Skip to contentSkip to conversation

ipv6: prevent possible fib6 leaks

PATCH In review Eric Dumazet 5 participants 5 messages May 16, 2019

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ipv6: prevent possible fib6 leaks

At ipv6 route dismantle, fib6_drop_pcpu_from() is responsible for finding all percpu routes and set their ->from pointer to NULL, so that fib6_ref can reach its expected value (1).

The problem right now is that other cpus can still catch the route being deleted, since there is no rcu grace period between the route deletion and call to fib6_drop_pcpu_from()

This can leak the fib6 and associated resources, since no notifier will take care of removing the last reference(s).

I decided to add another boolean (fib6_destroying) instead of reusing/renaming exception_bucket_flushed to ease stable backports, and properly document the memory barriers used to implement this fix.

This patch has been co-developped with Wei Wang.

Fixes 93531c674315 ("net/ipv6: separate handling of FIB entries from dst based routes")
Cc Wei Wang <[email protected]>
Cc David Ahern <[email protected]>
Cc Martin Lau <[email protected]>
---
 include/net/ip6_fib.h |  3 ++-
 net/ipv6/ip6_fib.c    | 12 +++++++++---
 net/ipv6/route.c      |  7 +++++++
 3 files changed, 18 insertions(+), 4 deletions(-)
include/net/ip6_fib.h +2 -1
index 40105738e2f6b8e37adac1ff46879ce6c09381b8..525f701653ca69596b941f5f3b4a438d634c4e6c 100644
@@ -167,7 +167,8 @@ struct fib6_info {
 					dst_nocount:1,
 					dst_nopolicy:1,
 					dst_host:1,
-					unused:3;
+					fib6_destroying:1,
+					unused:2;
 
 	struct fib6_nh			fib6_nh;
 	struct rcu_head			rcu;
net/ipv6/ip6_fib.c +9 -3
index 08e0390e001c270ae21013f3fd3ef3bf2a52e039..008421b550c6bfd449665aa5e7ba5505fcabe53d 100644
@@ -904,6 +904,12 @@ static void fib6_drop_pcpu_from(struct fib6_info *f6i,
 {
 	int cpu;
 
+	/* Make sure rt6_make_pcpu_route() wont add other percpu routes
+	 * while we are cleaning them here.
+	 */
+	f6i->fib6_destroying = 1;
+	mb(); /* paired with the cmpxchg() in rt6_make_pcpu_route() */
+
 	/* release the reference to this fib entry from
 	 * all of its cached pcpu routes
 	 */
@@ -927,6 +933,9 @@ static void fib6_purge_rt(struct fib6_info *rt, struct fib6_node *fn,
 {
 	struct fib6_table *table = rt->fib6_table;
 
+	if (rt->rt6i_pcpu)
+		fib6_drop_pcpu_from(rt, table);
+
 	if (refcount_read(&rt->fib6_ref) != 1) {
 		/* This route is used as dummy address holder in some split
 		 * nodes. It is not leaked, but it still holds other resources,
@@ -948,9 +957,6 @@ static void fib6_purge_rt(struct fib6_info *rt, struct fib6_node *fn,
 			fn = rcu_dereference_protected(fn->parent,
 				    lockdep_is_held(&table->tb6_lock));
 		}
-
-		if (rt->rt6i_pcpu)
-			fib6_drop_pcpu_from(rt, table);
 	}
 }
 
net/ipv6/route.c +7 -0
index 23a20d62daac29e3252725b8cf95d1d1c2b567c4..27c0cc5d9d30e3689ebe6b8428cd4c586669d808 100644
@@ -1295,6 +1295,13 @@ static struct rt6_info *rt6_make_pcpu_route(struct net *net,
 	prev = cmpxchg(p, NULL, pcpu_rt);
 	BUG_ON(prev);
 
+	if (res->f6i->fib6_destroying) {
+		struct fib6_info *from;
+
+		from = xchg((__force struct fib6_info **)&pcpu_rt->from, NULL);
+		fib6_info_release(from);
+	}
+
 	return pcpu_rt;
 }
 
Signature · 1 line
2.21.0.1020.gf2820cf01a-goog
Acked-by on v1 1/1
#1 lore raw
Quoted 18 lines · Eric Dumazet

At ipv6 route dismantle, fib6_drop_pcpu_from() is responsible for finding all percpu routes and set their ->from pointer to NULL, so that fib6_ref can reach its expected value (1).

The problem right now is that other cpus can still catch the route being deleted, since there is no rcu grace period between the route deletion and call to fib6_drop_pcpu_from()

This can leak the fib6 and associated resources, since no notifier will take care of removing the last reference(s).

I decided to add another boolean (fib6_destroying) instead of reusing/renaming exception_bucket_flushed to ease stable backports, and properly document the memory barriers used to implement this fix.

This patch has been co-developped with Wei Wang.

Quoted 7 lines
Fixes 93531c674315 ("net/ipv6: separate handling of FIB entries from dst based routes")
Cc Wei Wang <[email protected]>
Cc David Ahern <[email protected]>
Cc Martin Lau <[email protected]>

---

Thanks for the fix Eric.

Quoted 77 lines
 include/net/ip6_fib.h |  3 ++-
 net/ipv6/ip6_fib.c    | 12 +++++++++---
 net/ipv6/route.c      |  7 +++++++
 3 files changed, 18 insertions(+), 4 deletions(-)
include/net/ip6_fib.h +2 -1
index 40105738e2f6b8e37adac1ff46879ce6c09381b8..525f701653ca69596b941f5f3b4a438d634c4e6c 100644
@@ -167,7 +167,8 @@ struct fib6_info {
                                        dst_nocount:1,
                                        dst_nopolicy:1,
                                        dst_host:1,
-                                       unused:3;
+                                       fib6_destroying:1,
+                                       unused:2;
 
        struct fib6_nh                  fib6_nh;
        struct rcu_head                 rcu;
net/ipv6/ip6_fib.c +9 -3
index 08e0390e001c270ae21013f3fd3ef3bf2a52e039..008421b550c6bfd449665aa5e7ba5505fcabe53d 100644
@@ -904,6 +904,12 @@ static void fib6_drop_pcpu_from(struct fib6_info *f6i,
 {
        int cpu;
 
+       /* Make sure rt6_make_pcpu_route() wont add other percpu routes
+        * while we are cleaning them here.
+        */
+       f6i->fib6_destroying = 1;
+       mb(); /* paired with the cmpxchg() in rt6_make_pcpu_route() */
+
        /* release the reference to this fib entry from
         * all of its cached pcpu routes
         */
@@ -927,6 +933,9 @@ static void fib6_purge_rt(struct fib6_info *rt, struct fib6_node *fn,
 {
        struct fib6_table *table = rt->fib6_table;
 
+       if (rt->rt6i_pcpu)
+               fib6_drop_pcpu_from(rt, table);
+
        if (refcount_read(&rt->fib6_ref) != 1) {
                /* This route is used as dummy address holder in some split
                 * nodes. It is not leaked, but it still holds other resources,
@@ -948,9 +957,6 @@ static void fib6_purge_rt(struct fib6_info *rt, struct fib6_node *fn,
                        fn = rcu_dereference_protected(fn->parent,
                                    lockdep_is_held(&table->tb6_lock));
                }
-
-               if (rt->rt6i_pcpu)
-                       fib6_drop_pcpu_from(rt, table);
        }
 }
 
net/ipv6/route.c +7 -1
index 23a20d62daac29e3252725b8cf95d1d1c2b567c4..27c0cc5d9d30e3689ebe6b8428cd4c586669d808 100644
@@ -1295,6 +1295,13 @@ static struct rt6_info *rt6_make_pcpu_route(struct net *net,
        prev = cmpxchg(p, NULL, pcpu_rt);
        BUG_ON(prev);
 
+       if (res->f6i->fib6_destroying) {
+               struct fib6_info *from;
+
+               from = xchg((__force struct fib6_info **)&pcpu_rt->from, NULL);
+               fib6_info_release(from);
+       }
+
        return pcpu_rt;
 }
 
--

2.21.0.1020.gf2820cf01a-goog

Acked-by on v1 1/1
#2 lore raw
In reply to Eric Dumazet
Quoted 8 lines · Eric Dumazet

At ipv6 route dismantle, fib6_drop_pcpu_from() is responsible for finding all percpu routes and set their ->from pointer to NULL, so that fib6_ref can reach its expected value (1).

The problem right now is that other cpus can still catch the route being deleted, since there is no rcu grace period between the route deletion and call to fib6_drop_pcpu_from()

Quoted 8 lines

This can leak the fib6 and associated resources, since no notifier will take care of removing the last reference(s).

I decided to add another boolean (fib6_destroying) instead of reusing/renaming exception_bucket_flushed to ease stable backports, and properly document the memory barriers used to implement this fix.

This patch has been co-developped with Wei Wang.

Nice fix!

Reviewed-by on v1 1/1
#3 lore raw
In reply to Eric Dumazet
Quoted 23 lines · Eric Dumazet

At ipv6 route dismantle, fib6_drop_pcpu_from() is responsible for finding all percpu routes and set their ->from pointer to NULL, so that fib6_ref can reach its expected value (1).

The problem right now is that other cpus can still catch the route being deleted, since there is no rcu grace period between the route deletion and call to fib6_drop_pcpu_from()

This can leak the fib6 and associated resources, since no notifier will take care of removing the last reference(s).

I decided to add another boolean (fib6_destroying) instead of reusing/renaming exception_bucket_flushed to ease stable backports, and properly document the memory barriers used to implement this fix.

This patch has been co-developped with Wei Wang.

Fixes 93531c674315 ("net/ipv6: separate handling of FIB entries from dst based routes")
Cc Wei Wang <[email protected]>
Cc David Ahern <[email protected]>
Cc Martin Lau <[email protected]>
Quoted 6 lines
---
 include/net/ip6_fib.h |  3 ++-
 net/ipv6/ip6_fib.c    | 12 +++++++++---
 net/ipv6/route.c      |  7 +++++++
 3 files changed, 18 insertions(+), 4 deletions(-)

Thanks, Eric. Attaching 'from' to rt6_info makes v6 much more complicated than ipv4.

maintainer
#4 lore raw
In reply to Eric Dumazet

From: Eric Dumazet <[email protected]>
Date: Wed, 15 May 2019 19:39:52 -0700

Quoted 15 lines

At ipv6 route dismantle, fib6_drop_pcpu_from() is responsible for finding all percpu routes and set their ->from pointer to NULL, so that fib6_ref can reach its expected value (1).

The problem right now is that other cpus can still catch the route being deleted, since there is no rcu grace period between the route deletion and call to fib6_drop_pcpu_from()

This can leak the fib6 and associated resources, since no notifier will take care of removing the last reference(s).

I decided to add another boolean (fib6_destroying) instead of reusing/renaming exception_bucket_flushed to ease stable backports, and properly document the memory barriers used to implement this fix.

Quoted 5 lines

This patch has been co-developped with Wei Wang.

Fixes 93531c674315 ("net/ipv6: separate handling of FIB entries from dst based routes")

Applied and queued up for -stable.