RFC: tc flower support

[daniel-noland]

I wanted to float the idea of support for tc flower filters.

flower is of special significance because some network cards (e.g. cards using the mlx5 driver or mlxsw switches) can offload significant portions of it.

I am happy to try my hand at implementing flower in this crate (and in downstream crates) but I wanted to make sure that I am on the same page with the maintainers before I implement a moderately large feature.

I have a very early code sketch in #111 which I post early to facilitate any needed discussion.

My plan is to

• implement flower in sympathy with the pattern set down in cls_u32.rs
• make a commit for each new key / component of filter I create

Let me know if you think I should approach this differently.

[cathay4t]

I have reviewed your patch, it looks very good. Please provide unit test case like other, it also verifying you are parsing and emitting as expected. You find notes about this in top README.md

[daniel-noland]

I have added a number of tests and am breaking apart the PR into tc-actions support and then tc-flower support as per discussion and to make this easier to review.

We need to make sure that this messy git commit history is fixed and #122 is merged before this can be merged.