ipmr_expire_timer;
struct list_head mfc_unres_queue; struct list_head mfc_unres_queue;
struct list_head mfc_cache_array[MFC_LINES]; struct list_head mfc_cache_array[MFC_LINES];
@@ -867,6 +867,7 @@ static int ipmr_cache_report(struct mr_table *mrt,
const int ihl = ip_hdrlen(pkt); const int ihl = ip_hdrlen(pkt);
struct igmphdr *igmp; struct igmphdr *igmp;
struct igmpmsg *msg; struct igmpmsg *msg;
struct sock *mroute_sk;
int ret; int ret;
#ifdef CONFIG_IP_PIMSM #ifdef CONFIG_IP_PIMSM
@@ -925,7 +926,10 @@ static int ipmr_cache_report(struct mr_table *mrt,
skb->transport_header = skb->network_header; skb->transport_header = skb->network_header;
} }
if (mrt->mroute_sk == NULL) { rcu_read_lock();
mroute_sk = rcu_dereference(mrt->mroute_sk);
if (mroute_sk == NULL) {
rcu_read_unlock();
kfree_skb(skb); kfree_skb(skb);
return -EINVAL; return -EINVAL;
} }
@@ -933,7 +937,8 @@ static int ipmr_cache_report(struct mr_table *mrt,
/* /*
* Deliver to mrouted * Deliver to mrouted
*/ */
ret = sock_queue_rcv_skb(mrt->mroute_sk, skb); ret = sock_queue_rcv_skb(mroute_sk, skb);
rcu_read_unlock();
if (ret < 0) { if (ret < 0) {
if (net_ratelimit()) if (net_ratelimit())
printk(KERN_WARNING "mroute: pending queue full, dropping entries.\n"); printk(KERN_WARNING "mroute: pending queue full, dropping entries.\n");
@@ -1164,6 +1169,9 @@ static void mroute_clean_tables(struct mr_table *mrt)
} }
} }
/* called from ip_ra_control(), before an RCU grace period,
* we dont need to call synchronize_rcu() here
*/
static void mrtsock_destruct(struct sock *sk) static void mrtsock_destruct(struct sock *sk)
{ {
struct net *net = sock_net(sk); struct net *net = sock_net(sk);
@@ -1171,13 +1179,9 @@ static void mrtsock_destruct(struct sock *sk)
rtnl_lock(); rtnl_lock();
ipmr_for_each_table(mrt, net) { ipmr_for_each_table(mrt, net) {
if (sk == mrt->mroute_sk) { if (sk == rtnl_dereference(mrt->mroute_sk)) {
IPV4_DEVCONF_ALL(net, MC_FORWARDING)--; IPV4_DEVCONF_ALL(net, MC_FORWARDING)--;
rcu_assign_pointer(mrt->mroute_sk, NULL);
write_lock_bh(&mrt_lock);
mrt->mroute_sk = NULL;
write_unlock_bh(&mrt_lock);
mroute_clean_tables(mrt); mroute_clean_tables(mrt);
} }
} }
@@ -1204,7 +1208,8 @@ int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, unsi
return -ENOENT; return -ENOENT;
if (optname != MRT_INIT) { if (optname != MRT_INIT) {
if (sk != mrt->mroute_sk && !capable(CAP_NET_ADMIN)) if (sk != rcu_dereference_raw(mrt->mroute_sk) &&
!capable(CAP_NET_ADMIN))
return -EACCES; return -EACCES;
} }
@@ -1217,23 +1222,20 @@ int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, unsi
return -ENOPROTOOPT; return -ENOPROTOOPT;
rtnl_lock(); rtnl_lock();
if (mrt->mroute_sk) { if (rtnl_dereference(mrt->mroute_sk)) {
rtnl_unlock(); rtnl_unlock();
return -EADDRINUSE; return -EADDRINUSE;
} }
ret = ip_ra_control(sk, 1, mrtsock_destruct); ret = ip_ra_control(sk, 1, mrtsock_destruct);
if (ret == 0) { if (ret == 0) {
write_lock_bh(&mrt_lock); rcu_assign_pointer(mrt->mroute_sk, sk);
mrt->mroute_sk = sk;
write_unlock_bh(&mrt_lock);
IPV4_DEVCONF_ALL(net, MC_FORWARDING)++; IPV4_DEVCONF_ALL(net, MC_FORWARDING)++;
} }
rtnl_unlock(); rtnl_unlock();
return ret; return ret;
case MRT_DONE: case MRT_DONE:
if (sk != mrt->mroute_sk) if (sk != rcu_dereference_raw(mrt->mroute_sk))
return -EACCES; return -EACCES;
return ip_ra_control(sk, 0, NULL); return ip_ra_control(sk, 0, NULL);
case MRT_ADD_VIF: case MRT_ADD_VIF:
@@ -1246,7 +1248,8 @@ int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, unsi
return -ENFILE; return -ENFILE;
rtnl_lock(); rtnl_lock();
if (optname == MRT_ADD_VIF) { if (optname == MRT_ADD_VIF) {
ret = vif_add(net, mrt, &vif, sk == mrt->mroute_sk); ret = vif_add(net, mrt, &vif,
sk == rtnl_dereference(mrt->mroute_sk));
} else { } else {
ret = vif_delete(mrt, vif.vifc_vifi, 0, NULL); ret = vif_delete(mrt, vif.vifc_vifi, 0, NULL);
} }
@@ -1267,7 +1270,8 @@ int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, unsi
if (optname == MRT_DEL_MFC) if (optname == MRT_DEL_MFC)
ret = ipmr_mfc_delete(mrt, &mfc); ret = ipmr_mfc_delete(mrt, &mfc);
else else
ret = ipmr_mfc_add(net, mrt, &mfc, sk == mrt->mroute_sk); ret = ipmr_mfc_add(net, mrt, &mfc,
sk == rtnl_dereference(mrt->mroute_sk));
rtnl_unlock(); rtnl_unlock();
return ret; return ret;
/* /*
@@ -1309,14 +1313,16 @@ int ip_mroute_setsockopt(struct sock *sk, int optname, char __user *optval, unsi
return -EINVAL; return -EINVAL;
if (get_user(v, (u32 __user *)optval)) if (get_user(v, (u32 __user *)optval))
return -EFAULT; return -EFAULT;
if (sk == mrt->mroute_sk)
return -EBUSY;
rtnl_lock(); rtnl_lock();
ret = 0; ret = 0;
if (!ipmr_new_table(net, v)) if (sk == rtnl_dereference(mrt->mroute_sk)) {
ret = -ENOMEM; ret = -EBUSY;
raw_sk(sk)->ipmr_table = v; } else {
if (!ipmr_new_table(net, v))
ret = -ENOMEM;
raw_sk(sk)->ipmr_table = v;
}
rtnl_unlock(); rtnl_unlock();
return ret; return ret;
} }
@@ -1713,6 +1719,7 @@ dont_forward:
/* /*
* Multicast packets for forwarding arrive here * Multicast packets for forwarding arrive here
* Called with rcu_read_lock();
*/ */
int ip_mr_input(struct sk_buff *skb) int ip_mr_input(struct sk_buff *skb)
@@ -1726,7 +1733,7 @@ int ip_mr_input(struct sk_buff *skb)
/* Packet is looped back after forward, it should not be /* Packet is looped back after forward, it should not be
forwarded second time, but still can be delivered locally. forwarded second time, but still can be delivered locally.
*/ */
if (IPCB(skb)->flags&IPSKB_FORWARDED) if (IPCB(skb)->flags & IPSKB_FORWARDED)
goto dont_forward; goto dont_forward;
err = ipmr_fib_lookup(net, &skb_rtable(skb)->fl, &mrt); err = ipmr_fib_lookup(net, &skb_rtable(skb)->fl, &mrt);
@@ -1736,24 +1743,24 @@ int ip_mr_input(struct sk_buff *skb)
} }
if (!local) { if (!local) {
if (IPCB(skb)->opt.router_alert) { if (IPCB(skb)->opt.router_alert) {
if (ip_call_ra_chain(skb)) if (ip_call_ra_chain(skb))
return 0; return 0;
} else if (ip_hdr(skb)->protocol == IPPROTO_IGMP){ } else if (ip_hdr(skb)->protocol == IPPROTO_IGMP) {
/* IGMPv1 (and broken IGMPv2 implementations sort of /* IGMPv1 (and broken IGMPv2 implementations sort of
Cisco IOS <= 11.2(8)) do not put router alert * Cisco IOS <= 11.2(8)) do not put router alert
option to IGMP packets destined to routable * option to IGMP packets destined to routable
groups. It is very bad, because it means * groups. It is very bad, because it means
that we can forward NO IGMP messages. * that we can forward NO IGMP messages.
*/ */
read_lock(&mrt_lock); struct sock *mroute_sk;
if (mrt->mroute_sk) {
nf_reset(skb); mroute_sk = rcu_dereference(mrt->mroute_sk);
raw_rcv(mrt->mroute_sk, skb); if (mroute_sk) {
read_unlock(&mrt_lock); nf_reset(skb);
return 0; raw_rcv(mroute_sk, skb);
} return 0;
read_unlock(&mrt_lock); }
} }
} }