[PATCH net-next] skbuff: revert "skbuff: remove some unnecessary operation in skb_segment_list()"

[Paolo Abeni]

the commit 1ddc3229ad3c ("skbuff: remove some unnecessary operation
in skb_segment_list()") introduces an issue very similar to the
one already fixed by commit 53475c5dd856 ("net: fix use-after-free when
UDP GRO with shared fraglist").

If the GSO skb goes though skb_clone() and pskb_expand_head() before
entering skb_segment_list(), the latter  will unshare the frag_list
skbs and will release the old list. With the reverted commit in place,
when skb_segment_list() completes, skb->next points to the just
released list, and later on the kernel will hit UaF.

Note that since commit e0e3070a9bc9 ("udp: properly complete L4 GRO
over UDP tunnel packet") the critical scenario can be reproduced also
receiving UDP over vxlan traffic with:

NIC (NETIF_F_GRO_FRAGLIST enabled) -> vxlan -> UDP sink

Attaching a packet socket to the NIC will cause skb_clone() and the
tunnel decapsulation will call pskb_expand_head().

Fixes: 1ddc3229ad3c ("skbuff: remove some unnecessary operation in 
skb_segment_list()")
Signed-off-by: Paolo Abeni <pab...@redhat.com>
---
 net/core/skbuff.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 3ad9e8425ab2..14010c0eec48 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -3773,13 +3773,13 @@ struct sk_buff *skb_segment_list(struct sk_buff *skb,
        unsigned int tnl_hlen = skb_tnl_header_len(skb);
        unsigned int delta_truesize = 0;
        unsigned int delta_len = 0;
+       struct sk_buff *tail = NULL;
        struct sk_buff *nskb, *tmp;
        int err;
 
        skb_push(skb, -skb_network_offset(skb) + offset);
 
        skb_shinfo(skb)->frag_list = NULL;
-       skb->next = list_skb;
 
        do {
                nskb = list_skb;
@@ -3797,8 +3797,17 @@ struct sk_buff *skb_segment_list(struct sk_buff *skb,
                        }
                }
 
-               if (unlikely(err))
+               if (!tail)
+                       skb->next = nskb;
+               else
+                       tail->next = nskb;
+
+               if (unlikely(err)) {
+                       nskb->next = list_skb;
                        goto err_linearize;
+               }
+
+               tail = nskb;
 
                delta_len += nskb->len;
                delta_truesize += nskb->truesize;
@@ -3825,7 +3834,7 @@ struct sk_buff *skb_segment_list(struct sk_buff *skb,
 
        skb_gso_reset(skb);
 
-       skb->prev = nskb;
+       skb->prev = tail;
 
        if (skb_needs_linearize(skb, features) &&
            __skb_linearize(skb))
-- 
2.26.2