[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CANn89iLXPQhCsnO6o6Kvaa3Hpftdp7k77Fsa3o9fns0xfXQWQQ@mail.gmail.com>
Date: Mon, 15 Jul 2024 08:04:46 -0700
From: Eric Dumazet <edumazet@...gle.com>
To: Kuniyuki Iwashima <kuniyu@...zon.com>
Cc: "David S. Miller" <davem@...emloft.net>, Jakub Kicinski <kuba@...nel.org>,
Paolo Abeni <pabeni@...hat.com>, David Ahern <dsahern@...nel.org>,
Dmitry Safonov <0x7f454c46@...il.com>, Kuniyuki Iwashima <kuni1840@...il.com>, netdev@...r.kernel.org,
syzkaller <syzkaller@...glegroups.com>
Subject: Re: [PATCH v1 net] tcp: Don't access uninit tcp_rsk(req)->ao_keyid in tcp_create_openreq_child().
On Sun, Jul 14, 2024 at 9:17 AM Kuniyuki Iwashima <kuniyu@...zon.com> wrote:
>
> syzkaller reported KMSAN splat in tcp_create_openreq_child(). [0]
>
> The uninit variable is tcp_rsk(req)->ao_keyid.
>
> tcp_rsk(req)->ao_keyid is initialised only when tcp_conn_request() finds
> a valid TCP AO option in SYN. Then, tcp_rsk(req)->used_tcp_ao is set
> accordingly.
>
> Let's not read tcp_rsk(req)->ao_keyid when tcp_rsk(req)->used_tcp_ao is
> false.
>
>
> Fixes: 06b22ef29591 ("net/tcp: Wire TCP-AO to request sockets")
> Reported-by: syzkaller <syzkaller@...glegroups.com>
> Signed-off-by: Kuniyuki Iwashima <kuniyu@...zon.com>
> ---
Reviewed-by: Eric Dumazet <edumazet@...gle.com>
Powered by blists - more mailing lists