Tracking Issue: drop support for OpenSSL

[marten-seemann]

Motivation: see #1952.

TLDR: OpenSSL support doesn't seem to provide any (significant) performance benefits outside of microbenchmarks. To stop it from continuing to consume a disproportionately of maintainer attention we should remove it.

Done criteria:

• remove OpenSSL code paths from go-libp2p: core/crypto: drop all OpenSSL code paths #1953
• archive https://github.com/libp2p/go-openssl
• write this up in the release notes for the next release
• maybe write a blog post? It might be interesting to explain what benchmarks we did and how we arrived at our conclusion. How to migrate the IPFS bootstrappers (migrate bootstrapper away from RSA ipfs/kubo#9507) away from RSA might also be an interesting story.

[marten-seemann]

Closing this as completed. The PR has been merged, the repo archived and I wrote up release notes. Not sure if we want to actually write a blog post, given that we don't expect any measurable performance impact.