There are header files in crypto/ that are used by a number of crypto/
submodules.  Move those to crypto/include/internal and adapt the
affected source code and Makefiles.

The header files that got moved are:

crypto/cryptolib.h
crypto/md32_common.h
Reviewed-by: NRich Salz <rsalz@openssl.org>

For a local variable:
        TYPE *p;
Allocations like this are "risky":
        p = OPENSSL_malloc(sizeof(TYPE));
if the type of p changes, and the malloc call isn't updated, you
could get memory corruption.  Instead do this:
        p = OPENSSL_malloc(sizeof(*p));
Also fixed a few memset() calls that I noticed while doing this.
Reviewed-by: NRichard Levitte <levitte@openssl.org>

./config --strict-warnings enable-deprecated
Reviewed-by: NTim Hudson <tjh@openssl.org>

Reviewed-by: NTim Hudson <tjh@openssl.org>

Reviewed-by: NTim Hudson <tjh@openssl.org>

Reviewed-by: NTim Hudson <tjh@openssl.org>

Reviewed-by: NTim Hudson <tjh@openssl.org>

Several files #include stdio.h and don't need it.
Also, per tjh, remove BN_COUNT
Reviewed-by: NEmilia Kasper <emilia@openssl.org>

crypto.h if needed.

Modify source tree to handle change.

OpenSSL malloc dependencies.

("perl util/ck_errf.pl */*.c */*/*.c" still reports many more.)

been done, and took care of the other one (which hadn't).

Submitted by: Nils Larsch
Reviewed by: Geoff Thorpe

else then at least so bc doesn't have problems parsing the output from
bntest :-).

little TODO list in there as well as the debugging code (only enabled if
BN_CTX_DEBUG is defined).

I'd appreciate as much review and testing as can be spared for this. I'll
commit some changes to other parts of the bignum code shortly to make
better use of this implementation (no more fixed size limitations). Note
also that under identical optimisations, I'm seeing a noticable speed
increase over openssl-0.9.7 - so any feedback to confirm/deny this on other
systems would also be most welcome.

change to work properly; BN_zero() should set 'neg' to zero as well as
'top' to match the behaviour of BN_new().

return a "zero" bignum as BN_new() does - so reset 'top'. During
BN_CTX_end(), released bignums should be consistent so enforce this in
debug builds. Also, reduce the number of wasted BN_clear_free() calls from
BN_CTX_end() (typically by 75% or so).

Submitted by: Nils Larsch
Reviewed by: Geoff Thorpe, Ulf Möller

is itself experimental, and in addition may cause execution to break on
existing openssl "bugs" that previously were harmless or at least
invisible.

left in an inconsistent state when they are released for later reuse. This
change resets the BIGNUMs when they are released back to the context.

except internally to the allocator BN_CTX_new(), as such this deprecates
the use of BN_CTX_init() in the API. Moreover, the structure definition of
BN_CTX is taken out of bn_lcl.h and moved into bn_ctx.c itself.

NDEBUG should probably only be "forced" in the top-level configuration, but
until it is I will avoid removing it from bn_ctx.c which might surprise
people with massive slow-downs in their keygens. So I've left it in
bn_ctx.c but tidied up the preprocessor logic a touch and made it more
tolerant of debugging efforts.

Incease the number of BIGNUMs in a BN_CTX.

like Malloc, Realloc and especially Free conflict with already existing names
on some operating systems or other packages.  That is reason enough to change
the names of the OpenSSL memory allocation macros to something that has a
better chance of being unique, like prepending them with OPENSSL_.

This change includes all the name changes needed throughout all C files.

and initialize too_many because memset(..., 0, ...) is not used here.

Avoid enabled 'assert()' in production library.

temporary BIGNUMs. BN_CTX still uses a fixed number of BIGNUMs, but
the BN_CTX implementation could now easily be changed.

used anywhere in OpenSSL, but might be used by libdes applications.

Never use des_set_key (it depends on the global variable des_check_key),
but usually des_set_key_unchecked.
Only destest.c bothered to look at the return values of des_set_key,
but it did not set des_check_key -- if it had done so,
most checks would have failed because of wrong parity and
because of weak keys.

(meaning pointer to char) to des_cblock * (meaning pointer to
array with 8 char elements), which allows the compiler to
do more typechecking.  (The changed argument types were of type
des_cblock * back in SSLeay, and a lot of ugly casts were
used then to turn them into pointers to elements; but it can be
done without those casts.)

Introduce new type const_des_cblock -- before, the pointers rather
than the elements pointed to were declared const, and for
some reason gcc did not complain about this (but some other
compilers did).