Matthias St. Pierre

@mcr here comes the explanation I promised in https://github.com/openssl/openssl/pull/17926#discussion_r837918808: ### Example For demonstration purposes, I added two new public symbols by "implementing" a public function `BIO_set_dgram_foo()` and a public function-like...

Um, no. It was your idea to document those public symbols and I really like your initial draft. So please continue, I didn't intend to steal your thunder or even...

IIRC, the OSSL_/ossl_ prefix policy applies strictly only to new 'namespaces', not to existing ones like 'x509'. In the case of `x509.h` it would be quite inconsistent to start adding...

Would it be asked too much to [add a simple fuzzing test](https://twitter.com/hanno/status/1587775675397726209), as suggested by @hannob on Twitter?

> In a separate PR ? A separate PR is ok, as long as it happens. We should not make the same mistake twice.

A compromise would be to add whirlpool and other (outdated, unsafe, ...) algorithms to the disabled-by-default list of algorithms in the next major release.

@asitde are you still interested in completing this pr? You have one approval of two, but still some outstanding change requests. And your branch needs to be rebased in order...

@bernd-edlinger why are the aes-cbc numbers for OpenSSL-1_1_1-stable much lower than for all other branches, including your 111 feature branch? Did you backport a new hardware implementation?

Is the speedup caused by the bit-sliced implementation?

The [buildbot failure](https://ci.buildbot.openssl.org/#/builders/45/builds/3033) is unrelated: ``` builtins.PermissionError: [WinError 32] The process cannot access the file because it is being used by another process: 'E:\\workers\\ossl-win10-x86_64-5\\master_windows-win10-x86\\build\\test-runs\\test_genpkey' program finished with exit code -1...