Add types and macros for ICMPv6 filtering

[Skyb0rg007]

Description

This commit adds the icmp6_filter type and the corresponding ICMP6_FILTER macros.
The API is defined in RFC 2292; I split-off this code from #4954 because the testing code here is more substantial.

These functions are tested in a similar manner to the CMSG macros, as the implementation differs between Linux and BSDs.

Sources

https://github.com/bminor/glibc/blob/04e750e75b73957cf1c791535a3f4319534a52fc/inet/netinet/icmp6.h
https://github.com/kraj/musl/blob/kraj/master/include/netinet/icmp6.h
https://github.com/freebsd/freebsd-src/blob/9ae367d11de8abbdf53884836c9ba30908c5c8db/sys/netinet/icmp6.h
https://github.com/apple-oss-distributions/xnu/blob/main/bsd/netinet/icmp6.h

Also in the RFC: https://datatracker.ietf.org/doc/html/rfc2292#section-3.2

Checklist

• Relevant tests in libc-test/semver have been updated
• No placeholder or unstable values like *LAST or *MAX are
  included (see #3131)
• Tested locally (cd libc-test && cargo test --target mytarget);
  especially relevant for platforms that may not be checked in CI

@rustbot label +stable-nominated

[rustbot]

Some changes occurred in a solarish module

cc @jclulow, @pfmooney

[Skyb0rg007]

Sorry for spamming the CI; I couldn't get it to work on my fork.

Anyways, I'm mostly confident with this PR. The missing parts are:

1. A way to initialize icmp6_filter. Right now the only field is private, so you need to use an unsafe method like std::mem::zeroed. I don't think this is a big deal.
2. The FreeBSD's netinet/icmp6.h header uses u_int32_t types, which I couldn't figure out how to properly include. The test code therefore includes the typedefs, which is somewhat inelegant.

Some notes on the implementation:

• The ICMP_FILTER functions are marked as safe. I think this is correct, since the functions are all total (due to marking the typ argument as u8).
• The icmp6_filter struct does not export its field. The meaning of a set bit differs on Linux and BSD, so there shouldn't be a good reason to not use the accessor functions.
• The test cases are somewhat slow. It takes ~1s to run them, but its because I am testing all 255 types for pass/fail. This could possibly be improved by using memcmp or similar.

[rustbot]

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.