[dpdk-dev] Promiscuous Mode Feature

[David Liu]

On Thu, Jun 25, 2020 at 4:23 PM Ajit Khaparde <ajit.khaparde at broadcom.com>
wrote:

>
>
> On Wed, Jun 24, 2020 at 1:03 PM Thomas Monjalon <thomas at monjalon.net>
> wrote:
>
>> Hi,
>>
>> A bit of context: David is going to implement a test in DTS
>> for promiscuous mode:
>> http://doc.dpdk.org/guides/nics/features.html#promiscuous-mode
>>
>> 24/06/2020 18:18, David Liu:
>> > Hi all,
>> >
>> > From my understanding, the promiscuous mode feature will verify the
>> > functionality of promiscuous mode when it is on or off.
>>
>> You need to check in which condition the promiscuous mode can be enabled.
>> Some contexts may forbid this mode. untrusted VF? flow isolation mode?
>>
> But there is no way other than some logs to indicate these.
> Maybe we should add some state in port info to indicate this?
>
>
 From what I know, I not sure if that is possible to add these states into
the test case?
 But let me know what you think.

>
>> > The plan for testing this feature is as follows:
>> >
>> > Sending over a packet to verify the ports are working properly before
>> any
>> > testing. First to verify the promiscuous mode is enabled. With the
>> > promiscuous being enabled, send over a packet and check if the packet
>> will
>> > be received when sending it to a different destination mac address than
>> the
>> > host/card. Then disable the promiscuous mode. The packet will only be
>> > received if the mac address is the destination mac address. Otherwise,
>> the
>> > test case will be considered a failure.
>>
>> I assume you will test positive and negative in both on/off cases.
>>
>> Yes, in the test case. It covered sending a packet with same/different
destination mac address
when promiscuous is on/off.

> Please let me know if there is anything I need to add on or if there are
>> > certain cases I need to be aware of.
>>
>> Anyone else has ideas about what to test and corner cases?
>>
>> >
>> > Thanks,
>> > David Liu
>> > UNH Interoperability Lab
>>
>>
>>