Re: [dpdk-stable] [PATCH] net/virtio-user: fix error run close(0)

[Maxime Coquelin <maxime.coquelin@redhat.com>]

On 12/9/20 12:31 PM, Xia, Chenbo wrote:
> Hi Jiawei,
> 
> Thanks for catching this!
> Comments inline.
> 
>> -----Original Message-----
>> From: Jiawei Zhu <17826875952@163.com>
>> Sent: Saturday, November 28, 2020 10:50 PM
>> To: dev@dpdk.org
>> Cc: liweifeng2@huawei.com; zhujiawei12@huawei.com; maxime.coquelin@redhat.com;
>> Xia, Chenbo <chenbo.xia@intel.com>
>> Subject: [PATCH] net/virtio-user: fix error run close(0)
>>
>> From: Jiawei Zhu <zhujiawei12@huawei.com>
>>
>> When i < VIRTIO_MAX_VIRTQUEUES and j == i,
>> dev->callfds[i] and dev->kickfds[i] are default 0.
>> So it will close(0), close the standard input (stdin).
>>
>> Fixes: e6e7ad8b3024 ("net/virtio-user: move eventfd open/close into
>> init/uninit")
>> Cc: stable@dpdk.org
>>
>> Signed-off-by: Jiawei Zhu <zhujiawei12@huawei.com>
>> ---
>>  drivers/net/virtio/virtio_user/virtio_user_dev.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/net/virtio/virtio_user/virtio_user_dev.c
>> b/drivers/net/virtio/virtio_user/virtio_user_dev.c
>> index 053f026..1bfd223 100644
>> --- a/drivers/net/virtio/virtio_user/virtio_user_dev.c
>> +++ b/drivers/net/virtio/virtio_user/virtio_user_dev.c
>> @@ -284,7 +284,7 @@ int virtio_user_stop_device(struct virtio_user_dev *dev)
>>  	}
>>
>>  	if (i < VIRTIO_MAX_VIRTQUEUES) {
>> -		for (j = 0; j <= i; ++j) {
>> +		for (j = 0; j < i; ++j) {
> 
> With the help of your patch, I notice another defect that if the code fails in kickfd
> creation, we will leave one callfd not closed. Since you are here, could you help solve
> this too? A potential solution could be doing 'dev->callfds[i] = callfd' just after callfd
> creation, keeping 'j <= i' and adding checks before close(). What do you think?

+1, I noticed the same discussing the patch with David.

> Btw, I noticed that you have sent multiple patches that have same content. If you want to
> send new version. Please --in-reply-to this patch as this is the one that shows in patchwork.
> (http://patchwork.dpdk.org/patch/84626/)

I personally don't care about the --in-reply-to, but I agree for the
versioning.

Thanks,
Maxime

> Thanks!
> Chenbo
> 
>>  			close(dev->callfds[j]);
>>  			close(dev->kickfds[j]);
>>  		}
>> --
>> 1.8.3.1
>