dccp/tcp: do not inherit mc_list from parent
authorEric Dumazet <edumazet@google.com>
Tue, 9 May 2017 13:29:19 +0000 (06:29 -0700)
committerWilly Tarreau <w@1wt.eu>
Wed, 7 Jun 2017 22:47:11 +0000 (00:47 +0200)
commit 657831ffc38e30092a2d5f03d385d710eb88b09a upstream.

syzkaller found a way to trigger double frees from ip_mc_drop_socket()

It turns out that leave a copy of parent mc_list at accept() time,
which is very bad.

Very similar to commit 8b485ce69876 ("tcp: do not inherit
fastopen_req from parent")

Initial report from Pray3r, completed by Andrey one.
Thanks a lot to them !

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Pray3r <pray3r.z@gmail.com>
Reported-by: Andrey Konovalov <andreyknvl@google.com>
Tested-by: Andrey Konovalov <andreyknvl@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Willy Tarreau <w@1wt.eu>
net/ipv4/inet_connection_sock.c

index 6acb541c90910204f02449e7500138362da6998a..40ac1e2cbb34c377de67e5d8393a291b823dc5a6 100644 (file)
@@ -688,6 +688,8 @@ struct sock *inet_csk_clone_lock(const struct sock *sk,
                inet_sk(newsk)->inet_sport = inet_rsk(req)->loc_port;
                newsk->sk_write_space = sk_stream_write_space;
 
+               inet_sk(newsk)->mc_list = NULL;
+
                newicsk->icsk_retransmits = 0;
                newicsk->icsk_backoff     = 0;
                newicsk->icsk_probes_out  = 0;