[PATCH] digest-openssl: improve OpenSSL v3 support

[Theo de Raadt]

Damien Miller <djm at mindrot.org> wrote:

> On Sun, 31 Aug 2025, Dimitri John Ledkov wrote:
> 
> > From OpenSSL v3 documentation https://docs.openssl.org/3.0/man3/EVP_sha1/#notes:
> > 
> >   Developers should be aware of the negative performance implications
> >   of calling this function multiple times and should consider using
> >   EVP_MD_fetch(3) with EVP_MD-SHA1(7) instead. See "Performance" in
> >   crypto(7) for further information.
> 
> hash/MAC instantiation doesn't happen very often in OpenSSH. Does this
> change yield any observable performance benefit? ("make unit-bench"
> might help here).

I think that cpp (meaning #ifdef) is an amazing programming language
with extremely strong negative incentive to not test the other codepath
when a change is made.  Dangerously so.  There's got to be strong
benefit, before this kind of thing gets done.