netns xfrm: per-netns xfrm_state_all list

This is done to get
a) simple "something leaked" check
b) cover possible DoSes when other netns puts many, many xfrm_states
   onto a list.
c) not miss "alien xfrm_state" check in some of list iterators in future.

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

diff --git a/include/net/netns/xfrm.h b/include/net/netns/xfrm.h
index 1cb0024a3b47003cd6c3aeef48c1bc6f18846f82..6ae234a1651715c8deea9d6c01da7585cf3fef29 100644 (file)
--- a/include/net/netns/xfrm.h
+++ b/include/net/netns/xfrm.h
@@ -1,7 +1,10 @@
 #ifndef __NETNS_XFRM_H
 #define __NETNS_XFRM_H
 
+#include <linux/list.h>
+
 struct netns_xfrm {
+       struct list_head        state_all;
 };
 
 #endif

diff --git a/net/xfrm/xfrm_state.c b/net/xfrm/xfrm_state.c
index 81bde76d049c30a7a7dde62c4cbd0a8b247241b1..85bb85484b70a63c68e72c359615851d872a341a 100644 (file)
--- a/net/xfrm/xfrm_state.c
+++ b/net/xfrm/xfrm_state.c
@@ -50,7 +50,6 @@ static DEFINE_SPINLOCK(xfrm_state_lock);
  * Main use is finding SA after policy selected tunnel or transport mode.
  * Also, it can be used by ah/esp icmp error handler to find offending SA.
  */
-static LIST_HEAD(xfrm_state_all);
 static struct hlist_head *xfrm_state_bydst __read_mostly;
 static struct hlist_head *xfrm_state_bysrc __read_mostly;
 static struct hlist_head *xfrm_state_byspi __read_mostly;
@@ -855,7 +854,7 @@ xfrm_state_find(xfrm_address_t *daddr, xfrm_address_t *saddr,
 
                if (km_query(x, tmpl, pol) == 0) {
                        x->km.state = XFRM_STATE_ACQ;
-                       list_add(&x->km.all, &xfrm_state_all);
+                       list_add(&x->km.all, &init_net.xfrm.state_all);
                        hlist_add_head(&x->bydst, xfrm_state_bydst+h);
                        h = xfrm_src_hash(daddr, saddr, family);
                        hlist_add_head(&x->bysrc, xfrm_state_bysrc+h);
@@ -924,7 +923,7 @@ static void __xfrm_state_insert(struct xfrm_state *x)
 
        x->genid = ++xfrm_state_genid;
 
-       list_add(&x->km.all, &xfrm_state_all);
+       list_add(&x->km.all, &init_net.xfrm.state_all);
 
        h = xfrm_dst_hash(&x->id.daddr, &x->props.saddr,
                          x->props.reqid, x->props.family);
@@ -1053,7 +1052,7 @@ static struct xfrm_state *__find_acq_core(unsigned short family, u8 mode, u32 re
                xfrm_state_hold(x);
                x->timer.expires = jiffies + sysctl_xfrm_acq_expires*HZ;
                add_timer(&x->timer);
-               list_add(&x->km.all, &xfrm_state_all);
+               list_add(&x->km.all, &init_net.xfrm.state_all);
                hlist_add_head(&x->bydst, xfrm_state_bydst+h);
                h = xfrm_src_hash(daddr, saddr, family);
                hlist_add_head(&x->bysrc, xfrm_state_bysrc+h);
@@ -1559,10 +1558,10 @@ int xfrm_state_walk(struct xfrm_state_walk *walk,
 
        spin_lock_bh(&xfrm_state_lock);
        if (list_empty(&walk->all))
-               x = list_first_entry(&xfrm_state_all, struct xfrm_state_walk, all);
+               x = list_first_entry(&init_net.xfrm.state_all, struct xfrm_state_walk, all);
        else
                x = list_entry(&walk->all, struct xfrm_state_walk, all);
-       list_for_each_entry_from(x, &xfrm_state_all, all) {
+       list_for_each_entry_from(x, &init_net.xfrm.state_all, all) {
                if (x->state == XFRM_STATE_DEAD)
                        continue;
                state = container_of(x, struct xfrm_state, km);
@@ -2085,6 +2084,8 @@ int __net_init xfrm_state_init(struct net *net)
 {
        unsigned int sz;
 
+       INIT_LIST_HEAD(&net->xfrm.state_all);
+
        sz = sizeof(struct hlist_head) * 8;
 
        xfrm_state_bydst = xfrm_hash_alloc(sz);
@@ -2100,6 +2101,7 @@ int __net_init xfrm_state_init(struct net *net)
 
 void xfrm_state_fini(struct net *net)
 {
+       WARN_ON(!list_empty(&net->xfrm.state_all));
 }
 
 #ifdef CONFIG_AUDITSYSCALL