xps: must clear sender_cpu before forwarding

John reported that my previous commit added a regression
on his router.

This is because sender_cpu & napi_id share a common location,
so get_xps_queue() can see garbage and perform an out of bound access.

We need to make sure sender_cpu is cleared before doing the transmit,
otherwise any NIC busy poll enabled (skb_mark_napi_id()) can trigger
this bug.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: John <jw@nuclearfallout.net>
Bisected-by: John <jw@nuclearfallout.net>
Fixes: 2bd82484bb4c ("xps: fix xps for stacked devices")
Signed-off-by: David S. Miller <davem@davemloft.net>

3 files changed, 3 insertions, 1 deletions

diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index f80507823531..434e78e5254d 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -4173,7 +4173,7 @@ void skb_scrub_packet(struct sk_buff *skb, bool xnet)
         skb->ignore_df = 0;
         skb_dst_drop(skb);
         skb->mark = 0;
-        skb->sender_cpu = 0;
+        skb_sender_cpu_clear(skb);
         skb_init_secmark(skb);
         secpath_reset(skb);
         nf_reset(skb);
diff --git a/net/ipv4/ip_forward.c b/net/ipv4/ip_forward.c
index 787b3c294ce6..d9bc28ac5d1b 100644
--- a/net/ipv4/ip_forward.c
+++ b/net/ipv4/ip_forward.c
@@ -67,6 +67,7 @@ static int ip_forward_finish(struct sk_buff *skb)
         if (unlikely(opt->optlen))
                 ip_forward_options(skb);
 
+        skb_sender_cpu_clear(skb);
         return dst_output(skb);
 }
 
diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index 0a04a37305d5..7e80b61b51ff 100644
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -318,6 +318,7 @@ static int ip6_forward_proxy_check(struct sk_buff *skb)
 
 static inline int ip6_forward_finish(struct sk_buff *skb)
 {
+        skb_sender_cpu_clear(skb);
         return dst_output(skb);
 }