On 02/29/2016 05:27 PM, Thomas Monjalon wrote:
> 2016-02-29 17:19, Panu Matilainen:
>> On 02/29/2016 01:35 PM, Ferruh Yigit wrote:
>>> On 2/29/2016 11:06 AM, Thomas Monjalon wrote:
>>>> Hi,
>>>> I totally agree with Avi's comments.
>>>> This topic is really important for the future of DPDK.
>>>> So I think we must give some time to continue the discussion
>>>> and have netdev involved in the choices done.
>>>> As a consequence, these series should not be merged in the release 16.04.
>>>> Thanks for continuing the work.
>>>>
>>> Hi Thomas,
>>>
>>> It is great to have some discussion and feedbacks.
>>> But I doubt not merging in this release will help to have more discussion.
>>>
>>> It is better to have them in this release and let people experiment it,
>>> this gives more chance to better discussion.
>>>
>>> These features are replacement of KNI, and KNI is not intended to be
>>> removed in this release, so who are using KNI as solution can continue
>>> to use KNI and can test KCP/KDP, so that we can get more feedbacks.
>>
>> So make the work available from a separate git repo and make it easy for
>> people to experiment with it. Code doesn't have to be in a release for
>> the sake of experimenting, and removing code is much harder than not
>> adding it in the first place, witness KNI.
>
> Good idea.
> What about a -next tree to experiment on kernel interactions?
Here's another, related but more radical (and rather unbaked) idea:
Move all the kernel modules and their associated libraries (thinking of
KNI here) to a separate repo with perhaps more relaxed rules, but OTOH
require upstream kernel support for any features to be included in dpdk
itself. Carrot-and-stick of sorts :)
- Panu -
