such operation can be considered as breaking binary compatibility. However!
OPNESSL_ia32cap_P is accessed by application through pointer returned by
OPENSSL_ia32cap_loc() and such change of *internal* OPENSSL_ia32cap_P
declaration is possible specifically on little-endian platforms, such as
x86[_64] ones in question. In addition, if 32-bit application calls
OPENSSL_ia32cap_loc(), it clears upper half of capability vector maintaining
the illusion that it's still 32 bits wide.
called whrlpool is not a typo, but a way to keep the names shorter than
8 characters. Remaining TODO list comprises adding OID, EVP, corresponding
flag to apps/openssl dgst, benchmark, engage assembler...
x86*_nw.pl will be deleted. In addition this update implements initseg
on several additional [in addition to ELF] platforms. Functions registered
with initseg are supposed to be called prior main().
is to have a placeholder to small routines, which can be written only
in assembler. In IA-32 case this includes processor capability
identification and access to Time-Stamp Counter. As discussed earlier
OPENSSL_ia32cap is introduced to control recently added SSE2 code
pathes (see docs/crypto/OPENSSL_ia32cap.pod). For the moment the
code is operational on ELF platforms only. I haven't checked it yet,
but I have all reasons to believe that Windows build should fail to
link too. I'll be looking into it shortly...
the one to be used to denote local labels in single function scope.
Problem is that SHA uses same label set across functions, therefore I
have to switch back to $ prefix.
assembler for various X86 platforms including Win32. It can output object files
that VC++ will tolerate so it could be used to provide assembly language support
to Win32 without the need for MASM.
This is preliminary stuff: it doesn't even work yet.
