[DRAFT] Update benchmarks

[justsmth]

Description of changes:

Adds a new aws-lc-rs-bench crate — a standalone benchmarking tool that measures CPU instruction counts using Valgrind's callgrind. Unlike wall-clock benchmarks, instruction counts are deterministic and noise-free, which makes them well-suited for catching unexpected performance changes in CI.

The tool covers AEAD, digest, HMAC, HKDF, key agreement, and signature operations (55 benchmarks total). A GitHub Actions workflow runs these on every PR and pushes a comparison report as a PR comment.

Also cleans up aws-lc-rs-testing:

• Makes ring and openssl truly optional (previously they were unconditional dev-deps behind empty feature flags)
• Gates the cross-library comparison tests with #![cfg(feature = "...")]
• Modernizes the RSA verify benchmarks to use UnparsedPublicKey instead of the deprecated VerificationAlgorithm trait
• Inlines the agree_ephemeral calls in agreement benchmarks to accommodate the differing aws-lc-rs / ring signatures

Call-outs:

• On this PR the "Benchmarks" CI jobs will fail b/c there is nothing to compare them to.
• aws-lc-rs-bench is Linux-only for instruction counting (requires Valgrind). A walltime subcommand is available for quick local testing on any platform.
• The crabgrind crate is pulled in only on Linux via cfg(target_os = "linux").
• The aws-lc-rs-testing dependency changes mean cargo test -p aws-lc-rs-testing (without features) now compiles the ring/openssl comparison tests to zero test cases. Pass --features ring-benchmarks,openssl-benchmarks to run them.

Testing:

• cargo check -p aws-lc-rs-bench and cargo test -p aws-lc-rs-testing --features ring-benchmarks,openssl-benchmarks both pass.
• Verified walltime mode runs all 55 benchmarks end-to-end.
• Verified that without features, all aws-lc-rs-testing test targets correctly compile to 0 tests (feature gates work).

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license and the ISC license.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 50.00000% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 92.61%. Comparing base (c358484) to head (a1b8e0d).
⚠️ Report is 351 commits behind head on main.

Files with missing lines	Patch %	Lines
aws-lc-rs/src/cipher.rs	50.00%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1014      +/-   ##
==========================================
- Coverage   95.80%   92.61%   -3.19%     
==========================================
  Files          61       71      +10     
  Lines        8143    10051    +1908     
  Branches        0    10051   +10051     
==========================================
+ Hits         7801     9309    +1508     
- Misses        342      450     +108     
- Partials        0      292     +292     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

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