[dpdk-stable] [dpdk-dev] [PATCH] virtio-user: fix backend selection if stat fails

[Kevin Traynor]

On 20/10/2020 10:11, Maxime Coquelin wrote:
> 
> 
> On 10/20/20 11:01 AM, Kevin Traynor wrote:
>> On 20/10/2020 08:16, Adrian Moreno wrote:
>>> If stat fails it means the backend must be vhost-user in server mode
>>>
>>> Bugzilla ID: 559
>>> Fixes: f908b22ea47a ("net/virtio: move backend type selection to ethdev")
>>> Cc: stable at dpdk.org
>>>
>>> Signed-off-by: Adrian Moreno <amorenoz at redhat.com>
>>> ---
>>>  drivers/net/virtio/virtio_user_ethdev.c | 5 +++--
>>>  1 file changed, 3 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/net/virtio/virtio_user_ethdev.c b/drivers/net/virtio/virtio_user_ethdev.c
>>> index 042665bc0..ce74d08ab 100644
>>> --- a/drivers/net/virtio/virtio_user_ethdev.c
>>> +++ b/drivers/net/virtio/virtio_user_ethdev.c
>>> @@ -560,9 +560,10 @@ virtio_user_backend_type(const char *path)
>>>  	struct stat sb;
>>>  
>>>  	if (stat(path, &sb) == -1) {
>>> -		PMD_INIT_LOG(ERR, "Stat fails: %s (%s)\n", path,
>>> +		PMD_INIT_LOG(INFO, "Stat fails: %s (%s)\n", path,
>>>  			     strerror(errno));
>>
>> It may be accurate, but a 'fail' in the logs can be confusing for users
>> when it is an INFO log and normal operation. Suggest to reword to
>> something softer like 'Unable to stat' or 'Not able to get file status'
> 
> 
> We may want to:
> - only return VIRTIO_USER_BACKEND_VHOST_USER if -ENOENT, and log that
> we assume this is Vhost-user backend in server mode at INFO level.

It will mean that sometimes the backend type is logged and sometimes
not, but maybe you make a distinction because there is an assumption
being made in this case?

> - return VIRTIO_USER_BACKEND_UNKNOWN otherwise and print an error
> message with the strerror(errno).
> 

yes, it seems better like that.

> What do you think?
> 
>>> -		return VIRTIO_USER_BACKEND_UNKNOWN;
>>> +		/* Must be vhost-user in server mode */
>>> +		return VIRTIO_USER_BACKEND_VHOST_USER;
>>>  	}
>>>  
>>>  	if (S_ISSOCK(sb.st_mode)) {
>>>
>>