[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20200729.132842.190888844026802233.davem@davemloft.net>
Date: Wed, 29 Jul 2020 13:28:42 -0700 (PDT)
From: David Miller <davem@...emloft.net>
To: saeedm@...lanox.com
Cc: songliubraving@...com, hawk@...nel.org, kafai@...com,
kpsingh@...omium.org, john.fastabend@...il.com, leon@...nel.org,
linux-kernel@...r.kernel.org, ast@...nel.org,
linux-rdma@...r.kernel.org, xiongx18@...an.edu.cn, yhs@...com,
andriin@...com, kuba@...nel.org, daniel@...earbox.net,
bpf@...r.kernel.org, netdev@...r.kernel.org,
xiyuyang19@...an.edu.cn, tanxin.ctf@...il.com,
yuanxzhang@...an.edu.cn
Subject: Re: [PATCH] net/mlx5e: fix bpf_prog refcnt leaks in mlx5e_alloc_rq
From: Saeed Mahameed <saeedm@...lanox.com>
Date: Wed, 29 Jul 2020 19:02:15 +0000
> On Wed, 2020-07-29 at 20:33 +0800, Xin Xiong wrote:
>> The function invokes bpf_prog_inc(), which increases the refcount of
>> a
>> bpf_prog object "rq->xdp_prog" if the object isn't NULL.
>>
>> The refcount leak issues take place in two error handling paths. When
>> mlx5_wq_ll_create() or mlx5_wq_cyc_create() fails, the function
>> simply
>> returns the error code and forgets to drop the refcount increased
>> earlier, causing a refcount leak of "rq->xdp_prog".
>>
>> Fix this issue by jumping to the error handling path
>> err_rq_wq_destroy
>> when either function fails.
>>
>
> Fixes: 422d4c401edd ("net/mlx5e: RX, Split WQ objects for different RQ
> types")
Saeed, are you going to take this into your tree or would you like me to
apply it directly?
Thanks.
Powered by blists - more mailing lists