[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <tencent_EECBD37DC379497A63A1C455B773377AC605@qq.com>
Date: Wed, 4 Sep 2024 08:52:19 +0800
From: Edward Adam Davis <eadavis@...com>
To: edumazet@...gle.com
Cc: davem@...emloft.net,
eadavis@...com,
geliang@...nel.org,
kuba@...nel.org,
linux-kernel@...r.kernel.org,
martineau@...nel.org,
matttbe@...nel.org,
mptcp@...ts.linux.dev,
netdev@...r.kernel.org,
pabeni@...hat.com,
syzbot+f3a31fb909db9b2a5c4d@...kaller.appspotmail.com,
syzkaller-bugs@...glegroups.com
Subject: Re: [PATCH] mptcp: pm: Fix uaf in __timer_delete_sync
On Tue, 3 Sep 2024 17:18:42 +0200, Eric Dumazet <edumazet@...gle.com> wrote:
>On Tue, Sep 3, 2024 at 5:10 PM Edward Adam Davis <eadavis@...com> wrote:
>>
>> There are two paths to access mptcp_pm_del_add_timer, result in a race
>> condition:
>>
>> CPU1 CPU2
>> ==== ====
>> net_rx_action
>> napi_poll netlink_sendmsg
>> __napi_poll netlink_unicast
>> process_backlog netlink_unicast_kernel
>> __netif_receive_skb genl_rcv
>> __netif_receive_skb_one_core netlink_rcv_skb
>> NF_HOOK genl_rcv_msg
>> ip_local_deliver_finish genl_family_rcv_msg
>> ip_protocol_deliver_rcu genl_family_rcv_msg_doit
>> tcp_v4_rcv mptcp_pm_nl_flush_addrs_doit
>> tcp_v4_do_rcv mptcp_nl_remove_addrs_list
>> tcp_rcv_established mptcp_pm_remove_addrs_and_subflows
>> tcp_data_queue remove_anno_list_by_saddr
>> mptcp_incoming_options mptcp_pm_del_add_timer
>> mptcp_pm_del_add_timer kfree(entry)
>>
>> In remove_anno_list_by_saddr(running on CPU2), after leaving the critical
>> zone protected by "pm.lock", the entry will be released, which leads to the
>> occurrence of uaf in the mptcp_pm_del_add_timer(running on CPU1).
>>
>> Reported-and-tested-by: syzbot+f3a31fb909db9b2a5c4d@...kaller.appspotmail.com
>> Closes: https://syzkaller.appspot.com/bug?extid=f3a31fb909db9b2a5c4d
>> Signed-off-by: Edward Adam Davis <eadavis@...com>
>> ---
>> net/mptcp/pm_netlink.c | 3 ++-
>> 1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c
>> index 3e4ad801786f..d28bf0c9ad66 100644
>> --- a/net/mptcp/pm_netlink.c
>> +++ b/net/mptcp/pm_netlink.c
>> @@ -336,11 +336,12 @@ mptcp_pm_del_add_timer(struct mptcp_sock *msk,
>> entry = mptcp_lookup_anno_list_by_saddr(msk, addr);
>> if (entry && (!check_id || entry->addr.id == addr->id))
>> entry->retrans_times = ADD_ADDR_RETRANS_MAX;
>> - spin_unlock_bh(&msk->pm.lock);
>>
>> if (entry && (!check_id || entry->addr.id == addr->id))
>> sk_stop_timer_sync(sk, &entry->add_timer);
>>
>> + spin_unlock_bh(&msk->pm.lock);
>
>
>mptcp_pm_add_timer() needs to lock msk->pm.lock
>
>Your patch might add a deadlock, because sk_stop_timer_sync() is
Got it. MPTCP CI reported it.
>calling del_timer_sync()
>
>What is preventing this ?
I will change the strategy for protecting entry and use pm.lock to
synchronize when it is released in remove_anno_list_by_saddr.
Link: https://syzkaller.appspot.com/text?tag=Patch&x=124282ab980000
BR,
Edward
Powered by blists - more mailing lists