Re: ssl code going astray

On Wed, Apr 07, 2021 at 12:45:42AM +0200, Thomas Klausner wrote:
> On my NetBSD 9.99.81/amd64 however, this dumps core, with a weird
> backtrace[2]: In the middle of the openssl part of the stack (frames
> 16-5), functions from nss are called instead of the ones of the same
> name from openssl (frames 4-0).

It's potentially even worse, since they are also libc names. At least
OpenSSL is patched in NetBSD base to not use its own implementation as
it is using a larger context than libc and would therefore results in
buffer overflows all over the place.

Joerg