Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: unregister_netdevice: waiting for DEV to become free (2)
To:     Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>,
        David Ahern <dsahern@gmail.com>,
        "David S. Miller" <davem@davemloft.net>
Cc:     Julian Anastasov <ja@ssi.bg>, Cong Wang <xiyou.wangcong@gmail.com>,
        syzbot <syzbot+30209ea299c09d8785c9@syzkaller.appspotmail.com>,
        ddstreet@ieee.org, dvyukov@google.com,
        linux-kernel@vger.kernel.org, netdev@vger.kernel.org,
        syzkaller-bugs@googlegroups.com
References: <0000000000007d22100573d66078@google.com>
 <alpine.LFD.2.20.1808201527230.2758@ja.home.ssi.bg>
 <4684eef5-ea50-2965-86a0-492b8b1e4f52@I-love.SAKURA.ne.jp>
 <9d430543-33c3-0d9b-dc77-3a179a8e3919@I-love.SAKURA.ne.jp>
 <920ebaf1-ee87-0dbb-6805-660c1cbce3d0@I-love.SAKURA.ne.jp>
 <cc054b5c-4e95-8d30-d4bf-9c85f7e20092@gmail.com>
 <15b353e9-49a2-f08b-dc45-2e9bad3abfe2@i-love.sakura.ne.jp>
 <057735f0-4475-7a7b-815f-034b1095fa6c@gmail.com>
 <6e57bc11-1603-0898-dfd4-0f091901b422@i-love.sakura.ne.jp>
From:   Eric Dumazet <eric.dumazet@gmail.com>
Message-ID: <a9075b87-4f02-fe78-e86d-34f67bdf8b2a@gmail.com>
Date:   Sun, 28 Apr 2019 08:04:18 -0700
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.6.1
MIME-Version: 1.0
In-Reply-To: <6e57bc11-1603-0898-dfd4-0f091901b422@i-love.sakura.ne.jp>
Content-Type: text/plain; charset=utf-8
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


On 4/27/19 9:22 PM, Tetsuo Handa wrote:
> On 2019/04/28 8:52, Eric Dumazet wrote:
>> On 4/27/19 3:33 PM, Tetsuo Handa wrote:
>>>
>>> I'm waiting for davem why it is safe to move the dst entry from
>>> "a device to unregister" to "a loopback device in that namespace".
>>> I'm waiting for an explanation how the dst entry which was moved to
>>> "a loopback device in that namespace" is released (i.e. what the
>>> expected shutdown sequence is).
>>
>> The most probable explanation is that we make sure the loopback device
>> is the last one to be dismantled at netns deletion,
>> and this would obviously happen after all dst have been released.
>>
> 
> rt_flush_dev() becomes a no-op if "dev" == "a loopback device in that
> namespace". And according to debug printk(), rt_flush_dev() is called
> on "a loopback device in that namespace" itself.
> 

This is the design yes. We can not let a dst having a pointer to some garbage memory.
(since we are going to free it very soon)

dst can be long lived objects. netdev (but loopback) are not.

> If "a loopback device in that namespace" is the last "one" (== "a network
> device in that namespace" ?), which shutdown sequence should have called
> dev_put("a loopback device in that namespace") before unregistration of
> "a loopback device in that namespace" starts?

You'll have to study all the netdev notifiers to answer this question.
They are many of them, and they have a priority to let them run in a given order.

> 
> Since I'm not a netdev person, I appreciate if you can explain
> that shutdown sequence using a flow chart.

I am a netdev person, but I have no time to explain this at this moment.