[PATCH net] tcp: do not inherit mc_list from parent

[PATCH net] tcp: do not inherit mc_list from parent

[Eric Dumazet]

From: Eric Dumazet <edumazet@google.com>

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>
---
Notes:
 - day-0 bug.
 - Not sure if it makes sense for TCP socket to be able to join MC
group ?

 net/ipv4/tcp_minisocks.c |    1 +
 1 file changed, 1 insertion(+)

diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index 717be4de53248352c758b50557987d898340dd4f..03035e2857fc8b6e4cd8af6e46e81048d4de9105 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -448,6 +448,7 @@ struct sock *tcp_create_openreq_child(const struct sock *sk,
 		minmax_reset(&newtp->rtt_min, tcp_time_stamp, ~0U);
 		newicsk->icsk_rto = TCP_TIMEOUT_INIT;
 		newicsk->icsk_ack.lrcvtime = tcp_time_stamp;
+		newicsk->icsk_inet.mc_list = NULL;
 
 		newtp->packets_out = 0;
 		newtp->retrans_out = 0;

Re: [PATCH net] tcp: do not inherit mc_list from parent

[Eric Dumazet]

On Tue, 2017-05-09 at 05:17 -0700, Eric Dumazet wrote:
> From: Eric Dumazet <edumazet@google.com>
> 
> 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>
> ---
> Notes:
>  - day-0 bug.
>  - Not sure if it makes sense for TCP socket to be able to join MC
> group ?


I will send a V2, putting the fix in inet_csk_clone_lock() so that DCCP
is also fixed ;)

Re: [PATCH net] tcp: do not inherit mc_list from parent

[David Miller]

From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Tue, 09 May 2017 06:23:35 -0700

> I will send a V2, putting the fix in inet_csk_clone_lock() so that DCCP
> is also fixed ;)

Thank you :)

[PATCH v2 net] dccp/tcp: do not inherit mc_list from parent

[Eric Dumazet]

From: Eric Dumazet <edumazet@google.com>

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>
---
v2: fix moved into inet_csk_clone_lock() to fix both DCCP and TCP

 net/ipv4/inet_connection_sock.c |    2 ++
 1 file changed, 2 insertions(+)

diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
index 5e313c1ac94fc88eca5fe3a0e9e46e551e955ff0..1054d330bf9df3189a21dbb08e27c0e6ad136775 100644
--- a/net/ipv4/inet_connection_sock.c
+++ b/net/ipv4/inet_connection_sock.c
@@ -794,6 +794,8 @@ struct sock *inet_csk_clone_lock(const struct sock *sk,
 		/* listeners have SOCK_RCU_FREE, not the children */
 		sock_reset_flag(newsk, SOCK_RCU_FREE);
 
+		inet_sk(newsk)->mc_list = NULL;
+
 		newsk->sk_mark = inet_rsk(req)->ir_mark;
 		atomic64_set(&newsk->sk_cookie,
 			     atomic64_read(&inet_rsk(req)->ir_cookie));

Re: [PATCH v2 net] dccp/tcp: do not inherit mc_list from parent

[David Miller]

From: Eric Dumazet <eric.dumazet@gmail.com>
Date: Tue, 09 May 2017 06:29:19 -0700

> From: Eric Dumazet <edumazet@google.com>
> 
> 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>
> ---
> v2: fix moved into inet_csk_clone_lock() to fix both DCCP and TCP

Applied and queued up for -stable, thanks Eric.

Re: [PATCH v2 net] dccp/tcp: do not inherit mc_list from parent

[Cong Wang]

On Tue, May 9, 2017 at 6:29 AM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> From: Eric Dumazet <edumazet@google.com>
>
> 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>
> ---
> v2: fix moved into inet_csk_clone_lock() to fix both DCCP and TCP
>
>  net/ipv4/inet_connection_sock.c |    2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
> index 5e313c1ac94fc88eca5fe3a0e9e46e551e955ff0..1054d330bf9df3189a21dbb08e27c0e6ad136775 100644
> --- a/net/ipv4/inet_connection_sock.c
> +++ b/net/ipv4/inet_connection_sock.c
> @@ -794,6 +794,8 @@ struct sock *inet_csk_clone_lock(const struct sock *sk,
>                 /* listeners have SOCK_RCU_FREE, not the children */
>                 sock_reset_flag(newsk, SOCK_RCU_FREE);
>
> +               inet_sk(newsk)->mc_list = NULL;
> +
>                 newsk->sk_mark = inet_rsk(req)->ir_mark;
>                 atomic64_set(&newsk->sk_cookie,
>                              atomic64_read(&inet_rsk(req)->ir_cookie));
>

I think IPv6 needs this too?

diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index aeb9497..b3611d9 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -1062,6 +1062,7 @@ static struct sock *tcp_v6_syn_recv_sock(const
struct sock *sk, struct sk_buff *
                newtp->af_specific = &tcp_sock_ipv6_mapped_specific;
 #endif

+               newnp->ipv6_mc_list = NULL;
                newnp->ipv6_ac_list = NULL;
                newnp->ipv6_fl_list = NULL;
                newnp->pktoptions  = NULL;

Re: [PATCH v2 net] dccp/tcp: do not inherit mc_list from parent

[Eric Dumazet]

On Tue, 2017-05-09 at 15:37 -0700, Cong Wang wrote:
> On Tue, May 9, 2017 at 6:29 AM, Eric Dumazet <eric.dumazet@gmail.com> wrote:
> > From: Eric Dumazet <edumazet@google.com>
> >
> > 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>
> > ---
> > v2: fix moved into inet_csk_clone_lock() to fix both DCCP and TCP
> >
> >  net/ipv4/inet_connection_sock.c |    2 ++
> >  1 file changed, 2 insertions(+)
> >
> > diff --git a/net/ipv4/inet_connection_sock.c b/net/ipv4/inet_connection_sock.c
> > index 5e313c1ac94fc88eca5fe3a0e9e46e551e955ff0..1054d330bf9df3189a21dbb08e27c0e6ad136775 100644
> > --- a/net/ipv4/inet_connection_sock.c
> > +++ b/net/ipv4/inet_connection_sock.c
> > @@ -794,6 +794,8 @@ struct sock *inet_csk_clone_lock(const struct sock *sk,
> >                 /* listeners have SOCK_RCU_FREE, not the children */
> >                 sock_reset_flag(newsk, SOCK_RCU_FREE);
> >
> > +               inet_sk(newsk)->mc_list = NULL;
> > +
> >                 newsk->sk_mark = inet_rsk(req)->ir_mark;
> >                 atomic64_set(&newsk->sk_cookie,
> >                              atomic64_read(&inet_rsk(req)->ir_cookie));
> >
> 
> I think IPv6 needs this too?
> 
> diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
> index aeb9497..b3611d9 100644
> --- a/net/ipv6/tcp_ipv6.c
> +++ b/net/ipv6/tcp_ipv6.c
> @@ -1062,6 +1062,7 @@ static struct sock *tcp_v6_syn_recv_sock(const
> struct sock *sk, struct sk_buff *
>                 newtp->af_specific = &tcp_sock_ipv6_mapped_specific;
>  #endif
> 
> +               newnp->ipv6_mc_list = NULL;
>                 newnp->ipv6_ac_list = NULL;
>                 newnp->ipv6_fl_list = NULL;
>                 newnp->pktoptions  = NULL;

Good point, but it looks like you patched on only IPv4 mapped sockets.

And DCCP would need fixes as well.