[PATCH v2] examples/l2fwd_acl: add ACL-based L2 forwarding example

[Harsh raj Singh]

Hi Konstantin,

Thanks for the clarification.
That makes sense. I agree that improving the existing l3fwd
documentation and comments would be more useful than introducing a
separate sample application.

I’ll drop this patch and instead prepare a new patch that focuses on
improving the l3fwd ACL documentation and adding clearer explanations
in the code where appropriate.

Thanks for the guidance.

Best regards,
Harsh


On Mon, Feb 23, 2026 at 3:09 PM Konstantin Ananyev
<konstantin.ananyev at huawei.com> wrote:
>
>
>
> >Thanks for the review.
> >The intention of this example is not to replace the existing l3fwd ACL mode, but to provide a minimal and easier starting point for >users who want to understand how to use the DPDK ACL library in isolation.
>
> For that thing, I think  we do have an ACL UT: app/test/*acl*
>
> >The l3fwd application is quite large and includes full L3 forwarding logic, routing tables, multiple lookup paths, and various >configuration options. Because of that, it can be difficult for new users to identify the specific steps required to
> >define ACL fields, create and build an ACL context, insert rules, and perform classification on packets.
> >This example focuses only on demonstrating ACL usage in a small, self-contained program, without routing or complex forwarding >logic. The goal is to make it easier for beginners to learn how ACL classification works before integrating it into larger applications.
> >If the maintainers feel that this use case is already sufficiently covered by existing examples, I’m happy to adjust or improve the patch >accordingly.
>
> I don't think there is a need to introduce a new sample app just for that.
> Feel free to submit a patch that improves l3fwd docs and/or comments to make
> It easier for newcomers to understand it.
>
> On a slightly different subject:
> when you replying to patch's comments - don't forget to hit 'reply all' button,
> and please use plain-text format.
>
> Thanks
> Konstantin
>
>
>