[PATCH net] ipv6: A few fixes on dereferencing rt->from

[PATCH net] ipv6: A few fixes on dereferencing rt->from

[Martin KaFai Lau]

It is a followup after the fix in
commit 9c69a1320515 ("route: Avoid crash from dereferencing NULL rt->from")

rt6_do_redirect():
1. NULL checking is needed on rt->from because a parallel
   fib6_info delete could happen that sets rt->from to NULL.
   (e.g. rt6_remove_exception() and fib6_drop_pcpu_from()).

2. fib6_info_hold() is not enough.  Same reason as (1).
   Meaning, holding dst->__refcnt cannot ensure
   rt->from is not NULL or rt->from->fib6_ref is not 0.

   Instead of using fib6_info_hold_safe() which ip6_rt_cache_alloc()
   is already doing, this patch chooses to extend the rcu section
   to keep "from" dereference-able after checking for NULL.

inet6_rtm_getroute():
1. NULL checking is also needed on rt->from for a similar reason.
   Note that inet6_rtm_getroute() is using RTNL_FLAG_DOIT_UNLOCKED.

Fixes: a68886a69180 ("net/ipv6: Make from in rt6_info rcu protected")
Signed-off-by: Martin KaFai Lau <kafai@fb.com>
---
 net/ipv6/route.c | 38 ++++++++++++++++++--------------------
 1 file changed, 18 insertions(+), 20 deletions(-)

diff --git a/net/ipv6/route.c b/net/ipv6/route.c
index b4899f0de0d0..73ef72c208af 100644
--- a/net/ipv6/route.c
+++ b/net/ipv6/route.c
@@ -3397,11 +3397,8 @@ static void rt6_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_bu
 
 	rcu_read_lock();
 	from = rcu_dereference(rt->from);
-	/* This fib6_info_hold() is safe here because we hold reference to rt
-	 * and rt already holds reference to fib6_info.
-	 */
-	fib6_info_hold(from);
-	rcu_read_unlock();
+	if (!from)
+		goto out;
 
 	nrt = ip6_rt_cache_alloc(from, &msg->dest, NULL);
 	if (!nrt)
@@ -3413,10 +3410,7 @@ static void rt6_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_bu
 
 	nrt->rt6i_gateway = *(struct in6_addr *)neigh->primary_key;
 
-	/* No need to remove rt from the exception table if rt is
-	 * a cached route because rt6_insert_exception() will
-	 * takes care of it
-	 */
+	/* rt6_insert_exception() will take care of duplicated exceptions */
 	if (rt6_insert_exception(nrt, from)) {
 		dst_release_immediate(&nrt->dst);
 		goto out;
@@ -3429,7 +3423,7 @@ static void rt6_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_bu
 	call_netevent_notifiers(NETEVENT_REDIRECT, &netevent);
 
 out:
-	fib6_info_release(from);
+	rcu_read_unlock();
 	neigh_release(neigh);
 }
 
@@ -5028,16 +5022,20 @@ static int inet6_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr *nlh,
 
 	rcu_read_lock();
 	from = rcu_dereference(rt->from);
-
-	if (fibmatch)
-		err = rt6_fill_node(net, skb, from, NULL, NULL, NULL, iif,
-				    RTM_NEWROUTE, NETLINK_CB(in_skb).portid,
-				    nlh->nlmsg_seq, 0);
-	else
-		err = rt6_fill_node(net, skb, from, dst, &fl6.daddr,
-				    &fl6.saddr, iif, RTM_NEWROUTE,
-				    NETLINK_CB(in_skb).portid, nlh->nlmsg_seq,
-				    0);
+	if (from) {
+		if (fibmatch)
+			err = rt6_fill_node(net, skb, from, NULL, NULL, NULL,
+					    iif, RTM_NEWROUTE,
+					    NETLINK_CB(in_skb).portid,
+					    nlh->nlmsg_seq, 0);
+		else
+			err = rt6_fill_node(net, skb, from, dst, &fl6.daddr,
+					    &fl6.saddr, iif, RTM_NEWROUTE,
+					    NETLINK_CB(in_skb).portid,
+					    nlh->nlmsg_seq, 0);
+	} else {
+		err = -ENETUNREACH;
+	}
 	rcu_read_unlock();
 
 	if (err < 0) {
-- 
2.17.1

Re: [PATCH net] ipv6: A few fixes on dereferencing rt->from

[Wei Wang]

On Tue, Apr 30, 2019 at 10:45 AM Martin KaFai Lau <kafai@fb.com> wrote:
>
> It is a followup after the fix in
> commit 9c69a1320515 ("route: Avoid crash from dereferencing NULL rt->from")
>
> rt6_do_redirect():
> 1. NULL checking is needed on rt->from because a parallel
>    fib6_info delete could happen that sets rt->from to NULL.
>    (e.g. rt6_remove_exception() and fib6_drop_pcpu_from()).
>
> 2. fib6_info_hold() is not enough.  Same reason as (1).
>    Meaning, holding dst->__refcnt cannot ensure
>    rt->from is not NULL or rt->from->fib6_ref is not 0.
>
>    Instead of using fib6_info_hold_safe() which ip6_rt_cache_alloc()
>    is already doing, this patch chooses to extend the rcu section
>    to keep "from" dereference-able after checking for NULL.
>
> inet6_rtm_getroute():
> 1. NULL checking is also needed on rt->from for a similar reason.
>    Note that inet6_rtm_getroute() is using RTNL_FLAG_DOIT_UNLOCKED.
>
> Fixes: a68886a69180 ("net/ipv6: Make from in rt6_info rcu protected")
> Signed-off-by: Martin KaFai Lau <kafai@fb.com>
> ---
Acked-by: Wei Wang <weiwan@google.com>

Nice fix. Thanks Martin.

>  net/ipv6/route.c | 38 ++++++++++++++++++--------------------
>  1 file changed, 18 insertions(+), 20 deletions(-)
>
> diff --git a/net/ipv6/route.c b/net/ipv6/route.c
> index b4899f0de0d0..73ef72c208af 100644
> --- a/net/ipv6/route.c
> +++ b/net/ipv6/route.c
> @@ -3397,11 +3397,8 @@ static void rt6_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_bu
>
>         rcu_read_lock();
>         from = rcu_dereference(rt->from);
> -       /* This fib6_info_hold() is safe here because we hold reference to rt
> -        * and rt already holds reference to fib6_info.
> -        */
> -       fib6_info_hold(from);
> -       rcu_read_unlock();
> +       if (!from)
> +               goto out;
>
>         nrt = ip6_rt_cache_alloc(from, &msg->dest, NULL);
>         if (!nrt)
> @@ -3413,10 +3410,7 @@ static void rt6_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_bu
>
>         nrt->rt6i_gateway = *(struct in6_addr *)neigh->primary_key;
>
> -       /* No need to remove rt from the exception table if rt is
> -        * a cached route because rt6_insert_exception() will
> -        * takes care of it
> -        */
> +       /* rt6_insert_exception() will take care of duplicated exceptions */
>         if (rt6_insert_exception(nrt, from)) {
>                 dst_release_immediate(&nrt->dst);
>                 goto out;
> @@ -3429,7 +3423,7 @@ static void rt6_do_redirect(struct dst_entry *dst, struct sock *sk, struct sk_bu
>         call_netevent_notifiers(NETEVENT_REDIRECT, &netevent);
>
>  out:
> -       fib6_info_release(from);
> +       rcu_read_unlock();
>         neigh_release(neigh);
>  }
>
> @@ -5028,16 +5022,20 @@ static int inet6_rtm_getroute(struct sk_buff *in_skb, struct nlmsghdr *nlh,
>
>         rcu_read_lock();
>         from = rcu_dereference(rt->from);
> -
> -       if (fibmatch)
> -               err = rt6_fill_node(net, skb, from, NULL, NULL, NULL, iif,
> -                                   RTM_NEWROUTE, NETLINK_CB(in_skb).portid,
> -                                   nlh->nlmsg_seq, 0);
> -       else
> -               err = rt6_fill_node(net, skb, from, dst, &fl6.daddr,
> -                                   &fl6.saddr, iif, RTM_NEWROUTE,
> -                                   NETLINK_CB(in_skb).portid, nlh->nlmsg_seq,
> -                                   0);
> +       if (from) {
> +               if (fibmatch)
> +                       err = rt6_fill_node(net, skb, from, NULL, NULL, NULL,
> +                                           iif, RTM_NEWROUTE,
> +                                           NETLINK_CB(in_skb).portid,
> +                                           nlh->nlmsg_seq, 0);
> +               else
> +                       err = rt6_fill_node(net, skb, from, dst, &fl6.daddr,
> +                                           &fl6.saddr, iif, RTM_NEWROUTE,
> +                                           NETLINK_CB(in_skb).portid,
> +                                           nlh->nlmsg_seq, 0);
> +       } else {
> +               err = -ENETUNREACH;
> +       }
>         rcu_read_unlock();
>
>         if (err < 0) {
> --
> 2.17.1
>

Re: [PATCH net] ipv6: A few fixes on dereferencing rt->from

[David Ahern]

On 4/30/19 11:45 AM, Martin KaFai Lau wrote:
> It is a followup after the fix in
> commit 9c69a1320515 ("route: Avoid crash from dereferencing NULL rt->from")
> 
> rt6_do_redirect():
> 1. NULL checking is needed on rt->from because a parallel
>    fib6_info delete could happen that sets rt->from to NULL.
>    (e.g. rt6_remove_exception() and fib6_drop_pcpu_from()).
> 
> 2. fib6_info_hold() is not enough.  Same reason as (1).
>    Meaning, holding dst->__refcnt cannot ensure
>    rt->from is not NULL or rt->from->fib6_ref is not 0.
> 
>    Instead of using fib6_info_hold_safe() which ip6_rt_cache_alloc()
>    is already doing, this patch chooses to extend the rcu section
>    to keep "from" dereference-able after checking for NULL.
> 
> inet6_rtm_getroute():
> 1. NULL checking is also needed on rt->from for a similar reason.
>    Note that inet6_rtm_getroute() is using RTNL_FLAG_DOIT_UNLOCKED.
> 
> Fixes: a68886a69180 ("net/ipv6: Make from in rt6_info rcu protected")
> Signed-off-by: Martin KaFai Lau <kafai@fb.com>
> ---
>  net/ipv6/route.c | 38 ++++++++++++++++++--------------------
>  1 file changed, 18 insertions(+), 20 deletions(-)
> 

Reviewed-by: David Ahern <dsahern@gmail.com>

Thanks, Martin.

Re: [PATCH net] ipv6: A few fixes on dereferencing rt->from

[Eric Dumazet]

On 4/30/19 10:45 AM, Martin KaFai Lau wrote:
> It is a followup after the fix in
> commit 9c69a1320515 ("route: Avoid crash from dereferencing NULL rt->from")
> 
> rt6_do_redirect():
> 1. NULL checking is needed on rt->from because a parallel
>    fib6_info delete could happen that sets rt->from to NULL.
>    (e.g. rt6_remove_exception() and fib6_drop_pcpu_from()).
> 
> 2. fib6_info_hold() is not enough.  Same reason as (1).
>    Meaning, holding dst->__refcnt cannot ensure
>    rt->from is not NULL or rt->from->fib6_ref is not 0.
> 
>    Instead of using fib6_info_hold_safe() which ip6_rt_cache_alloc()
>    is already doing, this patch chooses to extend the rcu section
>    to keep "from" dereference-able after checking for NULL.
> 
> inet6_rtm_getroute():
> 1. NULL checking is also needed on rt->from for a similar reason.
>    Note that inet6_rtm_getroute() is using RTNL_FLAG_DOIT_UNLOCKED.
> 
> Fixes: a68886a69180 ("net/ipv6: Make from in rt6_info rcu protected")
> Signed-off-by: Martin KaFai Lau <kafai@fb.com>
> ---
>

Reviewed-by: Eric Dumazet <edumazet@google.com>

Re: [PATCH net] ipv6: A few fixes on dereferencing rt->from

[David Miller]

From: Martin KaFai Lau <kafai@fb.com>
Date: Tue, 30 Apr 2019 10:45:12 -0700

> It is a followup after the fix in
> commit 9c69a1320515 ("route: Avoid crash from dereferencing NULL rt->from")
> 
> rt6_do_redirect():
> 1. NULL checking is needed on rt->from because a parallel
>    fib6_info delete could happen that sets rt->from to NULL.
>    (e.g. rt6_remove_exception() and fib6_drop_pcpu_from()).
> 
> 2. fib6_info_hold() is not enough.  Same reason as (1).
>    Meaning, holding dst->__refcnt cannot ensure
>    rt->from is not NULL or rt->from->fib6_ref is not 0.
> 
>    Instead of using fib6_info_hold_safe() which ip6_rt_cache_alloc()
>    is already doing, this patch chooses to extend the rcu section
>    to keep "from" dereference-able after checking for NULL.
> 
> inet6_rtm_getroute():
> 1. NULL checking is also needed on rt->from for a similar reason.
>    Note that inet6_rtm_getroute() is using RTNL_FLAG_DOIT_UNLOCKED.
> 
> Fixes: a68886a69180 ("net/ipv6: Make from in rt6_info rcu protected")
> Signed-off-by: Martin KaFai Lau <kafai@fb.com>

Applied and queued up for -stable, thanks.