Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <61ac01e70c293b7e0d378436f7617daaad552fcf.camel@codeconstruct.com.au>
Subject: Re: [PATCH v0] mctp: fix netdev reference bug
From:   Matt Johnston <matt@codeconstruct.com.au>
To:     Lin Ma <linma@zju.edu.cn>, jk@codeconstruct.com.au,
        davem@davemloft.net, kuba@kernel.org, pabeni@redhat.com,
        netdev@vger.kernel.org, linux-kernel@vger.kernel.org
Date:   Thu, 24 Mar 2022 11:17:56 +0800
In-Reply-To: <20220324023904.7173-1-linma@zju.edu.cn>
References: <20220324023904.7173-1-linma@zju.edu.cn>
Content-Type: text/plain; charset="UTF-8"
User-Agent: Evolution 3.40.4-1ubuntu2 
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Precedence: bulk

Hi Lin Ma,

On Thu, 2022-03-24 at 10:39 +0800, Lin Ma wrote:
> In extended addressing mode, function mctp_local_output() fetch netdev
> through dev_get_by_index_rcu, which won't increase netdev's reference
> counter. Hence, the reference may underflow when mctp_local_output calls
> dev_put(), results in possible use after free.
> 
> This patch adds dev_hold() to fix the reference bug.

This was already fixed in net-next to increment the refcount in
__mctp_dev_get() and use mctp_dev_put().

dc121c008491 ("mctp: make __mctp_dev_get() take a refcount hold")
e297db3eadd7 ("mctp: Fix incorrect netdev unref for extended addr")

Thanks,
Matt