Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Subject: Re: [RFC PATCH] vhost-vdpa: mark vhost device invalid to reflect vdpa
 device unregistration
From:   Jason Wang <jasowang@redhat.com>
To:     gautam.dawar@xilinx.com
Cc:     martinh@xilinx.com, hanand@xilinx.com, gdawar@xilinx.com,
        "Michael S. Tsirkin" <mst@redhat.com>, kvm@vger.kernel.org,
        virtualization@lists.linux-foundation.org, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org
References: <20210704205205.6132-1-gdawar@xilinx.com>
 <3d02b8f5-0a6b-e8d1-533d-8503da3fcc4e@redhat.com>
Message-ID: <d392646d-c871-8203-ae67-f21db0388380@redhat.com>
Date:   Mon, 5 Jul 2021 13:11:18 +0800
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:78.0)
 Gecko/20100101 Thunderbird/78.11.0
MIME-Version: 1.0
In-Reply-To: <3d02b8f5-0a6b-e8d1-533d-8503da3fcc4e@redhat.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Content-Language: en-US
Precedence: bulk


在 2021/7/5 上午11:48, Jason Wang 写道:
>
> 在 2021/7/5 上午4:52, gautam.dawar@xilinx.com 写道:
>>       vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
>> @@ -1091,11 +1122,13 @@ static void vhost_vdpa_remove(struct 
>> vdpa_device *vdpa)
>>           opened = atomic_cmpxchg(&v->opened, 0, 1);
>>           if (!opened)
>>               break;
>> -        wait_for_completion_timeout(&v->completion,
>> -                        msecs_to_jiffies(1000));
>> -        dev_warn_once(&v->dev,
>> -                  "%s waiting for/dev/%s to be closed\n",
>> -                  __func__, dev_name(&v->dev));
>> +        if (!wait_for_completion_timeout(&v->completion,
>> +                        msecs_to_jiffies(1000))) {
>> +            dev_warn(&v->dev,
>> +                 "%s/dev/%s in use, continue..\n",
>> +                 __func__, dev_name(&v->dev));
>> +            break;
>> +        }
>>       } while (1);
>>         put_device(&v->dev);
>> +    v->dev_invalid = true;
>
>
> Besides the mapping handling mentioned by Michael. I think this can 
> lead use-after-free. put_device may release the memory.
>
> Another fundamental issue, vDPA is the parent of vhost-vDPA device. 
> I'm not sure the device core can allow the parent to go away first.


Or this probably means you need couple the fd loosely with the 
vhost-vDPA device.

Thanks


>
> Thanks
>
>