supposedly smart compilers from optimizing memory cleanups away. Understood.
Ok, in case of an hypothetically super smart compiler, OPENSSL_cleanse() had
to be convoluted enough for the compiler not to recognize that this was
actually bzero() in disguise. Understood.
But then why there had been optimized assembler versions of OPENSSL_cleanse()
is beyond me. Did someone not trust the C obfuscation?
of des_old.h routines, acting as wrappers about the OpenSSL UI API.
Nothing should use these functions directly.
Riding the recent libcrypto major bump (in a `des' car).
are not please fix your operating system. Replace mem.c with an API-compatible
wrapper that just calls the system functions and does not allow a one word
modification of a variable in a running shared library to turn on memory
debug functions that expose things that should not be seen.
ok tedu@
The man page says "Compatibility des_ functions are provided for a short
while" and indeed even the original commit message says "The compatibility
functions will be removed in some future release, at the latest in
version 1.0." So here we are, a short while later.
Now I've only been an OpenBSD developer for 11 years, one year less than
this header has existed, but in that brief time, I've learned a thing or
two about deleting obsolete code. It doesn't delete itself. And worse,
people will continue using it until you force them onto a better path.
obsolete (and mostly internal) routines to be compiled out.
We don't expect any reasonable software to stick to these interfaces, so better
clean up the view and unifdef -DNO_ASN1_OLD.
The astute reader will notice the existence of NO_OLD_ASN1 which serves a
similar purpose, but is more entangled. Its time will come, soon.
library expands until it has its own dlfcn wrapper, and libcrypto is no
exception.
Remove the non-dlfcn DSO methods.
This causes public DSO_METHOD_{beos,dl,vms,win32} to disappear (major bump
coming soon). Note that portable software ought to use DSO_METHOD_openssl
instead of picking the backend directly (which makes one wonder why the
backends are exposed, as it is unlikely that more than one can work on
your system).
ok beck@ deraadt@
existing RAND interfaces unchanged.
All interfaces allowing external feed or seed of the RNG (either from a file
or a local entropy gathering daemon) are kept for ABI compatibility, but are
no longer do anything.
While the OpenSSL PRNG was required 15+ years ago when many systems lacked
proper entropy collection, things have evolved and one can reasonably assume
it is better to use the kernel (system global) entropy pool rather than trying
to build one's own and having to compensate for thread scheduling...
<RANT>
Whoever thought that RAND_screen(), feeding the PRNG with the contents of the
local workstation's display, under Win32, was a smart idea, ought to be banned
from security programming.
</RANT>
ok beck@ deraadt@ tedu@
as configuration files; split manpages and .pc files between libcrypto and
libssl.
No functional change, only there to make engineering easier, and libcrypto
sources are still found in libssl/src/crypto at the moment.
ok reyk@, also discussed with deraadt@ beck@ and the usual crypto suspects.
of the stride. don't overwrite past the end of the buffer, and also save
that amount for later so the array is completely filled. ok deraadt djm
reported by Dmitry Chestnykh (dchest)
miod