Re: [tipc-discussion] [PATCH net] tipc: wait and exit until all work queues are done

[Jon M. <jm...@re...>]

On 5/14/21 2:40 PM, Xin Long wrote:
> On some host, a crash could be triggered simply by repeating these
> commands several times:
>
>    # modprobe tipc
>    # tipc bearer enable media udp name UDP1 localip 127.0.0.1
>    # rmmod tipc
>
>    [] BUG: unable to handle kernel paging request at ffffffffc096bb00
>    [] Workqueue: events 0xffffffffc096bb00
>    [] Call Trace:
>    []  ? process_one_work+0x1a7/0x360
>    []  ? worker_thread+0x30/0x390
>    []  ? create_worker+0x1a0/0x1a0
>    []  ? kthread+0x116/0x130
>    []  ? kthread_flush_work_fn+0x10/0x10
>    []  ? ret_from_fork+0x35/0x40
>
> When removing the TIPC module, the UDP tunnel sock will be delayed to
> release in a work queue as sock_release() can't be done in rtnl_lock().
> If the work queue is schedule to run after the TIPC module is removed,
> kernel will crash as the work queue function cleanup_beareri() code no
> longer exists when trying to invoke it.
>
> To fix it, this patch introduce a member wq_count in tipc_net to track
> the numbers of work queues in schedule, and  wait and exit until all
> work queues are done in tipc_exit_net().
>
> Reported-by: Shuang Li <sh...@re...>
> Signed-off-by: Xin Long <luc...@gm...>
> ---
>   net/tipc/core.c      | 2 ++
>   net/tipc/core.h      | 2 ++
>   net/tipc/udp_media.c | 2 ++
>   3 files changed, 6 insertions(+)
>
> diff --git a/net/tipc/core.c b/net/tipc/core.c
> index 5cc1f03..72f3ac7 100644
> --- a/net/tipc/core.c
> +++ b/net/tipc/core.c
> @@ -119,6 +119,8 @@ static void __net_exit tipc_exit_net(struct net *net)
>   #ifdef CONFIG_TIPC_CRYPTO
>   	tipc_crypto_stop(&tipc_net(net)->crypto_tx);
>   #endif
> +	while (atomic_read(&tn->wq_count))
> +		cond_resched();
>   }
>   
>   static void __net_exit tipc_pernet_pre_exit(struct net *net)
> diff --git a/net/tipc/core.h b/net/tipc/core.h
> index 03de7b2..5741ae4 100644
> --- a/net/tipc/core.h
> +++ b/net/tipc/core.h
> @@ -149,6 +149,8 @@ struct tipc_net {
>   #endif
>   	/* Work item for net finalize */
>   	struct tipc_net_work final_work;
> +	/* The numbers of work queues in schedule */
> +	atomic_t wq_count;
>   };
>   
>   static inline struct tipc_net *tipc_net(struct net *net)
> diff --git a/net/tipc/udp_media.c b/net/tipc/udp_media.c
> index e556d2c..c2bb818 100644
> --- a/net/tipc/udp_media.c
> +++ b/net/tipc/udp_media.c
> @@ -814,6 +814,7 @@ static void cleanup_bearer(struct work_struct *work)
>   		kfree_rcu(rcast, rcu);
>   	}
>   
> +	atomic_dec(&tipc_net(sock_net(ub->ubsock->sk))->wq_count);
>   	dst_cache_destroy(&ub->rcast.dst_cache);
>   	udp_tunnel_sock_release(ub->ubsock);
>   	synchronize_net();
> @@ -834,6 +835,7 @@ static void tipc_udp_disable(struct tipc_bearer *b)
>   	RCU_INIT_POINTER(ub->bearer, NULL);
>   
>   	/* sock_release need to be done outside of rtnl lock */
> +	atomic_inc(&tipc_net(sock_net(ub->ubsock->sk))->wq_count);
>   	INIT_WORK(&ub->work, cleanup_bearer);
>   	schedule_work(&ub->work);
>   }

Acked-by: Jon Maloy <jm...@re...>