v4.0: xdp: switch to lpm trie for routing (backport of #11465)

[mergify]

The initial routing implementation in agave-xdp assumed a static routing table with only a handful of entries (one per interface + default).

DZ breaks that assumption by installing hundreds of routes (~550 on mnb today). The result is that routing when DZ is on doesn't work at all, the XDP thread is pegged at 100% cpu routing, which leads to skipped blocks and the XDP channel filling up.

This PR fixes the issue by reworking routing and implementing a 4-bit (nibble) trie for lookups.

before

after

---

This is an automatic backport of pull request #11465 done by [Mergify](https://mergify.com).

[mergify]

Cherry-pick of da1dcd7 has failed:

On branch mergify/bp/v4.0/pr-11465
Your branch is up to date with 'origin/v4.0'.

You are currently cherry-picking commit da1dcd767.
  (fix conflicts and run "git cherry-pick --continue")
  (use "git cherry-pick --skip" to skip this patch)
  (use "git cherry-pick --abort" to cancel the cherry-pick operation)

Changes to be committed:
	modified:   xdp/src/lib.rs
	new file:   xdp/src/lpm.rs
	modified:   xdp/src/netlink.rs
	modified:   xdp/src/route.rs
	modified:   xdp/src/route_monitor.rs

Unmerged paths:
  (use "git add <file>..." to mark resolution)
	both modified:   xdp/src/xdp_retransmitter.rs

To fix up this pull request, you can check it out locally. See documentation: https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/reviewing-changes-in-pull-requests/checking-out-pull-requests-locally

[codecov-commenter]

Codecov Report

❌ Patch coverage is 87.82490% with 89 lines in your changes missing coverage. Please review.
✅ Project coverage is 82.9%. Comparing base (ca034bd) to head (1c92675).

Additional details and impacted files

@@            Coverage Diff            @@
##             v4.0   #11538     +/-   ##
=========================================
- Coverage    83.0%    82.9%   -0.1%     
=========================================
  Files         838      839      +1     
  Lines      316793   317165    +372     
=========================================
- Hits       263229   263223      -6     
- Misses      53564    53942    +378     

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[gregcusack]

lgtm!

[bw-solana]

should we conditionally set this false on success only?

[alessandrod]

same as below, from_tables essentially never fails.

Even if it failed, from_tables(tables) with the same tables would fail in exactly the same way, unless the tables change, in which case dirty=true.

[bw-solana]

I'm afraid

[alessandrod]

this can never fail

inside from_tables we do

        router.cached_default_route = match router.default_route() {
            Ok(hop) => Some(hop),
            Err(RouteError::NoRouteFound(_)) => None,
            Err(e) => return Err(io::Error::other(e)),
        };

        Ok(router)

and default_route() fails only with NoRouteFound.

I'll rework the errors in master so this gets less scary but yeah it's effectively an impossible condition.