Re: [PATCH 1/1] RDMA/rxe: Fix Use-After-Free problem in rxe_net_del

["yanjun.zhu" <yanjun.zhu@linux.dev>]

On 5/18/26 4:39 AM, Leon Romanovsky wrote:
> On Sun, May 17, 2026 at 06:47:47AM +0200, Zhu Yanjun wrote:
>> +	rxe_ns_lock(net);
>> +
>>   	sk = rxe_ns_pernet_sk4(net);
>>   	if (sk)
>>   		rxe_sock_put(sk, rxe_ns_pernet_set_sk4, net);
>> @@ -663,6 +665,8 @@ void rxe_net_del(struct ib_device *dev)
>>   	if (sk)
>>   		rxe_sock_put(sk, rxe_ns_pernet_set_sk6, net);
>>   
>> +	rxe_ns_unlock(net);
>> +
>>   	dev_put(ndev);
>>   }
>>   
>> diff --git a/drivers/infiniband/sw/rxe/rxe_ns.c b/drivers/infiniband/sw/rxe/rxe_ns.c
>> index 8b9d734229b2..799a727bc1fe 100644
>> --- a/drivers/infiniband/sw/rxe/rxe_ns.c
>> +++ b/drivers/infiniband/sw/rxe/rxe_ns.c
>> @@ -16,6 +16,7 @@
>>   struct rxe_ns_sock {
>>   	struct sock __rcu *rxe_sk4;
>>   	struct sock __rcu *rxe_sk6;
>> +	struct mutex	release_lock;
> 
> This change renders the existing rcu_read_lock() and rcu_read_unlock()
> calls unnecessary.

Thanks, Leon. I fully agree with you.
In the next version, I will remove the existing rcu locks.

Zhu Yanjun

> 
> Thanks