Time to release OpenSSL 0.9.7 beta6.

The tag will be OpenSSL_0_9_7-beta6.

CHANGES

@@ -4,6 +4,20 @@
 
  Changes between 0.9.6h and 0.9.7  [XX xxx 2002]
 
+  *) The hw_ncipher.c engine requires dynamic locks.  Unfortunately, it
+     seems that in spite of existing for more than a year, many application
+     author have done nothing to provide the necessary callbacks, which
+     means that this particular engine will not work properly anywhere.
+     This is a very unfortunate situation which forces us, in the name
+     of usability, to give the hw_ncipher.c a static lock, which is part
+     of libcrypto.
+     NOTE: This is for the 0.9.7 series ONLY.  This hack will never
+     appear in 0.9.8 or later.  We EXPECT application authors to have
+     dealt properly with this when 0.9.8 is released (unless we actually
+     make such changes in the libcrypto locking code that changes will
+     have to be made anyway).
+     [Richard Levitte]
+
   *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
      octets have been read, EOF or an error occurs. Without this change
      some truncated ASN1 structures will not produce an error.

Configure

@@ -161,25 +161,25 @@ my %table=(
 # surrounds it with #APP #NO_APP comment pair which (at least Solaris
 # 7_x86) /usr/ccs/bin/as fails to assemble with "Illegal mnemonic"
 # error message.
-"solaris-x86-gcc","gcc:-O3 -fomit-frame-pointer -m486 -Wall -DL_ENDIAN -DOPENSSL_NO_INLINE_ASM::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}:${x86_sol_asm}:dlfcn:solaris-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"solaris-x86-gcc","gcc:-O3 -fomit-frame-pointer -m486 -Wall -DL_ENDIAN -DOPENSSL_NO_INLINE_ASM::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}:${x86_sol_asm}:dlfcn:solaris-shared:-fPIC:-shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 
 #### Solaris x86 with Sun C setups
 "solaris-x86-cc","cc:-fast -O -Xa::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_PTR DES_UNROLL BF_PTR::::::::::dlfcn:solaris-shared:-KPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 
 #### SPARC Solaris with GNU C setups
-"solaris-sparcv7-gcc","gcc:-O3 -fomit-frame-pointer -Wall -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR::::::::::dlfcn:solaris-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"solaris-sparcv8-gcc","gcc:-mv8 -O3 -fomit-frame-pointer -Wall -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:asm/sparcv8.o:::::::::dlfcn:solaris-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"solaris-sparcv7-gcc","gcc:-O3 -fomit-frame-pointer -Wall -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR::::::::::dlfcn:solaris-shared:-fPIC:-shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"solaris-sparcv8-gcc","gcc:-mv8 -O3 -fomit-frame-pointer -Wall -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:asm/sparcv8.o:::::::::dlfcn:solaris-shared:-fPIC:-shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 # -m32 should be safe to add as long as driver recognizes -mcpu=ultrasparc
-"solaris-sparcv9-gcc","gcc:-m32 -mcpu=ultrasparc -O3 -fomit-frame-pointer -Wall -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:asm/sparcv8plus.o:::asm/md5-sparcv8plus.o::::::dlfcn:solaris-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"solaris64-sparcv9-gcc31","gcc:-mcpu=ultrasparc -m64 -O3 -fomit-frame-pointer -Wall -DB_ENDIAN::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_INT DES_PTR DES_RISC1 DES_UNROLL BF_PTR::::asm/md5-sparcv9.o::::::dlfcn:solaris-shared:-fPIC:-m64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"solaris-sparcv9-gcc","gcc:-m32 -mcpu=ultrasparc -O3 -fomit-frame-pointer -Wall -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:asm/sparcv8plus.o:::asm/md5-sparcv8plus.o::::::dlfcn:solaris-shared:-fPIC:-shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"solaris64-sparcv9-gcc31","gcc:-mcpu=ultrasparc -m64 -O3 -fomit-frame-pointer -Wall -DB_ENDIAN::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_INT DES_PTR DES_RISC1 DES_UNROLL BF_PTR::::asm/md5-sparcv9.o::::::dlfcn:solaris-shared:-fPIC:-m64 -shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 # gcc pre-2.8 doesn't understand -mcpu=ultrasparc, so fall down to -mv8
 # but keep the assembler modules.
-"solaris-sparcv9-gcc27","gcc:-mv8 -O3 -fomit-frame-pointer -Wall -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:asm/sparcv8plus-gcc27.o:::asm/md5-sparcv8plus-gcc27.o::::::dlfcn:solaris-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"solaris64-sparcv9-gcc","gcc:-m64 -mcpu=ultrasparc -O3 -Wall -DB_ENDIAN::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_INT DES_PTR DES_RISC1 DES_UNROLL BF_PTR::::asm/md5-sparcv9.o::::::dlfcn:solaris-shared:-fPIC:-m64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"solaris-sparcv9-gcc27","gcc:-mv8 -O3 -fomit-frame-pointer -Wall -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:asm/sparcv8plus-gcc27.o:::asm/md5-sparcv8plus-gcc27.o::::::dlfcn:solaris-shared:-fPIC:-shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"solaris64-sparcv9-gcc","gcc:-m64 -mcpu=ultrasparc -O3 -Wall -DB_ENDIAN::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_INT DES_PTR DES_RISC1 DES_UNROLL BF_PTR::::asm/md5-sparcv9.o::::::dlfcn:solaris-shared:-fPIC:-m64 -shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 
 ####
-"debug-solaris-sparcv8-gcc","gcc:-DBN_DEBUG -DREF_CHECK -DCONF_DEBUG -DBN_CTX_DEBUG -DCRYPTO_MDEBUG_ALL -O -g -mv8 -Wall -DB_ENDIAN::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:asm/sparcv8.o:::::::::dlfcn:solaris-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"debug-solaris-sparcv9-gcc","gcc:-DBN_DEBUG -DREF_CHECK -DCONF_DEBUG -DBN_CTX_DEBUG -DCRYPTO_MDEBUG_ALL -O -g -mcpu=ultrasparc -Wall -DB_ENDIAN::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:asm/sparcv8plus.o:::::::::dlfcn:solaris-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"debug-solaris-sparcv8-gcc","gcc:-DBN_DEBUG -DREF_CHECK -DCONF_DEBUG -DBN_CTX_DEBUG -DCRYPTO_MDEBUG_ALL -O -g -mv8 -Wall -DB_ENDIAN::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:asm/sparcv8.o:::::::::dlfcn:solaris-shared:-fPIC:-shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"debug-solaris-sparcv9-gcc","gcc:-DBN_DEBUG -DREF_CHECK -DCONF_DEBUG -DBN_CTX_DEBUG -DCRYPTO_MDEBUG_ALL -O -g -mcpu=ultrasparc -Wall -DB_ENDIAN::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:asm/sparcv8plus.o:::::::::dlfcn:solaris-shared:-fPIC:-shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 
 #### SPARC Solaris with Sun C setups
 # DO NOT use /xO[34] on sparc with SC3.0.  It is broken, and will not pass the tests
@@ -203,9 +203,8 @@ my %table=(
 # it's a real mess with -mcpu=ultrasparc option under Linux, but
 # -Wa,-Av8plus should do the trick no matter what.
 "linux-sparcv9","gcc:-mcpu=ultrasparc -DB_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -Wall -Wa,-Av8plus -DBN_DIV2W::-D_REENTRANT:ULTRASPARC:-ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:asm/sparcv8plus.o:::asm/md5-sparcv8plus.o::::::dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-# !!!Folowing can't be even tested yet!!!
-#    We have to wait till 64-bit glibc for SPARC is operational!!!
-#"linux64-sparcv9","sparc64-linux-gcc:-m64 -mcpu=v9 -DB_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -Wall -DBN_DIV2W::-D_REENTRANT:ULTRASPARC::BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR::::asm/md5-sparcv9.o:",
+# GCC 3.1 is a requirement
+"linux64-sparcv9","gcc:-m64 -mcpu=ultrasparc -DB_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -Wall::-D_REENTRANT:ULTRASPARC:-ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR::::asm/md5-sparcv9.o::::::dlfcn:linux-shared:-fPIC:-m64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 
 # Sunos configs, assuming sparc for the gcc one.
 ##"sunos-cc", "cc:-O4 -DNOPROTO -DNOCONST::(unknown):SUNOS::DES_UNROLL:::",
@@ -392,7 +391,7 @@ my %table=(
 "linux-s390",	"gcc:-DB_ENDIAN -DTERMIO -DNO_ASM -O3 -fomit-frame-pointer -Wall::-D_REENTRANT::-ldl:BN_LLONG::::::::::dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "linux-s390x",	"gcc:-DB_ENDIAN -DTERMIO -DNO_ASM -O3 -fomit-frame-pointer -Wall::-D_REENTRANT::-ldl:SIXTY_FOUR_BIT_LONG::::::::::dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "linux-ia64",   "gcc:-DL_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -Wall::-D_REENTRANT::-ldl:SIXTY_FOUR_BIT_LONG RC4_CHUNK RC4_CHAR:asm/ia64.o:::::::::dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"linux-x86_64", "gcc:-DL_ENDIAN -DNO_ASM ::-D_REENTRANT::-ldl:SIXTY_FOUR_BIT_LONG::::::::::dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"linux-x86_64", "gcc:-m64 -DL_ENDIAN -DTERMIO -O3 -Wall -DMD32_REG_T=int::-D_REENTRANT::-ldl:SIXTY_FOUR_BIT_LONG RC4_CHUNK RC4_CHAR BF_PTR2 DES_INT DES_UNROLL:asm/x86_64-gcc.o:::::::::dlfcn:linux-shared:-fPIC:-m64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "NetBSD-sparc",	"gcc:-DTERMIOS -O3 -fomit-frame-pointer -mv8 -Wall -DB_ENDIAN::(unknown):::BN_LLONG MD2_CHAR RC4_INDEX DES_UNROLL::::::::::dlfcn:bsd-gcc-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "NetBSD-m68",	"gcc:-DTERMIOS -O3 -fomit-frame-pointer -Wall -DB_ENDIAN::(unknown):::BN_LLONG MD2_CHAR RC4_INDEX DES_UNROLL::::::::::dlfcn:bsd-gcc-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "NetBSD-x86",	"gcc:-DTERMIOS -O3 -fomit-frame-pointer -m486 -Wall::(unknown):::BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}::::::::::dlfcn:bsd-gcc-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
@@ -528,7 +527,7 @@ my %table=(
 "DJGPP", "gcc:-I/dev/env/WATT_ROOT/inc -DTERMIOS -DL_ENDIAN -fomit-frame-pointer -O2 -Wall:::MSDOS:-L/dev/env/WATT_ROOT/lib -lwatt:BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}::::::::::",
 
 # Ultrix from Bernhard Simon <simon@zid.tuwien.ac.at>
-"ultrix-cc","cc:-std1 -O -Olimit 1000 -DL_ENDIAN::(unknown):::::::",
+"ultrix-cc","cc:-std1 -O -Olimit 2500 -DL_ENDIAN::(unknown):::::::",
 "ultrix-gcc","gcc:-O3 -DL_ENDIAN::(unknown):::::::",
 # K&R C is no longer supported; you need gcc on old Ultrix installations
 ##"ultrix","cc:-O2 -DNOPROTO -DNOCONST -DL_ENDIAN::(unknown):::::::",
@@ -1084,6 +1083,11 @@ if ($no_asm)
 	$sha1_obj=$md5_obj=$rmd160_obj="";
 	}
 
+if (!$no_shared)
+	{
+	$cast_obj="";	# CAST assembler is not PIC
+	}
+
 if ($threads)
 	{
 	$cflags=$thread_cflags;
@@ -1099,20 +1103,16 @@ if ($zlib)
 
 # You will find shlib_mark1 and shlib_mark2 explained in Makefile.org
 my $shared_mark = "";
-if ($shared_target ne "")
+if ($shared_target eq "")
+	{
+	$no_shared = 1;
+	}
+if (!$no_shared)
 	{
 	if ($shared_cflag ne "")
 		{
 		$cflags = "$shared_cflag $cflags";
 		}
-	if (!$no_shared)
-		{
-		#$shared_mark = "\$(SHARED_LIBS)";
-		}
-	}
-else
-	{
-	$no_shared = 1;
 	}
 
 if ($sys_id ne "")

INSTALL

@@ -140,8 +140,8 @@
      standard headers).  If it is a problem with OpenSSL itself, please
      report the problem to <openssl-bugs@openssl.org> (note that your
      message will be recorded in the request tracker publicly readable
-     via http://www.openssl.org/rt2.html and will be forwarded to a public
-     mailing list). Include the output of "make report" in your message.
+     via http://www.openssl.org/support/rt2.html and will be forwarded to a
+     public mailing list). Include the output of "make report" in your message.
      Please check out the request tracker. Maybe the bug was already
      reported or has already been fixed.
 
@@ -162,7 +162,7 @@
      in Makefile.ssl and run "make clean; make". Please send a bug
      report to <openssl-bugs@openssl.org>, including the output of
      "make report" in order to be added to the request tracker at
-     http://www.openssl.org/rt2.html.
+     http://www.openssl.org/support/rt2.html.
 
   4. If everything tests ok, install OpenSSL with
 

Makefile.org

@@ -15,6 +15,11 @@ OPTIONS=
 CONFIGURE_ARGS=
 SHLIB_TARGET=
 
+# HERE indicates where this Makefile lives.  This can be used to indicate
+# where sub-Makefiles are expected to be.  Currently has very limited usage,
+# and should probably not be bothered with at all.
+HERE=.
+
 # INSTALL_PREFIX is for package builders so that they can configure
 # for, say, /usr/ and yet have everything installed to /tmp/somedir/usr/.
 # Normally it is left empty.
@@ -274,10 +279,7 @@ do_gnu-shared:
 	libs="$$libs -l$$i"; \
 	done
 
-DETECT_GNU_LD=${CC} -v 2>&1 | grep '^gcc' >/dev/null 2>&1 && \
-	my_ld=`${CC} -print-prog-name=ld 2>&1` && \
-	[ -n "$$my_ld" ] && \
-	$$my_ld -v 2>&1 | grep 'GNU ld' >/dev/null 2>&1
+DETECT_GNU_LD=(${CC} -Wl,-V /dev/null 2>&1 | grep '^GNU ld' )>/dev/null
 
 # For Darwin AKA Mac OS/X (dyld)
 do_darwin-shared: 
@@ -355,10 +357,13 @@ do_solaris-shared:
 	else \
 		libs='-L. ${SHLIBDEPS}'; for i in ${SHLIBDIRS}; do \
 		( PATH=/usr/ccs/bin:$$PATH ; export PATH; \
-		  set -x; ${CC} ${SHARED_LDFLAGS} \
-			-G -o lib$$i.so.${SHLIB_MAJOR}.${SHLIB_MINOR} \
+		  MINUSZ='-z '; \
+		  (${CC} -v 2>&1 | grep gcc) > /dev/null && MINUSZ='-Wl,-z,'; \
+		  set -x; ${CC} ${SHARED_LDFLAGS} -G -dy -z text \
+			-o lib$$i.so.${SHLIB_MAJOR}.${SHLIB_MINOR} \
 			-h lib$$i.so.${SHLIB_MAJOR}.${SHLIB_MINOR} \
-			-z allextract lib$$i.a $$libs ${EX_LIBS} -lc ) || exit 1; \
+			$${MINUSZ}allextract lib$$i.a $${MINUSZ}defaultextract \
+			$$libs ${EX_LIBS} -lc ) || exit 1; \
 		libs="$$libs -l$$i"; \
 		done; \
 	fi
@@ -597,7 +602,7 @@ rehash.time: certs
 	@(OPENSSL="`pwd`/apps/openssl"; OPENSSL_DEBUG_MEMORY=on; \
 		export OPENSSL OPENSSL_DEBUG_MEMORY; \
 		LD_LIBRARY_PATH="`pwd`"; DYLD_LIBRARY_PATH="`pwd`"; SHLIB_PATH="`pwd`"; LIBPATH="`pwd`"; \
-		if [ "$(PLATFORM)" == "DJGPP" ]; then PATH="`pwd`\;$$PATH";  \
+		if [ "$(PLATFORM)" = "DJGPP" ]; then PATH="`pwd`\;$$PATH";  \
 		elif [ "$(PLATFORM)" != "Cygwin" ];  then PATH="`pwd`:$$PATH"; fi; \
 		export LD_LIBRARY_PATH DYLD_LIBRARY_PATH SHLIB_PATH LIBPATH PATH; \
 		$(PERL) tools/c_rehash certs)
@@ -609,7 +614,7 @@ tests: rehash
 	@(cd test && echo "testing..." && \
 	$(MAKE) CC='${CC}' PLATFORM='${PLATFORM}' CFLAG='${CFLAG}' SDIRS='$(SDIRS)' INSTALLTOP='${INSTALLTOP}' PEX_LIBS='${PEX_LIBS}' EX_LIBS='${EX_LIBS}' BN_ASM='${BN_ASM}' DES_ENC='${DES_ENC}' BF_ENC='${BF_ENC}' CAST_ENC='${CAST_ENC}' RC4_ENC='${RC4_ENC}' RC5_ENC='${RC5_ENC}' SHA1_ASM_OBJ='${SHA1_ASM_OBJ}' MD5_ASM_OBJ='${MD5_ASM_OBJ}' RMD160_ASM_OBJ='${RMD160_ASM_OBJ}' AR='${AR}' PROCESSOR='${PROCESSOR}' PERL='${PERL}' RANLIB='${RANLIB}' TESTS='${TESTS}' KRB5_INCLUDES='${KRB5_INCLUDES}' LIBKRB5='${LIBKRB5}' EXE_EXT='${EXE_EXT}' SHARED_LIBS='${SHARED_LIBS}' SHLIB_EXT='${SHLIB_EXT}' SHLIB_TARGET='${SHLIB_TARGET}' TESTS='${TESTS}' OPENSSL_DEBUG_MEMORY=on tests );
 	@LD_LIBRARY_PATH="`pwd`"; DYLD_LIBRARY_PATH="`pwd`"; SHLIB_PATH="`pwd`"; LIBPATH="`pwd`"; \
-		if [ "$(PLATFORM)" == "DJGPP" ]; then PATH="`pwd`\;$$PATH";  \
+		if [ "$(PLATFORM)" = "DJGPP" ]; then PATH="`pwd`\;$$PATH";  \
 		elif [ "$(PLATFORM)" != "Cygwin" ];  then PATH="`pwd`:$$PATH"; fi; \
 		export LD_LIBRARY_PATH DYLD_LIBRARY_PATH SHLIB_PATH LIBPATH PATH; \
 		apps/openssl version -a
@@ -731,7 +736,7 @@ install: all install_docs
 			cp $$i $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new; \
 			$(RANLIB) $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new; \
 			chmod 644 $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new; \
-			mv $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i ); \
+			mv -f $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i ); \
 		fi; \
 	done;
 	@if [ -n "$(SHARED_LIBS)" ]; then \
@@ -743,15 +748,15 @@ install: all install_docs
 				if [ "$(PLATFORM)" != "Cygwin" ]; then \
 					cp $$i $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new; \
 					chmod 555 $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new; \
-					mv $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i; \
+					mv -f $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.new $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i; \
 				else \
 					c=`echo $$i | sed 's/^lib\(.*\)/cyg\1-$(SHLIB_VERSION_NUMBER)/'`; \
 					cp $$c $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c.new; \
 					chmod 755 $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c.new; \
-					mv $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c.new $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c; \
+					mv -f $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c.new $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$c; \
 					cp $$i.a $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.a.new; \
 					chmod 644 $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.a.new; \
-					mv $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.a.new $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.a; \
+					mv -f $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.a.new $(INSTALL_PREFIX)$(INSTALLTOP)/lib/$$i.a; \
 				fi ); \
 			fi; \
 		done; \
@@ -769,6 +774,11 @@ install_docs:
 		$(INSTALL_PREFIX)$(MANDIR)/man5 \
 		$(INSTALL_PREFIX)$(MANDIR)/man7
 	@pod2man="`cd util; ./pod2mantest $(PERL)`"; \
+	here="`pwd`"; \
+	filecase=; \
+	if [ "$(PLATFORM)" = "DJGPP" -o "$(PLATFORM)" = "Cygwin" ]; then \
+		filecase=-i; \
+	fi; \
 	for i in doc/apps/*.pod; do \
 		fn=`basename $$i .pod`; \
 		if [ "$$fn" = "config" ]; then sec=5; else sec=1; fi; \
@@ -778,10 +788,12 @@ install_docs:
 			--section=$$sec --center=OpenSSL \
 			--release=$(VERSION) `basename $$i`") \
 			>  $(INSTALL_PREFIX)$(MANDIR)/man$$sec/$$fn.$$sec; \
-		$(PERL) util/extract-names.pl < $$i | grep -v "^$$fn" | \
-			while read n; do \
-				util/point.sh $$fn.$$sec $(INSTALL_PREFIX)$(MANDIR)/man$$sec/$$n.$$sec; \
-			done; \
+		$(PERL) util/extract-names.pl < $$i | \
+			grep -v $$filecase "^$$fn\$$" | \
+			(cd $(INSTALL_PREFIX)$(MANDIR)/man$$sec/; \
+			 while read n; do \
+				$$here/util/point.sh $$fn.$$sec $$n.$$sec; \
+			 done); \
 	done; \
 	for i in doc/crypto/*.pod doc/ssl/*.pod; do \
 		fn=`basename $$i .pod`; \
@@ -792,10 +804,12 @@ install_docs:
 			--section=$$sec --center=OpenSSL \
 			--release=$(VERSION) `basename $$i`") \
 			>  $(INSTALL_PREFIX)$(MANDIR)/man$$sec/$$fn.$$sec; \
-		$(PERL) util/extract-names.pl < $$i | grep -v "^$$fn" | \
-			while read n; do \
-				util/point.sh $$fn.$$sec $(INSTALL_PREFIX)$(MANDIR)/man$$sec/$$n.$$sec; \
-			done; \
+		$(PERL) util/extract-names.pl < $$i | \
+			grep -v $$filecase "^$$fn\$$" | \
+			(cd $(INSTALL_PREFIX)$(MANDIR)/man$$sec/; \
+			 while read n; do \
+				$$here/util/point.sh $$fn.$$sec $$n.$$sec; \
+			 done); \
 	done
 
 # DO NOT DELETE THIS LINE -- make depend depends on it.

NEWS

@@ -18,11 +18,14 @@
       o New elliptic curve library section.
       o New AES (Rijndael) library section.
       o Support for new platforms: Windows CE, Tandem OSS, A/UX, AIX 64-bit,
-        Linux x86_64
+        Linux x86_64, Linux 64-bit on Sparc v9
       o Extended support for some platforms: VxWorks
       o Enhanced support for shared libraries.
+      o Now only builds PIC code when shared library support is requested.
       o Support for pkg-config.
       o Lots of new manuals.
+      o Makes symbolic links to or copies of manuals to cover all described
+        functions.
       o Change DES API to clean up the namespace (some applications link also
         against libdes providing similar functions having the same name).
         Provide macros for backward compatibility (will be removed in the

README

@@ -1,5 +1,5 @@
 
- OpenSSL 0.9.7-beta5  5 Dec 2002
+ OpenSSL 0.9.7-beta6 17 Dec 2002
 
  Copyright (c) 1998-2002 The OpenSSL Project
  Copyright (c) 1995-1998 Eric A. Young, Tim J. Hudson

STATUS

@@ -1,10 +1,11 @@
 
   OpenSSL STATUS                           Last modified at
-  ______________                           $Date: 2002/12/05 22:51:17 $
+  ______________                           $Date: 2002/12/17 14:24:51 $
 
   DEVELOPMENT STATE
 
     o  OpenSSL 0.9.8:  Under development...
+    o  OpenSSL 0.9.7-beta6: Released on December 17th, 2002
     o  OpenSSL 0.9.7-beta5: Released on December  5th, 2002
     o  OpenSSL 0.9.7-beta4: Released on November 19th, 2002
     o  OpenSSL 0.9.7-beta3: Released on July     30th, 2002
@@ -67,9 +68,8 @@
 	UTIL (a new set of library functions to support some higher level
 	      functionality that is currently missing).
 	Shared library support for VMS.
-	Kerberos 5 authentication
+	Kerberos 5 authentication (Heimdal)
 	Constification
-	OCSP
 
   NEEDS PATCH
 

TABLE

@@ -1945,7 +1945,7 @@ $rc5_obj      =
 $dso_scheme   = dlfcn
 $shared_target= solaris-shared
 $shared_cflag = -fPIC
-$shared_ldflag = 
+$shared_ldflag = -shared
 $shared_extension = .so.$(SHLIB_MAJOR).$(SHLIB_MINOR)
 $ranlib       = 
 $arflags      = 
@@ -1995,7 +1995,7 @@ $rc5_obj      =
 $dso_scheme   = dlfcn
 $shared_target= solaris-shared
 $shared_cflag = -fPIC
-$shared_ldflag = 
+$shared_ldflag = -shared
 $shared_extension = .so.$(SHLIB_MAJOR).$(SHLIB_MINOR)
 $ranlib       = 
 $arflags      = 
@@ -3377,13 +3377,13 @@ $arflags      =
 
 *** linux-x86_64
 $cc           = gcc
-$cflags       = -DL_ENDIAN -DNO_ASM
+$cflags       = -m64 -DL_ENDIAN -DTERMIO -O3 -Wall -DMD32_REG_T=int
 $unistd       = 
 $thread_cflag = -D_REENTRANT
 $sys_id       = 
 $lflags       = -ldl
-$bn_ops       = SIXTY_FOUR_BIT_LONG
-$bn_obj       = 
+$bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHUNK RC4_CHAR BF_PTR2 DES_INT DES_UNROLL
+$bn_obj       = asm/x86_64-gcc.o
 $des_obj      = 
 $bf_obj       = 
 $md5_obj      = 
@@ -3395,7 +3395,32 @@ $rc5_obj      =
 $dso_scheme   = dlfcn
 $shared_target= linux-shared
 $shared_cflag = -fPIC
-$shared_ldflag = 
+$shared_ldflag = -m64
+$shared_extension = .so.$(SHLIB_MAJOR).$(SHLIB_MINOR)
+$ranlib       = 
+$arflags      = 
+
+*** linux64-sparcv9
+$cc           = gcc
+$cflags       = -m64 -mcpu=ultrasparc -DB_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -Wall
+$unistd       = 
+$thread_cflag = -D_REENTRANT
+$sys_id       = ULTRASPARC
+$lflags       = -ldl
+$bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR
+$bn_obj       = 
+$des_obj      = 
+$bf_obj       = 
+$md5_obj      = asm/md5-sparcv9.o
+$sha1_obj     = 
+$cast_obj     = 
+$rc4_obj      = 
+$rmd160_obj   = 
+$rc5_obj      = 
+$dso_scheme   = dlfcn
+$shared_target= linux-shared
+$shared_cflag = -fPIC
+$shared_ldflag = -m64
 $shared_extension = .so.$(SHLIB_MAJOR).$(SHLIB_MINOR)
 $ranlib       = 
 $arflags      = 
@@ -3770,7 +3795,7 @@ $rc5_obj      =
 $dso_scheme   = dlfcn
 $shared_target= solaris-shared
 $shared_cflag = -fPIC
-$shared_ldflag = 
+$shared_ldflag = -shared
 $shared_extension = .so.$(SHLIB_MAJOR).$(SHLIB_MINOR)
 $ranlib       = 
 $arflags      = 
@@ -3820,7 +3845,7 @@ $rc5_obj      =
 $dso_scheme   = dlfcn
 $shared_target= solaris-shared
 $shared_cflag = -fPIC
-$shared_ldflag = 
+$shared_ldflag = -shared
 $shared_extension = .so.$(SHLIB_MAJOR).$(SHLIB_MINOR)
 $ranlib       = 
 $arflags      = 
@@ -3870,7 +3895,7 @@ $rc5_obj      =
 $dso_scheme   = dlfcn
 $shared_target= solaris-shared
 $shared_cflag = -fPIC
-$shared_ldflag = 
+$shared_ldflag = -shared
 $shared_extension = .so.$(SHLIB_MAJOR).$(SHLIB_MINOR)
 $ranlib       = 
 $arflags      = 
@@ -3895,7 +3920,7 @@ $rc5_obj      =
 $dso_scheme   = dlfcn
 $shared_target= solaris-shared
 $shared_cflag = -fPIC
-$shared_ldflag = 
+$shared_ldflag = -shared
 $shared_extension = .so.$(SHLIB_MAJOR).$(SHLIB_MINOR)
 $ranlib       = 
 $arflags      = 
@@ -3945,7 +3970,7 @@ $rc5_obj      = asm/r586-sol.o
 $dso_scheme   = dlfcn
 $shared_target= solaris-shared
 $shared_cflag = -fPIC
-$shared_ldflag = 
+$shared_ldflag = -shared
 $shared_extension = .so.$(SHLIB_MAJOR).$(SHLIB_MINOR)
 $ranlib       = 
 $arflags      = 
@@ -3995,7 +4020,7 @@ $rc5_obj      =
 $dso_scheme   = dlfcn
 $shared_target= solaris-shared
 $shared_cflag = -fPIC
-$shared_ldflag = -m64
+$shared_ldflag = -m64 -shared
 $shared_extension = .so.$(SHLIB_MAJOR).$(SHLIB_MINOR)
 $ranlib       = 
 $arflags      = 
@@ -4020,7 +4045,7 @@ $rc5_obj      =
 $dso_scheme   = dlfcn
 $shared_target= solaris-shared
 $shared_cflag = -fPIC
-$shared_ldflag = -m64
+$shared_ldflag = -m64 -shared
 $shared_extension = .so.$(SHLIB_MAJOR).$(SHLIB_MINOR)
 $ranlib       = 
 $arflags      = 
@@ -4077,7 +4102,7 @@ $arflags      =
 
 *** ultrix-cc
 $cc           = cc
-$cflags       = -std1 -O -Olimit 1000 -DL_ENDIAN
+$cflags       = -std1 -O -Olimit 2500 -DL_ENDIAN
 $unistd       = 
 $thread_cflag = (unknown)
 $sys_id       = 

apps/Makefile.ssl

@@ -99,17 +99,20 @@ install:
 	@for i in $(EXE); \
 	do  \
 	(echo installing $$i; \
-	 cp $$i $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$i; \
-	 chmod 755 $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$i ); \
+	 cp $$i $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$i.new; \
+	 chmod 755 $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$i.new; \
+	 mv -f $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$i.new $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$i ); \
 	 done;
 	@for i in $(SCRIPTS); \
 	do  \
 	(echo installing $$i; \
-	 cp $$i $(INSTALL_PREFIX)$(OPENSSLDIR)/misc/$$i; \
-	 chmod 755 $(INSTALL_PREFIX)$(OPENSSLDIR)/misc/$$i ); \
+	 cp $$i $(INSTALL_PREFIX)$(OPENSSLDIR)/misc/$$i.new; \
+	 chmod 755 $(INSTALL_PREFIX)$(OPENSSLDIR)/misc/$$i.new; \
+	 mv -f $(INSTALL_PREFIX)$(INSTALLTOP)/misc/$$i.new $(INSTALL_PREFIX)$(INSTALLTOP)/misc/$$i ); \
 	 done
-	@cp openssl.cnf $(INSTALL_PREFIX)$(OPENSSLDIR); \
-	chmod 644 $(INSTALL_PREFIX)$(OPENSSLDIR)/openssl.cnf
+	@cp openssl.cnf $(INSTALL_PREFIX)$(OPENSSLDIR)/openssl.cnf.new; \
+	chmod 644 $(INSTALL_PREFIX)$(OPENSSLDIR)/openssl.cnf.new; \
+	mv -f  $(INSTALL_PREFIX)$(OPENSSLDIR)/openssl.cnf.new $(INSTALL_PREFIX)$(OPENSSLDIR)/openssl.cnf
 
 tags:
 	ctags $(SRC)
@@ -148,7 +151,7 @@ $(PROGRAM): progs.h $(E_OBJ) $(PROGRAM).o $(DLIBCRYPTO) $(DLIBSSL)
 	fi
 	-(cd ..; OPENSSL="`pwd`/apps/openssl"; export OPENSSL; \
 		LIBPATH="`pwd`"; LD_LIBRARY_PATH="`pwd`"; DYLD_LIBRARY_PATH="`pwd`"; SHLIB_PATH="`pwd`"; \
-		if [ "$(PLATFORM)" == "DJGPP" ]; then PATH="`pwd`\;$$PATH";  \
+		if [ "$(PLATFORM)" = "DJGPP" ]; then PATH="`pwd`\;$$PATH";  \
 		elif [ "$(PLATFORM)" != "Cygwin" ];  then PATH="`pwd`:$$PATH"; fi; \
 		export LD_LIBRARY_PATH DYLD_LIBRARY_PATH SHLIB_PATH LIBPATH PATH; \
 		$(PERL) tools/c_rehash certs)

config

@@ -547,12 +547,13 @@ EOF
   ppc-apple-darwin*) OUT="darwin-ppc-cc" ;;
   i386-apple-darwin*) OUT="darwin-i386-cc" ;;
   sparc64-*-linux2)
-	#Before we can uncomment following lines we have to wait at least
-	#till 64-bit glibc for SPARC is operational:-(
-	#echo "WARNING! If you wish to build 64-bit library, then you have to"
-	#echo "         invoke './Configure linux64-sparcv9' *manually*."
-	#echo "         Type return if you want to continue, Ctrl-C to abort."
-	#read waste < /dev/tty
+	echo "WARNING! If *know* that your GNU C supports 64-bit/V9 ABI"
+	echo "         and wish to build 64-bit library, then you have to"
+	echo "         invoke './Configure linux64-sparcv9' *manually*."
+	if [ "$TEST" = "false" ]; then
+	  echo "          You have about 5 seconds to press Ctrl-C to abort."
+	  (stty -icanon min 0 time 50; read waste) < /dev/tty
+	fi
 	OUT="linux-sparcv9" ;;
   sparc-*-linux2)
 	KARCH=`awk '/^type/{print$3}' /proc/cpuinfo`

crypto/asn1/a_d2i_fp.c

@@ -149,7 +149,12 @@ static int asn1_d2i_read_bio(BIO *in, BUF_MEM **pb)
 	ASN1_CTX c;
 	int want=HEADER_SIZE;
 	int eos=0;
+#if defined(__GNUC__) && defined(__ia64)
+	/* pathetic compiler bug in all known versions as of Nov. 2002 */
+	long off=0;
+#else
 	int off=0;
+#endif
 	int len=0;
 
 	b=BUF_MEM_new();

crypto/bn/Makefile.ssl

@@ -136,6 +136,8 @@ asm/ia64-cpp.o:	asm/ia64.S
 	$(CC) $(ASFLAGS) -c -o asm/ia64-cpp.o /tmp/ia64.$$$$.s;	\
 	rm -f /tmp/ia64.$$$$.s
 
+asm/x86_64-gcc.o: asm/x86_64-gcc.c
+
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile.ssl >> $(TOP)/MINFO
 

crypto/bn/asm/x86_64-gcc.c

@@ -0,0 +1,575 @@
+/*
+ * x86_64 BIGNUM accelerator version 0.1, December 2002.
+ *
+ * Implemented by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+ * project.
+ *
+ * Rights for redistribution and usage in source and binary forms are
+ * granted according to the OpenSSL license. Warranty of any kind is
+ * disclaimed.
+ *
+ * Q. Version 0.1? It doesn't sound like Andy, he used to assign real
+ *    versions, like 1.0...
+ * A. Well, that's because this code is basically a quick-n-dirty
+ *    proof-of-concept hack. As you can see it's implemented with
+ *    inline assembler, which means that you're bound to GCC and that
+ *    there must be a room for fine-tuning.
+ *
+ * Q. Why inline assembler?
+ * A. x86_64 features own ABI I'm not familiar with. Which is why
+ *    I decided to let the compiler take care of subroutine
+ *    prologue/epilogue as well as register allocation.
+ *
+ * Q. How much faster does it get?
+ * A. Unfortunately people sitting on x86_64 hardware are prohibited
+ *    to disclose the performance numbers, so they (SuSE labs to be
+ *    specific) wouldn't tell me. However! Very similar coding technique
+ *    (reaching out for 128-bit result from 64x64-bit multiplication)
+ *    results in >3 times performance improvement on MIPS and I see no
+ *    reason why gain on x86_64 would be so much different:-)
+ */
+
+#define BN_ULONG unsigned long
+
+/*
+ * "m"(a), "+m"(r)	is the way to favor DirectPath <20>-code;
+ * "g"(0)		let the compiler to decide where does it
+ *			want to keep the value of zero;
+ */
+#define mul_add(r,a,word,carry) do {	\
+	register BN_ULONG high,low;	\
+	asm ("mulq %3"			\
+		: "=a"(low),"=d"(high)	\
+		: "a"(word),"m"(a)	\
+		: "cc");		\
+	asm ("addq %2,%0; adcq %3,%1"	\
+		: "+r"(carry),"+d"(high)\
+		: "a"(low),"g"(0)	\
+		: "cc");		\
+	asm ("addq %2,%0; adcq %3,%1"	\
+		: "+m"(r),"+d"(high)	\
+		: "r"(carry),"g"(0)	\
+		: "cc");		\
+	carry=high;			\
+	} while (0)
+
+#define mul(r,a,word,carry) do {	\
+	register BN_ULONG high,low;	\
+	asm ("mulq %3"			\
+		: "=a"(low),"=d"(high)	\
+		: "a"(word),"g"(a)	\
+		: "cc");		\
+	asm ("addq %2,%0; adcq %3,%1"	\
+		: "+r"(carry),"+d"(high)\
+		: "a"(low),"g"(0)	\
+		: "cc");		\
+	(r)=carry, carry=high;		\
+	} while (0)
+
+#define sqr(r0,r1,a)			\
+	asm ("mulq %2"			\
+		: "=a"(r0),"=d"(r1)	\
+		: "a"(a)		\
+		: "cc");
+
+BN_ULONG bn_mul_add_words(BN_ULONG *rp, BN_ULONG *ap, int num, BN_ULONG w)
+	{
+	BN_ULONG c1=0;
+
+	if (num <= 0) return(c1);
+
+	while (num&~3)
+		{
+		mul_add(rp[0],ap[0],w,c1);
+		mul_add(rp[1],ap[1],w,c1);
+		mul_add(rp[2],ap[2],w,c1);
+		mul_add(rp[3],ap[3],w,c1);
+		ap+=4; rp+=4; num-=4;
+		}
+	if (num)
+		{
+		mul_add(rp[0],ap[0],w,c1); if (--num==0) return c1;
+		mul_add(rp[1],ap[1],w,c1); if (--num==0) return c1;
+		mul_add(rp[2],ap[2],w,c1); return c1;
+		}
+	
+	return(c1);
+	} 
+
+BN_ULONG bn_mul_words(BN_ULONG *rp, BN_ULONG *ap, int num, BN_ULONG w)
+	{
+	BN_ULONG c1=0;
+
+	if (num <= 0) return(c1);
+
+	while (num&~3)
+		{
+		mul(rp[0],ap[0],w,c1);
+		mul(rp[1],ap[1],w,c1);
+		mul(rp[2],ap[2],w,c1);
+		mul(rp[3],ap[3],w,c1);
+		ap+=4; rp+=4; num-=4;
+		}
+	if (num)
+		{
+		mul(rp[0],ap[0],w,c1); if (--num == 0) return c1;
+		mul(rp[1],ap[1],w,c1); if (--num == 0) return c1;
+		mul(rp[2],ap[2],w,c1);
+		}
+	return(c1);
+	} 
+
+void bn_sqr_words(BN_ULONG *r, BN_ULONG *a, int n)
+        {
+	if (n <= 0) return;
+
+	while (n&~3)
+		{
+		sqr(r[0],r[1],a[0]);
+		sqr(r[2],r[3],a[1]);
+		sqr(r[4],r[5],a[2]);
+		sqr(r[6],r[7],a[3]);
+		a+=4; r+=8; n-=4;
+		}
+	if (n)
+		{
+		sqr(r[0],r[1],a[0]); if (--n == 0) return;
+		sqr(r[2],r[3],a[1]); if (--n == 0) return;
+		sqr(r[4],r[5],a[2]);
+		}
+	}
+
+BN_ULONG bn_div_words(BN_ULONG h, BN_ULONG l, BN_ULONG d)
+{	BN_ULONG ret,waste;
+
+	asm ("divq	%3"
+		: "=a"(ret),"=d"(waste)
+		: "a"(l),"d"(h),"g"(d)
+		: "cc");
+
+	return ret;
+}
+
+BN_ULONG bn_add_words (BN_ULONG *rp, BN_ULONG *ap, BN_ULONG *bp,int n)
+{ BN_ULONG ret,i;
+
+	if (n <= 0) return 0;
+
+	asm (
+	"	subq	%2,%2		\n"
+	".align 16			\n"
+	"1:	movq	(%4,%2,8),%0	\n"
+	"	adcq	(%5,%2,8),%0	\n"
+	"	movq	%0,(%3,%2,8)	\n"
+	"	leaq	1(%2),%2	\n"
+	"	loop	1b		\n"
+	"	sbbq	%0,%0		\n"
+		: "+a"(ret),"+c"(n),"+r"(i)
+		: "r"(rp),"r"(ap),"r"(bp)
+		: "cc"
+	);
+
+  return ret&1;
+}
+
+#ifndef SIMICS
+BN_ULONG bn_sub_words (BN_ULONG *rp, BN_ULONG *ap, BN_ULONG *bp,int n)
+{ BN_ULONG ret,i;
+
+	if (n <= 0) return 0;
+
+	asm (
+	"	subq	%2,%2		\n"
+	".align 16			\n"
+	"1:	movq	(%4,%2,8),%0	\n"
+	"	sbbq	(%5,%2,8),%0	\n"
+	"	movq	%0,(%3,%2,8)	\n"
+	"	leaq	1(%2),%2	\n"
+	"	loop	1b		\n"
+	"	sbbq	%0,%0		\n"
+		: "+a"(ret),"+c"(n),"+r"(i)
+		: "r"(rp),"r"(ap),"r"(bp)
+		: "cc"
+	);
+
+  return ret&1;
+}
+#else
+/* Simics 1.4<7 has buggy sbbq:-( */
+#define BN_MASK2 0xffffffffffffffffL
+BN_ULONG bn_sub_words(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n)
+        {
+	BN_ULONG t1,t2;
+	int c=0;
+
+	if (n <= 0) return((BN_ULONG)0);
+
+	for (;;)
+		{
+		t1=a[0]; t2=b[0];
+		r[0]=(t1-t2-c)&BN_MASK2;
+		if (t1 != t2) c=(t1 < t2);
+		if (--n <= 0) break;
+
+		t1=a[1]; t2=b[1];
+		r[1]=(t1-t2-c)&BN_MASK2;
+		if (t1 != t2) c=(t1 < t2);
+		if (--n <= 0) break;
+
+		t1=a[2]; t2=b[2];
+		r[2]=(t1-t2-c)&BN_MASK2;
+		if (t1 != t2) c=(t1 < t2);
+		if (--n <= 0) break;
+
+		t1=a[3]; t2=b[3];
+		r[3]=(t1-t2-c)&BN_MASK2;
+		if (t1 != t2) c=(t1 < t2);
+		if (--n <= 0) break;
+
+		a+=4;
+		b+=4;
+		r+=4;
+		}
+	return(c);
+	}
+#endif
+
+/* mul_add_c(a,b,c0,c1,c2)  -- c+=a*b for three word number c=(c2,c1,c0) */
+/* mul_add_c2(a,b,c0,c1,c2) -- c+=2*a*b for three word number c=(c2,c1,c0) */
+/* sqr_add_c(a,i,c0,c1,c2)  -- c+=a[i]^2 for three word number c=(c2,c1,c0) */
+/* sqr_add_c2(a,i,c0,c1,c2) -- c+=2*a[i]*a[j] for three word number c=(c2,c1,c0) */
+
+#if 0
+/* original macros are kept for reference purposes */
+#define mul_add_c(a,b,c0,c1,c2) {	\
+	BN_ULONG ta=(a),tb=(b);		\
+	t1 = ta * tb;			\
+	t2 = BN_UMULT_HIGH(ta,tb);	\
+	c0 += t1; t2 += (c0<t1)?1:0;	\
+	c1 += t2; c2 += (c1<t2)?1:0;	\
+	}
+
+#define mul_add_c2(a,b,c0,c1,c2) {	\
+	BN_ULONG ta=(a),tb=(b),t0;	\
+	t1 = BN_UMULT_HIGH(ta,tb);	\
+	t0 = ta * tb;			\
+	t2 = t1+t1; c2 += (t2<t1)?1:0;	\
+	t1 = t0+t0; t2 += (t1<t0)?1:0;	\
+	c0 += t1; t2 += (c0<t1)?1:0;	\
+	c1 += t2; c2 += (c1<t2)?1:0;	\
+	}
+#else
+#define mul_add_c(a,b,c0,c1,c2)	do {	\
+	asm ("mulq %3"			\
+		: "=a"(t1),"=d"(t2)	\
+		: "a"(a),"m"(b)		\
+		: "cc");		\
+	asm ("addq %2,%0; adcq %3,%1"	\
+		: "+r"(c0),"+d"(t2)	\
+		: "a"(t1),"g"(0)	\
+		: "cc");		\
+	asm ("addq %2,%0; adcq %3,%1"	\
+		: "+r"(c1),"+r"(c2)	\
+		: "d"(t2),"g"(0)	\
+		: "cc");		\
+	} while (0)
+
+#define sqr_add_c(a,i,c0,c1,c2)	do {	\
+	asm ("mulq %2"			\
+		: "=a"(t1),"=d"(t2)	\
+		: "a"(a[i])		\
+		: "cc");		\
+	asm ("addq %2,%0; adcq %3,%1"	\
+		: "+r"(c0),"+d"(t2)	\
+		: "a"(t1),"g"(0)	\
+		: "cc");		\
+	asm ("addq %2,%0; adcq %3,%1"	\
+		: "+r"(c1),"+r"(c2)	\
+		: "d"(t2),"g"(0)	\
+		: "cc");		\
+	} while (0)
+
+#define mul_add_c2(a,b,c0,c1,c2) do {	\
+	asm ("mulq %3"			\
+		: "=a"(t1),"=d"(t2)	\
+		: "a"(a),"m"(b)		\
+		: "cc");		\
+	asm ("addq %0,%0; adcq %2,%1"	\
+		: "+d"(t2),"+r"(c2)	\
+		: "g"(0)		\
+		: "cc");		\
+	asm ("addq %0,%0; adcq %2,%1"	\
+		: "+a"(t1),"+d"(t2)	\
+		: "g"(0)		\
+		: "cc");		\
+	asm ("addq %2,%0; adcq %3,%1"	\
+		: "+r"(c0),"+d"(t2)	\
+		: "a"(t1),"g"(0)	\
+		: "cc");		\
+	asm ("addq %2,%0; adcq %3,%1"	\
+		: "+r"(c1),"+r"(c2)	\
+		: "d"(t2),"g"(0)	\
+		: "cc");		\
+	} while (0)
+#endif
+
+#define sqr_add_c2(a,i,j,c0,c1,c2)	\
+	mul_add_c2((a)[i],(a)[j],c0,c1,c2)
+
+void bn_mul_comba8(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
+	{
+	BN_ULONG bl,bh;
+	BN_ULONG t1,t2;
+	BN_ULONG c1,c2,c3;
+
+	c1=0;
+	c2=0;
+	c3=0;
+	mul_add_c(a[0],b[0],c1,c2,c3);
+	r[0]=c1;
+	c1=0;
+	mul_add_c(a[0],b[1],c2,c3,c1);
+	mul_add_c(a[1],b[0],c2,c3,c1);
+	r[1]=c2;
+	c2=0;
+	mul_add_c(a[2],b[0],c3,c1,c2);
+	mul_add_c(a[1],b[1],c3,c1,c2);
+	mul_add_c(a[0],b[2],c3,c1,c2);
+	r[2]=c3;
+	c3=0;
+	mul_add_c(a[0],b[3],c1,c2,c3);
+	mul_add_c(a[1],b[2],c1,c2,c3);
+	mul_add_c(a[2],b[1],c1,c2,c3);
+	mul_add_c(a[3],b[0],c1,c2,c3);
+	r[3]=c1;
+	c1=0;
+	mul_add_c(a[4],b[0],c2,c3,c1);
+	mul_add_c(a[3],b[1],c2,c3,c1);
+	mul_add_c(a[2],b[2],c2,c3,c1);
+	mul_add_c(a[1],b[3],c2,c3,c1);
+	mul_add_c(a[0],b[4],c2,c3,c1);
+	r[4]=c2;
+	c2=0;
+	mul_add_c(a[0],b[5],c3,c1,c2);
+	mul_add_c(a[1],b[4],c3,c1,c2);
+	mul_add_c(a[2],b[3],c3,c1,c2);
+	mul_add_c(a[3],b[2],c3,c1,c2);
+	mul_add_c(a[4],b[1],c3,c1,c2);
+	mul_add_c(a[5],b[0],c3,c1,c2);
+	r[5]=c3;
+	c3=0;
+	mul_add_c(a[6],b[0],c1,c2,c3);
+	mul_add_c(a[5],b[1],c1,c2,c3);
+	mul_add_c(a[4],b[2],c1,c2,c3);
+	mul_add_c(a[3],b[3],c1,c2,c3);
+	mul_add_c(a[2],b[4],c1,c2,c3);
+	mul_add_c(a[1],b[5],c1,c2,c3);
+	mul_add_c(a[0],b[6],c1,c2,c3);
+	r[6]=c1;
+	c1=0;
+	mul_add_c(a[0],b[7],c2,c3,c1);
+	mul_add_c(a[1],b[6],c2,c3,c1);
+	mul_add_c(a[2],b[5],c2,c3,c1);
+	mul_add_c(a[3],b[4],c2,c3,c1);
+	mul_add_c(a[4],b[3],c2,c3,c1);
+	mul_add_c(a[5],b[2],c2,c3,c1);
+	mul_add_c(a[6],b[1],c2,c3,c1);
+	mul_add_c(a[7],b[0],c2,c3,c1);
+	r[7]=c2;
+	c2=0;
+	mul_add_c(a[7],b[1],c3,c1,c2);
+	mul_add_c(a[6],b[2],c3,c1,c2);
+	mul_add_c(a[5],b[3],c3,c1,c2);
+	mul_add_c(a[4],b[4],c3,c1,c2);
+	mul_add_c(a[3],b[5],c3,c1,c2);
+	mul_add_c(a[2],b[6],c3,c1,c2);
+	mul_add_c(a[1],b[7],c3,c1,c2);
+	r[8]=c3;
+	c3=0;
+	mul_add_c(a[2],b[7],c1,c2,c3);
+	mul_add_c(a[3],b[6],c1,c2,c3);
+	mul_add_c(a[4],b[5],c1,c2,c3);
+	mul_add_c(a[5],b[4],c1,c2,c3);
+	mul_add_c(a[6],b[3],c1,c2,c3);
+	mul_add_c(a[7],b[2],c1,c2,c3);
+	r[9]=c1;
+	c1=0;
+	mul_add_c(a[7],b[3],c2,c3,c1);
+	mul_add_c(a[6],b[4],c2,c3,c1);
+	mul_add_c(a[5],b[5],c2,c3,c1);
+	mul_add_c(a[4],b[6],c2,c3,c1);
+	mul_add_c(a[3],b[7],c2,c3,c1);
+	r[10]=c2;
+	c2=0;
+	mul_add_c(a[4],b[7],c3,c1,c2);
+	mul_add_c(a[5],b[6],c3,c1,c2);
+	mul_add_c(a[6],b[5],c3,c1,c2);
+	mul_add_c(a[7],b[4],c3,c1,c2);
+	r[11]=c3;
+	c3=0;
+	mul_add_c(a[7],b[5],c1,c2,c3);
+	mul_add_c(a[6],b[6],c1,c2,c3);
+	mul_add_c(a[5],b[7],c1,c2,c3);
+	r[12]=c1;
+	c1=0;
+	mul_add_c(a[6],b[7],c2,c3,c1);
+	mul_add_c(a[7],b[6],c2,c3,c1);
+	r[13]=c2;
+	c2=0;
+	mul_add_c(a[7],b[7],c3,c1,c2);
+	r[14]=c3;
+	r[15]=c1;
+	}
+
+void bn_mul_comba4(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b)
+	{
+	BN_ULONG bl,bh;
+	BN_ULONG t1,t2;
+	BN_ULONG c1,c2,c3;
+
+	c1=0;
+	c2=0;
+	c3=0;
+	mul_add_c(a[0],b[0],c1,c2,c3);
+	r[0]=c1;
+	c1=0;
+	mul_add_c(a[0],b[1],c2,c3,c1);
+	mul_add_c(a[1],b[0],c2,c3,c1);
+	r[1]=c2;
+	c2=0;
+	mul_add_c(a[2],b[0],c3,c1,c2);
+	mul_add_c(a[1],b[1],c3,c1,c2);
+	mul_add_c(a[0],b[2],c3,c1,c2);
+	r[2]=c3;
+	c3=0;
+	mul_add_c(a[0],b[3],c1,c2,c3);
+	mul_add_c(a[1],b[2],c1,c2,c3);
+	mul_add_c(a[2],b[1],c1,c2,c3);
+	mul_add_c(a[3],b[0],c1,c2,c3);
+	r[3]=c1;
+	c1=0;
+	mul_add_c(a[3],b[1],c2,c3,c1);
+	mul_add_c(a[2],b[2],c2,c3,c1);
+	mul_add_c(a[1],b[3],c2,c3,c1);
+	r[4]=c2;
+	c2=0;
+	mul_add_c(a[2],b[3],c3,c1,c2);
+	mul_add_c(a[3],b[2],c3,c1,c2);
+	r[5]=c3;
+	c3=0;
+	mul_add_c(a[3],b[3],c1,c2,c3);
+	r[6]=c1;
+	r[7]=c2;
+	}
+
+void bn_sqr_comba8(BN_ULONG *r, BN_ULONG *a)
+	{
+	BN_ULONG bl,bh;
+	BN_ULONG t1,t2;
+	BN_ULONG c1,c2,c3;
+
+	c1=0;
+	c2=0;
+	c3=0;
+	sqr_add_c(a,0,c1,c2,c3);
+	r[0]=c1;
+	c1=0;
+	sqr_add_c2(a,1,0,c2,c3,c1);
+	r[1]=c2;
+	c2=0;
+	sqr_add_c(a,1,c3,c1,c2);
+	sqr_add_c2(a,2,0,c3,c1,c2);
+	r[2]=c3;
+	c3=0;
+	sqr_add_c2(a,3,0,c1,c2,c3);
+	sqr_add_c2(a,2,1,c1,c2,c3);
+	r[3]=c1;
+	c1=0;
+	sqr_add_c(a,2,c2,c3,c1);
+	sqr_add_c2(a,3,1,c2,c3,c1);
+	sqr_add_c2(a,4,0,c2,c3,c1);
+	r[4]=c2;
+	c2=0;
+	sqr_add_c2(a,5,0,c3,c1,c2);
+	sqr_add_c2(a,4,1,c3,c1,c2);
+	sqr_add_c2(a,3,2,c3,c1,c2);
+	r[5]=c3;
+	c3=0;
+	sqr_add_c(a,3,c1,c2,c3);
+	sqr_add_c2(a,4,2,c1,c2,c3);
+	sqr_add_c2(a,5,1,c1,c2,c3);
+	sqr_add_c2(a,6,0,c1,c2,c3);
+	r[6]=c1;
+	c1=0;
+	sqr_add_c2(a,7,0,c2,c3,c1);
+	sqr_add_c2(a,6,1,c2,c3,c1);
+	sqr_add_c2(a,5,2,c2,c3,c1);
+	sqr_add_c2(a,4,3,c2,c3,c1);
+	r[7]=c2;
+	c2=0;
+	sqr_add_c(a,4,c3,c1,c2);
+	sqr_add_c2(a,5,3,c3,c1,c2);
+	sqr_add_c2(a,6,2,c3,c1,c2);
+	sqr_add_c2(a,7,1,c3,c1,c2);
+	r[8]=c3;
+	c3=0;
+	sqr_add_c2(a,7,2,c1,c2,c3);
+	sqr_add_c2(a,6,3,c1,c2,c3);
+	sqr_add_c2(a,5,4,c1,c2,c3);
+	r[9]=c1;
+	c1=0;
+	sqr_add_c(a,5,c2,c3,c1);
+	sqr_add_c2(a,6,4,c2,c3,c1);
+	sqr_add_c2(a,7,3,c2,c3,c1);
+	r[10]=c2;
+	c2=0;
+	sqr_add_c2(a,7,4,c3,c1,c2);
+	sqr_add_c2(a,6,5,c3,c1,c2);
+	r[11]=c3;
+	c3=0;
+	sqr_add_c(a,6,c1,c2,c3);
+	sqr_add_c2(a,7,5,c1,c2,c3);
+	r[12]=c1;
+	c1=0;
+	sqr_add_c2(a,7,6,c2,c3,c1);
+	r[13]=c2;
+	c2=0;
+	sqr_add_c(a,7,c3,c1,c2);
+	r[14]=c3;
+	r[15]=c1;
+	}
+
+void bn_sqr_comba4(BN_ULONG *r, BN_ULONG *a)
+	{
+	BN_ULONG bl,bh;
+	BN_ULONG t1,t2;
+	BN_ULONG c1,c2,c3;
+
+	c1=0;
+	c2=0;
+	c3=0;
+	sqr_add_c(a,0,c1,c2,c3);
+	r[0]=c1;
+	c1=0;
+	sqr_add_c2(a,1,0,c2,c3,c1);
+	r[1]=c2;
+	c2=0;
+	sqr_add_c(a,1,c3,c1,c2);
+	sqr_add_c2(a,2,0,c3,c1,c2);
+	r[2]=c3;
+	c3=0;
+	sqr_add_c2(a,3,0,c1,c2,c3);
+	sqr_add_c2(a,2,1,c1,c2,c3);
+	r[3]=c1;
+	c1=0;
+	sqr_add_c(a,2,c2,c3,c1);
+	sqr_add_c2(a,3,1,c2,c3,c1);
+	r[4]=c2;
+	c2=0;
+	sqr_add_c2(a,3,2,c3,c1,c2);
+	r[5]=c3;
+	c3=0;
+	sqr_add_c(a,3,c1,c2,c3);
+	r[6]=c1;
+	r[7]=c2;
+	}

crypto/bn/bn_div.c

@@ -150,6 +150,20 @@ int BN_div(BIGNUM *dv, BIGNUM *rem, const BIGNUM *m, const BIGNUM *d,
 	    q;					\
 	})
 #  define REMAINDER_IS_ALREADY_CALCULATED
+#  elif defined(__x86_64) && defined(SIXTY_FOUR_BIT_LONG)
+   /*
+    * Same story here, but it's 128-bit by 64-bit division. Wow!
+    *					<appro@fy.chalmers.se>
+    */
+#  define bn_div_words(n0,n1,d0)		\
+	({  asm volatile (			\
+		"divq	%4"			\
+		: "=a"(q), "=d"(rem)		\
+		: "a"(n1), "d"(n0), "g"(d0)	\
+		: "cc");			\
+	    q;					\
+	})
+#  define REMAINDER_IS_ALREADY_CALCULATED
 #  endif /* __<cpu> */
 # endif /* __GNUC__ */
 #endif /* OPENSSL_NO_ASM */
@@ -296,7 +310,9 @@ int BN_div(BIGNUM *dv, BIGNUM *rm, const BIGNUM *num, const BIGNUM *divisor,
 			rem=(n1-q*d0)&BN_MASK2;
 #endif
 
-#ifdef BN_UMULT_HIGH
+#if defined(BN_UMULT_LOHI)
+			BN_UMULT_LOHI(t2l,t2h,d1,q);
+#elif defined(BN_UMULT_HIGH)
 			t2l = d1 * q;
 			t2h = BN_UMULT_HIGH(d1,q);
 #else

crypto/bn/bn_lcl.h

@@ -230,6 +230,21 @@ struct bignum_ctx
 	     : "r"(a), "r"(b));		\
 	ret;			})
 #  endif	/* compiler */
+# elif defined(__x86_64) && defined(SIXTY_FOUR_BIT_LONG)
+#  if defined(__GNUC__)
+#   define BN_UMULT_HIGH(a,b)	({	\
+	register BN_ULONG ret,discard;	\
+	asm ("mulq	%3"		\
+	     : "=a"(discard),"=d"(ret)	\
+	     : "a"(a), "g"(b)		\
+	     : "cc");			\
+	ret;			})
+#   define BN_UMULT_LOHI(low,high,a,b)	\
+	asm ("mulq	%3"		\
+		: "=a"(low),"=d"(high)	\
+		: "a"(a),"g"(b)		\
+		: "cc");
+#  endif
 # endif		/* cpu */
 #endif		/* OPENSSL_NO_ASM */
 
@@ -337,7 +352,7 @@ struct bignum_ctx
 
 #define LBITS(a)	((a)&BN_MASK2l)
 #define HBITS(a)	(((a)>>BN_BITS4)&BN_MASK2l)
-#define	L2HBITS(a)	((BN_ULONG)((a)&BN_MASK2l)<<BN_BITS4)
+#define	L2HBITS(a)	(((a)<<BN_BITS4)&BN_MASK2)
 
 #define LLBITS(a)	((a)&BN_MASKl)
 #define LHBITS(a)	(((a)>>BN_BITS2)&BN_MASKl)
@@ -418,10 +433,13 @@ void bn_sqr_comba4(BN_ULONG *r,const BN_ULONG *a);
 int bn_cmp_words(const BN_ULONG *a,const BN_ULONG *b,int n);
 int bn_cmp_part_words(const BN_ULONG *a, const BN_ULONG *b,
 	int cl, int dl);
+#if 0
+/* bn_mul.c rollback <appro> */
 void bn_mul_recursive(BN_ULONG *r,BN_ULONG *a,BN_ULONG *b,int n2,
 	int dna,int dnb,BN_ULONG *t);
 void bn_mul_part_recursive(BN_ULONG *r,BN_ULONG *a,BN_ULONG *b,
 	int n,int tna,int tnb,BN_ULONG *t);
+#endif
 void bn_sqr_recursive(BN_ULONG *r,const BN_ULONG *a, int n2, BN_ULONG *t);
 void bn_mul_low_normal(BN_ULONG *r,BN_ULONG *a,BN_ULONG *b, int n);
 void bn_mul_low_recursive(BN_ULONG *r,BN_ULONG *a,BN_ULONG *b,int n2,

crypto/bn/bn_mul.c

@@ -56,325 +56,10 @@
  * [including the GNU Public Licence.]
  */
 
-#ifndef BN_DEBUG
-# undef NDEBUG /* avoid conflicting definitions */
-# define NDEBUG
-#endif
-
 #include <stdio.h>
-#include <assert.h>
 #include "cryptlib.h"
 #include "bn_lcl.h"
 
-#if defined(OPENSSL_NO_ASM) || !(defined(__i386) || defined(__i386__)) || defined(__DJGPP__) /* Assembler implementation exists only for x86 */
-/* Here follows specialised variants of bn_add_words() and
-   bn_sub_words().  They have the property performing operations on
-   arrays of different sizes.  The sizes of those arrays is expressed through
-   cl, which is the common length ( basicall, min(len(a),len(b)) ), and dl,
-   which is the delta between the two lengths, calculated as len(a)-len(b).
-   All lengths are the number of BN_ULONGs...  For the operations that require
-   a result array as parameter, it must have the length cl+abs(dl).
-   These functions should probably end up in bn_asm.c as soon as there are
-   assembler counterparts for the systems that use assembler files.  */
-
-BN_ULONG bn_sub_part_words(BN_ULONG *r,
-	const BN_ULONG *a, const BN_ULONG *b,
-	int cl, int dl)
-	{
-	BN_ULONG c, t;
-
-	assert(cl >= 0);
-	c = bn_sub_words(r, a, b, cl);
-
-	if (dl == 0)
-		return c;
-
-	r += cl;
-	a += cl;
-	b += cl;
-
-	if (dl < 0)
-		{
-#ifdef BN_COUNT
-		fprintf(stderr, "  bn_sub_part_words %d + %d (dl < 0, c = %d)\n", cl, dl, c);
-#endif
-		for (;;)
-			{
-			t = b[0];
-			r[0] = (0-t-c)&BN_MASK2;
-			if (t != 0) c=1;
-			if (++dl >= 0) break;
-
-			t = b[1];
-			r[1] = (0-t-c)&BN_MASK2;
-			if (t != 0) c=1;
-			if (++dl >= 0) break;
-
-			t = b[2];
-			r[2] = (0-t-c)&BN_MASK2;
-			if (t != 0) c=1;
-			if (++dl >= 0) break;
-
-			t = b[3];
-			r[3] = (0-t-c)&BN_MASK2;
-			if (t != 0) c=1;
-			if (++dl >= 0) break;
-
-			b += 4;
-			r += 4;
-			}
-		}
-	else
-		{
-		int save_dl = dl;
-#ifdef BN_COUNT
-		fprintf(stderr, "  bn_sub_part_words %d + %d (dl > 0, c = %d)\n", cl, dl, c);
-#endif
-		while(c)
-			{
-			t = a[0];
-			r[0] = (t-c)&BN_MASK2;
-			if (t != 0) c=0;
-			if (--dl <= 0) break;
-
-			t = a[1];
-			r[1] = (t-c)&BN_MASK2;
-			if (t != 0) c=0;
-			if (--dl <= 0) break;
-
-			t = a[2];
-			r[2] = (t-c)&BN_MASK2;
-			if (t != 0) c=0;
-			if (--dl <= 0) break;
-
-			t = a[3];
-			r[3] = (t-c)&BN_MASK2;
-			if (t != 0) c=0;
-			if (--dl <= 0) break;
-
-			save_dl = dl;
-			a += 4;
-			r += 4;
-			}
-		if (dl > 0)
-			{
-#ifdef BN_COUNT
-			fprintf(stderr, "  bn_sub_part_words %d + %d (dl > 0, c == 0)\n", cl, dl);
-#endif
-			if (save_dl > dl)
-				{
-				switch (save_dl - dl)
-					{
-				case 1:
-					r[1] = a[1];
-					if (--dl <= 0) break;
-				case 2:
-					r[2] = a[2];
-					if (--dl <= 0) break;
-				case 3:
-					r[3] = a[3];
-					if (--dl <= 0) break;
-					}
-				a += 4;
-				r += 4;
-				}
-			}
-		if (dl > 0)
-			{
-#ifdef BN_COUNT
-			fprintf(stderr, "  bn_sub_part_words %d + %d (dl > 0, copy)\n", cl, dl);
-#endif
-			for(;;)
-				{
-				r[0] = a[0];
-				if (--dl <= 0) break;
-				r[1] = a[1];
-				if (--dl <= 0) break;
-				r[2] = a[2];
-				if (--dl <= 0) break;
-				r[3] = a[3];
-				if (--dl <= 0) break;
-
-				a += 4;
-				r += 4;
-				}
-			}
-		}
-	return c;
-	}
-#endif
-
-BN_ULONG bn_add_part_words(BN_ULONG *r,
-	const BN_ULONG *a, const BN_ULONG *b,
-	int cl, int dl)
-	{
-	BN_ULONG c, l, t;
-
-	assert(cl >= 0);
-	c = bn_add_words(r, a, b, cl);
-
-	if (dl == 0)
-		return c;
-
-	r += cl;
-	a += cl;
-	b += cl;
-
-	if (dl < 0)
-		{
-		int save_dl = dl;
-#ifdef BN_COUNT
-		fprintf(stderr, "  bn_add_part_words %d + %d (dl < 0, c = %d)\n", cl, dl, c);
-#endif
-		while (c)
-			{
-			l=(c+b[0])&BN_MASK2;
-			c=(l < c);
-			r[0]=l;
-			if (++dl >= 0) break;
-
-			l=(c+b[1])&BN_MASK2;
-			c=(l < c);
-			r[1]=l;
-			if (++dl >= 0) break;
-
-			l=(c+b[2])&BN_MASK2;
-			c=(l < c);
-			r[2]=l;
-			if (++dl >= 0) break;
-
-			l=(c+b[3])&BN_MASK2;
-			c=(l < c);
-			r[3]=l;
-			if (++dl >= 0) break;
-
-			save_dl = dl;
-			b+=4;
-			r+=4;
-			}
-		if (dl < 0)
-			{
-#ifdef BN_COUNT
-			fprintf(stderr, "  bn_add_part_words %d + %d (dl < 0, c == 0)\n", cl, dl);
-#endif
-			if (save_dl < dl)
-				{
-				switch (dl - save_dl)
-					{
-				case 1:
-					r[1] = b[1];
-					if (++dl >= 0) break;
-				case 2:
-					r[2] = b[2];
-					if (++dl >= 0) break;
-				case 3:
-					r[3] = b[3];
-					if (++dl >= 0) break;
-					}
-				b += 4;
-				r += 4;
-				}
-			}
-		if (dl < 0)
-			{
-#ifdef BN_COUNT
-			fprintf(stderr, "  bn_add_part_words %d + %d (dl < 0, copy)\n", cl, dl);
-#endif
-			for(;;)
-				{
-				r[0] = b[0];
-				if (++dl >= 0) break;
-				r[1] = b[1];
-				if (++dl >= 0) break;
-				r[2] = b[2];
-				if (++dl >= 0) break;
-				r[3] = b[3];
-				if (++dl >= 0) break;
-
-				b += 4;
-				r += 4;
-				}
-			}
-		}
-	else
-		{
-		int save_dl = dl;
-#ifdef BN_COUNT
-		fprintf(stderr, "  bn_add_part_words %d + %d (dl > 0)\n", cl, dl);
-#endif
-		while (c)
-			{
-			t=(a[0]+c)&BN_MASK2;
-			c=(t < c);
-			r[0]=t;
-			if (--dl <= 0) break;
-
-			t=(a[1]+c)&BN_MASK2;
-			c=(t < c);
-			r[1]=t;
-			if (--dl <= 0) break;
-
-			t=(a[2]+c)&BN_MASK2;
-			c=(t < c);
-			r[2]=t;
-			if (--dl <= 0) break;
-
-			t=(a[3]+c)&BN_MASK2;
-			c=(t < c);
-			r[3]=t;
-			if (--dl <= 0) break;
-
-			save_dl = dl;
-			a+=4;
-			r+=4;
-			}
-#ifdef BN_COUNT
-		fprintf(stderr, "  bn_add_part_words %d + %d (dl > 0, c == 0)\n", cl, dl);
-#endif
-		if (dl > 0)
-			{
-			if (save_dl > dl)
-				{
-				switch (save_dl - dl)
-					{
-				case 1:
-					r[1] = a[1];
-					if (--dl <= 0) break;
-				case 2:
-					r[2] = a[2];
-					if (--dl <= 0) break;
-				case 3:
-					r[3] = a[3];
-					if (--dl <= 0) break;
-					}
-				a += 4;
-				r += 4;
-				}
-			}
-		if (dl > 0)
-			{
-#ifdef BN_COUNT
-			fprintf(stderr, "  bn_add_part_words %d + %d (dl > 0, copy)\n", cl, dl);
-#endif
-			for(;;)
-				{
-				r[0] = a[0];
-				if (--dl <= 0) break;
-				r[1] = a[1];
-				if (--dl <= 0) break;
-				r[2] = a[2];
-				if (--dl <= 0) break;
-				r[3] = a[3];
-				if (--dl <= 0) break;
-
-				a += 4;
-				r += 4;
-				}
-			}
-		}
-	return c;
-	}
-
 #ifdef BN_RECURSION
 /* Karatsuba recursive multiplication algorithm
  * (cf. Knuth, The Art of Computer Programming, Vol. 2) */
@@ -390,15 +75,14 @@ BN_ULONG bn_add_part_words(BN_ULONG *r,
  * a[1]*b[1]
  */
 void bn_mul_recursive(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n2,
-	int dna, int dnb, BN_ULONG *t)
+	     BN_ULONG *t)
 	{
 	int n=n2/2,c1,c2;
-	int tna=n+dna, tnb=n+dnb;
 	unsigned int neg,zero;
 	BN_ULONG ln,lo,*p;
 
 # ifdef BN_COUNT
-	fprintf(stderr," bn_mul_recursive %d * %d\n",n2,n2);
+	printf(" bn_mul_recursive %d * %d\n",n2,n2);
 # endif
 # ifdef BN_MUL_COMBA
 #  if 0
@@ -408,40 +92,34 @@ void bn_mul_recursive(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n2,
 		return;
 		}
 #  endif
-	/* Only call bn_mul_comba 8 if n2 == 8 and the
-	 * two arrays are complete [steve]
-	 */
-	if (n2 == 8 && dna == 0 && dnb == 0)
+	if (n2 == 8)
 		{
 		bn_mul_comba8(r,a,b);
 		return; 
 		}
 # endif /* BN_MUL_COMBA */
-	/* Else do normal multiply */
 	if (n2 < BN_MUL_RECURSIVE_SIZE_NORMAL)
 		{
-		bn_mul_normal(r,a,n2+dna,b,n2+dnb);
-		if ((dna + dnb) < 0)
-			memset(&r[2*n2 + dna + dnb], 0,
-				sizeof(BN_ULONG) * -(dna + dnb));
+		/* This should not happen */
+		bn_mul_normal(r,a,n2,b,n2);
 		return;
 		}
 	/* r=(a[0]-a[1])*(b[1]-b[0]) */
-	c1=bn_cmp_part_words(a,&(a[n]),tna,n-tna);
-	c2=bn_cmp_part_words(&(b[n]),b,tnb,tnb-n);
+	c1=bn_cmp_words(a,&(a[n]),n);
+	c2=bn_cmp_words(&(b[n]),b,n);
 	zero=neg=0;
 	switch (c1*3+c2)
 		{
 	case -4:
-		bn_sub_part_words(t,      &(a[n]),a,      tna,tna-n); /* - */
-		bn_sub_part_words(&(t[n]),b,      &(b[n]),tnb,n-tnb); /* - */
+		bn_sub_words(t,      &(a[n]),a,      n); /* - */
+		bn_sub_words(&(t[n]),b,      &(b[n]),n); /* - */
 		break;
 	case -3:
 		zero=1;
 		break;
 	case -2:
-		bn_sub_part_words(t,      &(a[n]),a,      tna,tna-n); /* - */
-		bn_sub_part_words(&(t[n]),&(b[n]),b,      tnb,tnb-n); /* + */
+		bn_sub_words(t,      &(a[n]),a,      n); /* - */
+		bn_sub_words(&(t[n]),&(b[n]),b,      n); /* + */
 		neg=1;
 		break;
 	case -1:
@@ -450,22 +128,21 @@ void bn_mul_recursive(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n2,
 		zero=1;
 		break;
 	case 2:
-		bn_sub_part_words(t,      a,      &(a[n]),tna,n-tna); /* + */
-		bn_sub_part_words(&(t[n]),b,      &(b[n]),tnb,n-tnb); /* - */
+		bn_sub_words(t,      a,      &(a[n]),n); /* + */
+		bn_sub_words(&(t[n]),b,      &(b[n]),n); /* - */
 		neg=1;
 		break;
 	case 3:
 		zero=1;
 		break;
 	case 4:
-		bn_sub_part_words(t,      a,      &(a[n]),tna,n-tna);
-		bn_sub_part_words(&(t[n]),&(b[n]),b,      tnb,tnb-n);
+		bn_sub_words(t,      a,      &(a[n]),n);
+		bn_sub_words(&(t[n]),&(b[n]),b,      n);
 		break;
 		}
 
 # ifdef BN_MUL_COMBA
-	if (n == 4 && dna == 0 && dnb == 0) /* XXX: bn_mul_comba4 could take
-					       extra args to do this well */
+	if (n == 4)
 		{
 		if (!zero)
 			bn_mul_comba4(&(t[n2]),t,&(t[n]));
@@ -475,9 +152,7 @@ void bn_mul_recursive(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n2,
 		bn_mul_comba4(r,a,b);
 		bn_mul_comba4(&(r[n2]),&(a[n]),&(b[n]));
 		}
-	else if (n == 8 && dna == 0 && dnb == 0) /* XXX: bn_mul_comba8 could
-						    take extra args to do this
-						    well */
+	else if (n == 8)
 		{
 		if (!zero)
 			bn_mul_comba8(&(t[n2]),t,&(t[n]));
@@ -492,11 +167,11 @@ void bn_mul_recursive(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n2,
 		{
 		p= &(t[n2*2]);
 		if (!zero)
-			bn_mul_recursive(&(t[n2]),t,&(t[n]),n,0,0,p);
+			bn_mul_recursive(&(t[n2]),t,&(t[n]),n,p);
 		else
 			memset(&(t[n2]),0,n2*sizeof(BN_ULONG));
-		bn_mul_recursive(r,a,b,n,0,0,p);
-		bn_mul_recursive(&(r[n2]),&(a[n]),&(b[n]),n,dna,dnb,p);
+		bn_mul_recursive(r,a,b,n,p);
+		bn_mul_recursive(&(r[n2]),&(a[n]),&(b[n]),n,p);
 		}
 
 	/* t[32] holds (a[0]-a[1])*(b[1]-b[0]), c1 is the sign
@@ -545,39 +220,39 @@ void bn_mul_recursive(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n2,
 
 /* n+tn is the word length
  * t needs to be n*4 is size, as does r */
-void bn_mul_part_recursive(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n,
-	     int tna, int tnb, BN_ULONG *t)
+void bn_mul_part_recursive(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int tn,
+	     int n, BN_ULONG *t)
 	{
 	int i,j,n2=n*2;
 	unsigned int c1,c2,neg,zero;
 	BN_ULONG ln,lo,*p;
 
 # ifdef BN_COUNT
-	fprintf(stderr," bn_mul_part_recursive (%d+%d) * (%d+%d)\n",
-		tna, n, tnb, n);
+	printf(" bn_mul_part_recursive %d * %d\n",tn+n,tn+n);
 # endif
 	if (n < 8)
 		{
-		bn_mul_normal(r,a,n+tna,b,n+tnb);
+		i=tn+n;
+		bn_mul_normal(r,a,i,b,i);
 		return;
 		}
 
 	/* r=(a[0]-a[1])*(b[1]-b[0]) */
-	c1=bn_cmp_part_words(a,&(a[n]),tna,n-tna);
-	c2=bn_cmp_part_words(&(b[n]),b,tnb,tnb-n);
+	c1=bn_cmp_words(a,&(a[n]),n);
+	c2=bn_cmp_words(&(b[n]),b,n);
 	zero=neg=0;
 	switch (c1*3+c2)
 		{
 	case -4:
-		bn_sub_part_words(t,      &(a[n]),a,      tna,tna-n); /* - */
-		bn_sub_part_words(&(t[n]),b,      &(b[n]),tnb,n-tnb); /* - */
+		bn_sub_words(t,      &(a[n]),a,      n); /* - */
+		bn_sub_words(&(t[n]),b,      &(b[n]),n); /* - */
 		break;
 	case -3:
 		zero=1;
 		/* break; */
 	case -2:
-		bn_sub_part_words(t,      &(a[n]),a,      tna,tna-n); /* - */
-		bn_sub_part_words(&(t[n]),&(b[n]),b,      tnb,tnb-n); /* + */
+		bn_sub_words(t,      &(a[n]),a,      n); /* - */
+		bn_sub_words(&(t[n]),&(b[n]),b,      n); /* + */
 		neg=1;
 		break;
 	case -1:
@@ -586,16 +261,16 @@ void bn_mul_part_recursive(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n,
 		zero=1;
 		/* break; */
 	case 2:
-		bn_sub_part_words(t,      a,      &(a[n]),tna,n-tna); /* + */
-		bn_sub_part_words(&(t[n]),b,      &(b[n]),tnb,n-tnb); /* - */
+		bn_sub_words(t,      a,      &(a[n]),n); /* + */
+		bn_sub_words(&(t[n]),b,      &(b[n]),n); /* - */
 		neg=1;
 		break;
 	case 3:
 		zero=1;
 		/* break; */
 	case 4:
-		bn_sub_part_words(t,      a,      &(a[n]),tna,n-tna);
-		bn_sub_part_words(&(t[n]),&(b[n]),b,      tnb,tnb-n);
+		bn_sub_words(t,      a,      &(a[n]),n);
+		bn_sub_words(&(t[n]),&(b[n]),b,      n);
 		break;
 		}
 		/* The zero case isn't yet implemented here. The speedup
@@ -614,59 +289,54 @@ void bn_mul_part_recursive(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n,
 		{
 		bn_mul_comba8(&(t[n2]),t,&(t[n]));
 		bn_mul_comba8(r,a,b);
-		bn_mul_normal(&(r[n2]),&(a[n]),tna,&(b[n]),tnb);
-		memset(&(r[n2+tna+tnb]),0,sizeof(BN_ULONG)*(n2-tna-tnb));
+		bn_mul_normal(&(r[n2]),&(a[n]),tn,&(b[n]),tn);
+		memset(&(r[n2+tn*2]),0,sizeof(BN_ULONG)*(n2-tn*2));
 		}
 	else
 		{
 		p= &(t[n2*2]);
-		bn_mul_recursive(&(t[n2]),t,&(t[n]),n,0,0,p);
-		bn_mul_recursive(r,a,b,n,0,0,p);
+		bn_mul_recursive(&(t[n2]),t,&(t[n]),n,p);
+		bn_mul_recursive(r,a,b,n,p);
 		i=n/2;
 		/* If there is only a bottom half to the number,
 		 * just do it */
-		if (tna > tnb)
-			j = tna - i;
-		else
-			j = tnb - i;
+		j=tn-i;
 		if (j == 0)
 			{
-			bn_mul_recursive(&(r[n2]),&(a[n]),&(b[n]),
-				i,tna-i,tnb-i,p);
+			bn_mul_recursive(&(r[n2]),&(a[n]),&(b[n]),i,p);
 			memset(&(r[n2+i*2]),0,sizeof(BN_ULONG)*(n2-i*2));
 			}
 		else if (j > 0) /* eg, n == 16, i == 8 and tn == 11 */
 				{
 				bn_mul_part_recursive(&(r[n2]),&(a[n]),&(b[n]),
-					i,tna-i,tnb-i,p);
-				memset(&(r[n2+tna+tnb]),0,
-					sizeof(BN_ULONG)*(n2-tna-tnb));
+					j,i,p);
+				memset(&(r[n2+tn*2]),0,
+					sizeof(BN_ULONG)*(n2-tn*2));
 				}
 		else /* (j < 0) eg, n == 16, i == 8 and tn == 5 */
 			{
 			memset(&(r[n2]),0,sizeof(BN_ULONG)*n2);
-			if (tna < BN_MUL_RECURSIVE_SIZE_NORMAL
-				&& tnb < BN_MUL_RECURSIVE_SIZE_NORMAL)
+			if (tn < BN_MUL_RECURSIVE_SIZE_NORMAL)
 				{
-				bn_mul_normal(&(r[n2]),&(a[n]),tna,&(b[n]),tnb);
+				bn_mul_normal(&(r[n2]),&(a[n]),tn,&(b[n]),tn);
 				}
 			else
 				{
 				for (;;)
 					{
 					i/=2;
-					if (i < tna && i < tnb)
+					if (i < tn)
 						{
 						bn_mul_part_recursive(&(r[n2]),
 							&(a[n]),&(b[n]),
-							i,tna-i,tnb-i,p);
+							tn-i,i,p);
 						break;
 						}
-					else if (i <= tna && i <= tnb)
+					else if (i == tn)
 						{
 						bn_mul_recursive(&(r[n2]),
 							&(a[n]),&(b[n]),
-							i,tna-i,tnb-i,p);
+							i,p);
 						break;
 						}
 					}
@@ -727,10 +397,10 @@ void bn_mul_low_recursive(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n2,
 	int n=n2/2;
 
 # ifdef BN_COUNT
-	fprintf(stderr," bn_mul_low_recursive %d * %d\n",n2,n2);
+	printf(" bn_mul_low_recursive %d * %d\n",n2,n2);
 # endif
 
-	bn_mul_recursive(r,a,b,n,0,0,&(t[0]));
+	bn_mul_recursive(r,a,b,n,&(t[0]));
 	if (n >= BN_MUL_LOW_RECURSIVE_SIZE_NORMAL)
 		{
 		bn_mul_low_recursive(&(t[0]),&(a[0]),&(b[n]),n,&(t[n2]));
@@ -761,7 +431,7 @@ void bn_mul_high(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, BN_ULONG *l, int n2,
 	BN_ULONG ll,lc,*lp,*mp;
 
 # ifdef BN_COUNT
-	fprintf(stderr," bn_mul_high %d * %d\n",n2,n2);
+	printf(" bn_mul_high %d * %d\n",n2,n2);
 # endif
 	n=n2/2;
 
@@ -814,8 +484,8 @@ void bn_mul_high(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, BN_ULONG *l, int n2,
 	else
 # endif
 		{
-		bn_mul_recursive(&(t[0]),&(r[0]),&(r[n]),n,0,0,&(t[n2]));
-		bn_mul_recursive(r,&(a[n]),&(b[n]),n,0,0,&(t[n2]));
+		bn_mul_recursive(&(t[0]),&(r[0]),&(r[n]),n,&(t[n2]));
+		bn_mul_recursive(r,&(a[n]),&(b[n]),n,&(t[n2]));
 		}
 
 	/* s0 == low(al*bl)
@@ -940,19 +610,19 @@ void bn_mul_high(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, BN_ULONG *l, int n2,
 
 int BN_mul(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, BN_CTX *ctx)
 	{
-	int ret=0;
 	int top,al,bl;
 	BIGNUM *rr;
+	int ret = 0;
 #if defined(BN_MUL_COMBA) || defined(BN_RECURSION)
 	int i;
 #endif
 #ifdef BN_RECURSION
-	BIGNUM *t=NULL;
-	int j=0,k;
+	BIGNUM *t;
+	int j,k;
 #endif
 
 #ifdef BN_COUNT
-	fprintf(stderr,"BN_mul %d * %d\n",a->top,b->top);
+	printf("BN_mul %d * %d\n",a->top,b->top);
 #endif
 
 	bn_check_top(a);
@@ -1005,55 +675,21 @@ int BN_mul(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, BN_CTX *ctx)
 #ifdef BN_RECURSION
 	if ((al >= BN_MULL_SIZE_NORMAL) && (bl >= BN_MULL_SIZE_NORMAL))
 		{
-		if (i >= -1 && i <= 1)
+		if (i == 1 && !BN_get_flags(b,BN_FLG_STATIC_DATA) && bl<b->dmax)
 			{
-			int sav_j =0;
-			/* Find out the power of two lower or equal
-			   to the longest of the two numbers */
-			if (i >= 0)
-				{
-				j = BN_num_bits_word((BN_ULONG)al);
-				}
-			if (i == -1)
-				{
-				j = BN_num_bits_word((BN_ULONG)bl);
-				}
-			sav_j = j;
-			j = 1<<(j-1);
-			assert(j <= al || j <= bl);
-			k = j+j;
-			t = BN_CTX_get(ctx);
-			if (al > j || bl > j)
-				{
-				bn_wexpand(t,k*4);
-				bn_wexpand(rr,k*4);
-				bn_mul_part_recursive(rr->d,a->d,b->d,
-					j,al-j,bl-j,t->d);
-				}
-			else	/* al <= j || bl <= j */
-				{
-				bn_wexpand(t,k*2);
-				bn_wexpand(rr,k*2);
-				bn_mul_recursive(rr->d,a->d,b->d,
-					j,al-j,bl-j,t->d);
-				}
-			rr->top=top;
-			goto end;
-			}
-#if 0
-		if (i == 1 && !BN_get_flags(b,BN_FLG_STATIC_DATA))
-			{
-			BIGNUM *tmp_bn = (BIGNUM *)b;
-			if (bn_wexpand(tmp_bn,al) == NULL) goto err;
-			tmp_bn->d[bl]=0;
+#if 0	/* tribute to const-ification, bl<b->dmax above covers for this */
+			if (bn_wexpand(b,al) == NULL) goto err;
+#endif
+			b->d[bl]=0;
 			bl++;
 			i--;
 			}
-		else if (i == -1 && !BN_get_flags(a,BN_FLG_STATIC_DATA))
+		else if (i == -1 && !BN_get_flags(a,BN_FLG_STATIC_DATA) && al<a->dmax)
 			{
-			BIGNUM *tmp_bn = (BIGNUM *)a;
-			if (bn_wexpand(tmp_bn,bl) == NULL) goto err;
-			tmp_bn->d[al]=0;
+#if 0	/* tribute to const-ification, al<a->dmax above covers for this */
+			if (bn_wexpand(a,bl) == NULL) goto err;
+#endif
+			a->d[al]=0;
 			al++;
 			i++;
 			}
@@ -1070,17 +706,26 @@ int BN_mul(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, BN_CTX *ctx)
 				if (bn_wexpand(t,k*2) == NULL) goto err;
 				if (bn_wexpand(rr,k*2) == NULL) goto err;
 				bn_mul_recursive(rr->d,a->d,b->d,al,t->d);
+				rr->top=top;
+				goto end;
 				}
+#if 0	/* tribute to const-ification, rsa/dsa performance is not affected */
 			else
 				{
-				if (bn_wexpand(t,k*4) == NULL) goto err;
-				if (bn_wexpand(rr,k*4) == NULL) goto err;
+				if (bn_wexpand(a,k) == NULL ) goto err;
+				if (bn_wexpand(b,k) == NULL ) goto err;
+				if (bn_wexpand(t,k*4) == NULL ) goto err;
+				if (bn_wexpand(rr,k*4) == NULL ) goto err;
+				for (i=a->top; i<k; i++)
+					a->d[i]=0;
+				for (i=b->top; i<k; i++)
+					b->d[i]=0;
 				bn_mul_part_recursive(rr->d,a->d,b->d,al-j,j,t->d);
 				}
 			rr->top=top;
 			goto end;
-			}
 #endif
+			}
 		}
 #endif /* BN_RECURSION */
 	if (bn_wexpand(rr,top) == NULL) goto err;
@@ -1103,7 +748,7 @@ void bn_mul_normal(BN_ULONG *r, BN_ULONG *a, int na, BN_ULONG *b, int nb)
 	BN_ULONG *rr;
 
 #ifdef BN_COUNT
-	fprintf(stderr," bn_mul_normal %d * %d\n",na,nb);
+	printf(" bn_mul_normal %d * %d\n",na,nb);
 #endif
 
 	if (na < nb)
@@ -1116,13 +761,7 @@ void bn_mul_normal(BN_ULONG *r, BN_ULONG *a, int na, BN_ULONG *b, int nb)
 
 		}
 	rr= &(r[na]);
-	if (nb <= 0)
-		{
-		(void)bn_mul_words(r,a,na,0);
-		return;
-		}
-	else
-		rr[0]=bn_mul_words(r,a,na,b[0]);
+	rr[0]=bn_mul_words(r,a,na,b[0]);
 
 	for (;;)
 		{
@@ -1143,7 +782,7 @@ void bn_mul_normal(BN_ULONG *r, BN_ULONG *a, int na, BN_ULONG *b, int nb)
 void bn_mul_low_normal(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n)
 	{
 #ifdef BN_COUNT
-	fprintf(stderr," bn_mul_low_normal %d * %d\n",n,n);
+	printf(" bn_mul_low_normal %d * %d\n",n,n);
 #endif
 	bn_mul_words(r,a,n,b[0]);
 

crypto/cryptlib.c

@@ -104,7 +104,8 @@ static const char* lock_names[CRYPTO_NUM_LOCKS] =
 	"dynlock",
 	"engine",
 	"ui",
-#if CRYPTO_NUM_LOCKS != 32
+	"hwcrhk",		/* This is a HACK which will disappear in 0.9.8 */
+#if CRYPTO_NUM_LOCKS != 33
 # error "Inconsistency between crypto.h and cryptlib.c"
 #endif
 	};
@@ -207,10 +208,18 @@ int CRYPTO_get_new_dynlockid(void)
 	i=sk_CRYPTO_dynlock_find(dyn_locks,NULL);
 	/* If there was none, push, thereby creating a new one */
 	if (i == -1)
-		i=sk_CRYPTO_dynlock_push(dyn_locks,pointer);
+		/* Since sk_push() returns the number of items on the
+		   stack, not the location of the pushed item, we need
+		   to transform the returned number into a position,
+		   by decreasing it.  */
+		i=sk_CRYPTO_dynlock_push(dyn_locks,pointer) - 1;
+	else
+		/* If we found a place with a NULL pointer, put our pointer
+		   in it.  */
+		sk_CRYPTO_dynlock_set(dyn_locks,i,pointer);
 	CRYPTO_w_unlock(CRYPTO_LOCK_DYNLOCK);
 
-	if (!i)
+	if (i == -1)
 		{
 		dynlock_destroy_callback(pointer->data,__FILE__,__LINE__);
 		OPENSSL_free(pointer);
@@ -402,15 +411,17 @@ void CRYPTO_lock(int mode, int type, const char *file, int line)
 #endif
 	if (type < 0)
 		{
-		struct CRYPTO_dynlock_value *pointer
-			= CRYPTO_get_dynlock_value(type);
-
-		if (pointer && dynlock_lock_callback)
+		if (dynlock_lock_callback != NULL)
 			{
-			dynlock_lock_callback(mode, pointer, file, line);
-			}
+			struct CRYPTO_dynlock_value *pointer
+				= CRYPTO_get_dynlock_value(type);
 
-		CRYPTO_destroy_dynlockid(type);
+			OPENSSL_assert(pointer != NULL);
+
+			dynlock_lock_callback(mode, pointer, file, line);
+
+			CRYPTO_destroy_dynlockid(type);
+			}
 		}
 	else
 		if (locking_callback != NULL)

crypto/crypto.h

@@ -127,7 +127,8 @@ extern "C" {
 #define CRYPTO_LOCK_DYNLOCK		29
 #define CRYPTO_LOCK_ENGINE		30
 #define CRYPTO_LOCK_UI			31
-#define CRYPTO_NUM_LOCKS		32
+#define CRYPTO_LOCK_HWCRHK		32 /* This is a HACK which will disappear in 0.9.8 */
+#define CRYPTO_NUM_LOCKS		33
 
 #define CRYPTO_LOCK		1
 #define CRYPTO_UNLOCK		2

crypto/des/Makefile.ssl

@@ -67,19 +67,27 @@ des: des.o cbc3_enc.o lib
 
 # elf
 asm/dx86-elf.o: asm/dx86unix.cpp
-	$(CPP) -DELF -x c asm/dx86unix.cpp | as -o asm/dx86-elf.o
+	$(CPP) -DELF	\
+		`(echo $(CFLAGS) | egrep -ie '-[fK]PIC') > /dev/null 2>&1 && echo -DPIC`\
+		-x c asm/dx86unix.cpp | as -o asm/dx86-elf.o
 
 asm/yx86-elf.o: asm/yx86unix.cpp
-	$(CPP) -DELF -x c asm/yx86unix.cpp | as -o asm/yx86-elf.o
+	$(CPP) -DELF	\
+		`(echo $(CFLAGS) | egrep -ie '-[fK]PIC') > /dev/null 2>&1 && echo -DPIC`\
+		-x c asm/yx86unix.cpp | as -o asm/yx86-elf.o
 
 # solaris
 asm/dx86-sol.o: asm/dx86unix.cpp
-	$(CC) -E -DSOL asm/dx86unix.cpp | sed 's/^#.*//' > asm/dx86-sol.s
+	$(CC) -E -DSOL	\
+		`(echo $(CFLAGS) | egrep -ie '-[fK]PIC') > /dev/null 2>&1 && echo -DPIC`\
+		asm/dx86unix.cpp | sed 's/^#.*//' > asm/dx86-sol.s
 	as -o asm/dx86-sol.o asm/dx86-sol.s
 	rm -f asm/dx86-sol.s
 
 asm/yx86-sol.o: asm/yx86unix.cpp
-	$(CC) -E -DSOL asm/yx86unix.cpp | sed 's/^#.*//' > asm/yx86-sol.s
+	$(CC) -E -DSOL	\
+		`(echo $(CFLAGS) | egrep -ie '-[fK]PIC') > /dev/null 2>&1 && echo -DPIC`\
+		asm/yx86unix.cpp | sed 's/^#.*//' > asm/yx86-sol.s
 	as -o asm/yx86-sol.o asm/yx86-sol.s
 	rm -f asm/yx86-sol.s
 

crypto/des/asm/crypt586.pl

@@ -26,11 +26,17 @@ sub fcrypt_body
 
 	&comment("");
 	&comment("Load the 2 words");
-	$ks="ebp";
+	$trans="ebp";
 
 	&xor(	$L,	$L);
 	&xor(	$R,	$R);
-	&mov($ks,&wparam(1));
+
+	# PIC-ification:-)
+	if ($cpp)	{ &picmeup("edx","DES_SPtrans");   }
+	else		{ &lea("edx",&DWP("DES_SPtrans")); }
+	&push("edx");	# becomes &swtmp(1)
+	#
+	&mov($trans,&wparam(1)); # reloaded with DES_SPtrans in D_ENCRYPT
 
 	&push(&DWC(25)); # add a variable
 
@@ -39,11 +45,11 @@ sub fcrypt_body
 		{
 		&comment("");
 		&comment("Round $i");
-		&D_ENCRYPT($i,$L,$R,$i*2,$ks,"DES_SPtrans","eax","ebx","ecx","edx");
+		&D_ENCRYPT($i,$L,$R,$i*2,$trans,"eax","ebx","ecx","edx");
 
 		&comment("");
 		&comment("Round ".sprintf("%d",$i+1));
-		&D_ENCRYPT($i+1,$R,$L,($i+1)*2,$ks,"DES_SPtrans","eax","ebx","ecx","edx");
+		&D_ENCRYPT($i+1,$R,$L,($i+1)*2,$trans,"eax","ebx","ecx","edx");
 		}
 	 &mov("ebx",	&swtmp(0));
 	&mov("eax",	$L);
@@ -61,14 +67,14 @@ sub fcrypt_body
 	&mov(&DWP(0,"edx","",0),"eax");
 	&mov(&DWP(4,"edx","",0),$L);
 
-	&pop("ecx");	# remove variable
+	&add("esp",8);	# remove variables
 
 	&function_end($name);
 	}
 
 sub D_ENCRYPT
 	{
-	local($r,$L,$R,$S,$ks,$desSP,$u,$tmp1,$tmp2,$t)=@_;
+	local($r,$L,$R,$S,$trans,$u,$tmp1,$tmp2,$t)=@_;
 
 	&mov(	$u,		&wparam(2));			# 2
 	&mov(	$t,		$R);
@@ -85,9 +91,9 @@ sub D_ENCRYPT
 	&shl(	$tmp2,		16); 				# 1
 	&xor(	$u,		$tmp1);				# 2
 	&xor(	$t,		$tmp2);				# 2
-	&mov(	$tmp1,		&DWP(&n2a($S*4),$ks,"",0));	# 2
+	&mov(	$tmp1,		&DWP(&n2a($S*4),$trans,"",0));	# 2
 	&xor(	$u,		$tmp1);
-	&mov(	$tmp2,		&DWP(&n2a(($S+1)*4),$ks,"",0));	# 2
+	&mov(	$tmp2,		&DWP(&n2a(($S+1)*4),$trans,"",0));	# 2
 	&xor(	$u,		$R);
 	&xor(	$t,		$R);
 	&xor(	$t,		$tmp2);
@@ -99,31 +105,28 @@ sub D_ENCRYPT
 	&movb(	&LB($tmp1),	&LB($u)	);
 	&movb(	&LB($tmp2),	&HB($u)	);
 	&rotr(	$t,		4		);
-	&mov(	$ks,		&DWP("      $desSP",$tmp1,"",0));
+	&mov(	$trans,		&swtmp(1));
+	&xor(	$L,		&DWP("     ",$trans,$tmp1,0));
 	&movb(	&LB($tmp1),	&LB($t)	);
-	&xor(	$L,		$ks);
-	&mov(	$ks,		&DWP("0x200+$desSP",$tmp2,"",0));
-	&xor(	$L,		$ks);
+	&xor(	$L,		&DWP("0x200",$trans,$tmp2,0));
 	&movb(	&LB($tmp2),	&HB($t)	);
 	&shr(	$u,		16);
-	&mov(	$ks,		&DWP("0x100+$desSP",$tmp1,"",0));
-	&xor(	$L,		$ks); 
+	&xor(	$L,		&DWP("0x100",$trans,$tmp1,0));
 	&movb(	&LB($tmp1),	&HB($u)	);
 	&shr(	$t,		16);
-	&mov(	$ks,		&DWP("0x300+$desSP",$tmp2,"",0));
-	&xor(	$L,		$ks);
-	&mov(	$ks,		&wparam(1));
+	&xor(	$L,		&DWP("0x300",$trans,$tmp2,0));
 	&movb(	&LB($tmp2),	&HB($t)	);
 	&and(	$u,		"0xff"	);
 	&and(	$t,		"0xff"	);
-	&mov(	$tmp1,		&DWP("0x600+$desSP",$tmp1,"",0));
+	&mov(	$tmp1,		&DWP("0x600",$trans,$tmp1,0));
 	&xor(	$L,		$tmp1);
-	&mov(	$tmp1,		&DWP("0x700+$desSP",$tmp2,"",0));
+	&mov(	$tmp1,		&DWP("0x700",$trans,$tmp2,0));
 	&xor(	$L,		$tmp1);
-	&mov(	$tmp1,		&DWP("0x400+$desSP",$u,"",0));
+	&mov(	$tmp1,		&DWP("0x400",$trans,$u,0));
 	&xor(	$L,		$tmp1);
-	&mov(	$tmp1,		&DWP("0x500+$desSP",$t,"",0));
+	&mov(	$tmp1,		&DWP("0x500",$trans,$t,0));
 	&xor(	$L,		$tmp1);
+	&mov(	$trans,		&wparam(1));
 	}
 
 sub n2a

crypto/des/asm/des-586.pl

@@ -40,7 +40,7 @@ sub DES_encrypt
 
 	&comment("");
 	&comment("Load the 2 words");
-	$ks="ebp";
+	$trans="ebp";
 
 	if ($do_ip)
 		{
@@ -72,7 +72,11 @@ sub DES_encrypt
 		&rotl($L,3);
 		}
 
-	&mov(	$ks,		&wparam(1)	);
+	# PIC-ification:-)
+	if ($cpp)	{ &picmeup($trans,"DES_SPtrans");   }
+	else		{ &lea($trans,&DWP("DES_SPtrans")); }
+
+	&mov(	"ecx",	&wparam(1)	);
 	&cmp("ebx","0");
 	&je(&label("start_decrypt"));
 
@@ -80,11 +84,11 @@ sub DES_encrypt
 		{
 		&comment("");
 		&comment("Round $i");
-		&D_ENCRYPT($i,$L,$R,$i*2,$ks,"DES_SPtrans","eax","ebx","ecx","edx");
+		&D_ENCRYPT($i,$L,$R,$i*2,$trans,"eax","ebx","ecx","edx");
 
 		&comment("");
 		&comment("Round ".sprintf("%d",$i+1));
-		&D_ENCRYPT($i+1,$R,$L,($i+1)*2,$ks,"DES_SPtrans","eax","ebx","ecx","edx");
+		&D_ENCRYPT($i+1,$R,$L,($i+1)*2,$trans,"eax","ebx","ecx","edx");
 		}
 	&jmp(&label("end"));
 
@@ -94,10 +98,10 @@ sub DES_encrypt
 		{
 		&comment("");
 		&comment("Round $i");
-		&D_ENCRYPT(15-$i,$L,$R,$i*2,$ks,"DES_SPtrans","eax","ebx","ecx","edx");
+		&D_ENCRYPT(15-$i,$L,$R,$i*2,$trans,"eax","ebx","ecx","edx");
 		&comment("");
 		&comment("Round ".sprintf("%d",$i-1));
-		&D_ENCRYPT(15-$i+1,$R,$L,($i-1)*2,$ks,"DES_SPtrans","eax","ebx","ecx","edx");
+		&D_ENCRYPT(15-$i+1,$R,$L,($i-1)*2,$trans,"eax","ebx","ecx","edx");
 		}
 
 	&set_label("end");
@@ -134,43 +138,36 @@ sub DES_encrypt
 
 sub D_ENCRYPT
 	{
-	local($r,$L,$R,$S,$ks,$desSP,$u,$tmp1,$tmp2,$t)=@_;
+	local($r,$L,$R,$S,$trans,$u,$tmp1,$tmp2,$t)=@_;
 
-	 &mov(	$u,		&DWP(&n2a($S*4),$ks,"",0));
+	 &mov(	$u,		&DWP(&n2a($S*4),$tmp2,"",0));
 	&xor(	$tmp1,		$tmp1);
-	 &mov(	$t,		&DWP(&n2a(($S+1)*4),$ks,"",0));
+	 &mov(	$t,		&DWP(&n2a(($S+1)*4),$tmp2,"",0));
 	&xor(	$u,		$R);
+	&xor(	$tmp2,		$tmp2);
 	 &xor(	$t,		$R);
 	&and(	$u,		"0xfcfcfcfc"	);
 	 &and(	$t,		"0xcfcfcfcf"	);
 	&movb(	&LB($tmp1),	&LB($u)	);
 	 &movb(	&LB($tmp2),	&HB($u)	);
 	&rotr(	$t,		4		);
-	&mov(	$ks,		&DWP("      $desSP",$tmp1,"",0));
+	&xor(	$L,		&DWP("     ",$trans,$tmp1,0));
 	 &movb(	&LB($tmp1),	&LB($t)	);
-	&xor(	$L,		$ks);
-	 &mov(	$ks,		&DWP("0x200+$desSP",$tmp2,"",0));
-	&xor(	$L,		$ks); ######
+	 &xor(	$L,		&DWP("0x200",$trans,$tmp2,0));
 	 &movb(	&LB($tmp2),	&HB($t)	);
 	&shr(	$u,		16);
-	 &mov(	$ks,		&DWP("0x100+$desSP",$tmp1,"",0));
-	&xor(	$L,		$ks); ######
+	 &xor(	$L,		&DWP("0x100",$trans,$tmp1,0));
 	 &movb(	&LB($tmp1),	&HB($u)	);
 	&shr(	$t,		16);
-	 &mov(	$ks,		&DWP("0x300+$desSP",$tmp2,"",0));
-	&xor(	$L,		$ks);
-	 &mov(	$ks,		&wparam(1)	);
+	 &xor(	$L,		&DWP("0x300",$trans,$tmp2,0));
 	&movb(	&LB($tmp2),	&HB($t)	);
 	 &and(	$u,		"0xff"	);
 	&and(	$t,		"0xff"	);
-	 &mov(	$tmp1,		&DWP("0x600+$desSP",$tmp1,"",0));
-	&xor(	$L,		$tmp1);
-	 &mov(	$tmp1,		&DWP("0x700+$desSP",$tmp2,"",0));
-	&xor(	$L,		$tmp1);
-	 &mov(	$tmp1,		&DWP("0x400+$desSP",$u,"",0));
-	&xor(	$L,		$tmp1);
-	 &mov(	$tmp1,		&DWP("0x500+$desSP",$t,"",0));
-	&xor(	$L,		$tmp1);
+	 &xor(	$L,		&DWP("0x600",$trans,$tmp1,0));
+	 &xor(	$L,		&DWP("0x700",$trans,$tmp2,0));
+	&mov(	$tmp2,		&wparam(1)	);
+	 &xor(	$L,		&DWP("0x400",$trans,$u,0));
+	 &xor(	$L,		&DWP("0x500",$trans,$t,0));
 	}
 
 sub n2a

crypto/des/des.h

@@ -56,8 +56,8 @@
  * [including the GNU Public Licence.]
  */
 
-#ifndef HEADER_DES_H
-#define HEADER_DES_H
+#ifndef HEADER_NEW_DES_H
+#define HEADER_NEW_DES_H
 
 #ifdef OPENSSL_NO_DES
 #error DES is disabled.

crypto/des/des_locl.h

@@ -162,7 +162,18 @@
 
 #if defined(OPENSSL_SYS_WIN32) && defined(_MSC_VER)
 #define	ROTATE(a,n)	(_lrotr(a,n))
-#else
+#elif defined(__GNUC__) && __GNUC__>=2 && !defined(__STRICT_ANSI__) && !defined(NO_ASM) && !defined(NO_INLINE_ASM)
+# if defined(__i386) || defined(__i386__) || defined(__x86_64) || defined(__x86_64__)
+#  define ROTATE(a,n)	({ register unsigned int ret;	\
+				asm ("rorl %1,%0"	\
+					: "=r"(ret)	\
+					: "I"(n),"0"(a)	\
+					: "cc");	\
+			   ret;				\
+			})
+# endif
+#endif
+#ifndef ROTATE
 #define	ROTATE(a,n)	(((a)>>(n))+((a)<<(32-(n))))
 #endif
 

crypto/des/des_old.h

@@ -88,14 +88,14 @@
  *
  */
 
-#ifndef HEADER_DES_OLD_H
-#define HEADER_DES_OLD_H
+#ifndef HEADER_DES_H
+#define HEADER_DES_H
 
 #ifdef OPENSSL_NO_DES
 #error DES is disabled.
 #endif
 
-#ifndef HEADER_DES_H
+#ifndef HEADER_NEW_DES_H
 #error You must include des.h, not des_old.h directly.
 #endif
 

crypto/dsa/dsatest.c

@@ -222,6 +222,7 @@ end:
 static int cb_exit(int ec)
 	{
 	EXIT(ec);
+	return(0);		/* To keep some compilers quiet */
 	}
 
 static void MS_CALLBACK dsa_cb(int p, int n, void *arg)

crypto/engine/hw_ncipher.c

@@ -91,11 +91,19 @@ static int hwcrhk_init(ENGINE *e);
 static int hwcrhk_finish(ENGINE *e);
 static int hwcrhk_ctrl(ENGINE *e, int cmd, long i, void *p, void (*f)()); 
 
-/* Functions to handle mutexes */
+/* Functions to handle mutexes if have dynamic locks */
 static int hwcrhk_mutex_init(HWCryptoHook_Mutex*, HWCryptoHook_CallerContext*);
 static int hwcrhk_mutex_lock(HWCryptoHook_Mutex*);
 static void hwcrhk_mutex_unlock(HWCryptoHook_Mutex*);
 static void hwcrhk_mutex_destroy(HWCryptoHook_Mutex*);
+#if 1 /* This is a HACK which will disappear in 0.9.8 */
+/* Functions to handle mutexes if only have static locks */
+static int hwcrhk_static_mutex_init(HWCryptoHook_Mutex *m,
+                                    HWCryptoHook_CallerContext *c);
+static int hwcrhk_static_mutex_lock(HWCryptoHook_Mutex *m);
+static void hwcrhk_static_mutex_unlock(HWCryptoHook_Mutex *m);
+static void hwcrhk_static_mutex_destroy(HWCryptoHook_Mutex *m);
+#endif
 
 /* BIGNUM stuff */
 static int hwcrhk_mod_exp(BIGNUM *r, const BIGNUM *a, const BIGNUM *p,
@@ -573,9 +581,17 @@ static int hwcrhk_init(ENGINE *e)
 			}
 		else if (CRYPTO_get_locking_callback() != NULL)
 			{
-			HWCRHKerr(HWCRHK_F_HWCRHK_INIT,HWCRHK_R_LOCKING_MISSING);
+			HWCRHKerr(HWCRHK_F_HWCRHK_INIT,HWCRHK_R_DYNAMIC_LOCKING_MISSING);
 			ERR_add_error_data(1,"You HAVE to add dynamic locking callbacks via CRYPTO_set_dynlock_{create,lock,destroy}_callback()");
+#if 1 /* This is a HACK which will disappear in 0.9.8 */
+			hwcrhk_globals.maxmutexes    = 1; /* Only have one lock */
+			hwcrhk_globals.mutex_init    = hwcrhk_static_mutex_init;
+			hwcrhk_globals.mutex_acquire = hwcrhk_static_mutex_lock;
+			hwcrhk_globals.mutex_release = hwcrhk_static_mutex_unlock;
+			hwcrhk_globals.mutex_destroy = hwcrhk_static_mutex_destroy;
+#else
 			goto err;
+#endif
 			}
 		}
 
@@ -1181,6 +1197,26 @@ static void hwcrhk_mutex_destroy(HWCryptoHook_Mutex *mt)
 	CRYPTO_destroy_dynlockid(mt->lockid);
 	}
 
+/* Mutex upcalls to use if the application does not support dynamic locks */
+
+static int hwcrhk_static_mutex_init(HWCryptoHook_Mutex *m,
+	HWCryptoHook_CallerContext *c)
+	{
+	return 0;
+	}
+static int hwcrhk_static_mutex_lock(HWCryptoHook_Mutex *m)
+	{
+	CRYPTO_w_lock(CRYPTO_LOCK_HWCRHK);
+	return 0;
+	}
+static void hwcrhk_static_mutex_unlock(HWCryptoHook_Mutex *m)
+	{
+	CRYPTO_w_unlock(CRYPTO_LOCK_HWCRHK);
+	}
+static void hwcrhk_static_mutex_destroy(HWCryptoHook_Mutex *m)
+	{
+	}
+
 static int hwcrhk_get_pass(const char *prompt_info,
 	int *len_io, char *buf,
 	HWCryptoHook_PassphraseContext *ppctx,

crypto/engine/hw_ncipher_err.c

@@ -86,7 +86,7 @@ static ERR_STRING_DATA HWCRHK_str_reasons[]=
 {HWCRHK_R_CHIL_ERROR                     ,"chil error"},
 {HWCRHK_R_CTRL_COMMAND_NOT_IMPLEMENTED   ,"ctrl command not implemented"},
 {HWCRHK_R_DSO_FAILURE                    ,"dso failure"},
-{HWCRHK_R_LOCKING_MISSING                ,"locking missing"},
+{HWCRHK_R_DYNAMIC_LOCKING_MISSING        ,"dynamic locking missing"},
 {HWCRHK_R_MISSING_KEY_COMPONENTS         ,"missing key components"},
 {HWCRHK_R_NOT_INITIALISED                ,"not initialised"},
 {HWCRHK_R_NOT_LOADED                     ,"not loaded"},

crypto/engine/hw_ncipher_err.h

@@ -84,7 +84,7 @@ static void ERR_HWCRHK_error(int function, int reason, char *file, int line);
 #define HWCRHK_R_CHIL_ERROR				 102
 #define HWCRHK_R_CTRL_COMMAND_NOT_IMPLEMENTED		 103
 #define HWCRHK_R_DSO_FAILURE				 104
-#define HWCRHK_R_LOCKING_MISSING			 114
+#define HWCRHK_R_DYNAMIC_LOCKING_MISSING		 114
 #define HWCRHK_R_MISSING_KEY_COMPONENTS			 105
 #define HWCRHK_R_NOT_INITIALISED			 106
 #define HWCRHK_R_NOT_LOADED				 107

crypto/engine/tb_cipher.c

@@ -81,7 +81,7 @@ int ENGINE_register_ciphers(ENGINE *e)
 		int num_nids = e->ciphers(e, NULL, &nids, 0);
 		if(num_nids > 0)
 			return engine_table_register(&cipher_table,
-					&engine_unregister_all_ciphers, e, nids,
+					engine_unregister_all_ciphers, e, nids,
 					num_nids, 0);
 		}
 	return 1;
@@ -103,7 +103,7 @@ int ENGINE_set_default_ciphers(ENGINE *e)
 		int num_nids = e->ciphers(e, NULL, &nids, 0);
 		if(num_nids > 0)
 			return engine_table_register(&cipher_table,
-					&engine_unregister_all_ciphers, e, nids,
+					engine_unregister_all_ciphers, e, nids,
 					num_nids, 1);
 		}
 	return 1;

crypto/engine/tb_dh.c

@@ -78,7 +78,7 @@ int ENGINE_register_DH(ENGINE *e)
 	{
 	if(e->dh_meth)
 		return engine_table_register(&dh_table,
-				&engine_unregister_all_DH, e, &dummy_nid, 1, 0);
+				engine_unregister_all_DH, e, &dummy_nid, 1, 0);
 	return 1;
 	}
 
@@ -94,7 +94,7 @@ int ENGINE_set_default_DH(ENGINE *e)
 	{
 	if(e->dh_meth)
 		return engine_table_register(&dh_table,
-				&engine_unregister_all_DH, e, &dummy_nid, 1, 1);
+				engine_unregister_all_DH, e, &dummy_nid, 1, 1);
 	return 1;
 	}
 

crypto/engine/tb_digest.c

@@ -81,7 +81,7 @@ int ENGINE_register_digests(ENGINE *e)
 		int num_nids = e->digests(e, NULL, &nids, 0);
 		if(num_nids > 0)
 			return engine_table_register(&digest_table,
-					&engine_unregister_all_digests, e, nids,
+					engine_unregister_all_digests, e, nids,
 					num_nids, 0);
 		}
 	return 1;
@@ -103,7 +103,7 @@ int ENGINE_set_default_digests(ENGINE *e)
 		int num_nids = e->digests(e, NULL, &nids, 0);
 		if(num_nids > 0)
 			return engine_table_register(&digest_table,
-					&engine_unregister_all_digests, e, nids,
+					engine_unregister_all_digests, e, nids,
 					num_nids, 1);
 		}
 	return 1;

crypto/engine/tb_dsa.c

@@ -78,7 +78,7 @@ int ENGINE_register_DSA(ENGINE *e)
 	{
 	if(e->dsa_meth)
 		return engine_table_register(&dsa_table,
-				&engine_unregister_all_DSA, e, &dummy_nid, 1, 0);
+				engine_unregister_all_DSA, e, &dummy_nid, 1, 0);
 	return 1;
 	}
 
@@ -94,7 +94,7 @@ int ENGINE_set_default_DSA(ENGINE *e)
 	{
 	if(e->dsa_meth)
 		return engine_table_register(&dsa_table,
-				&engine_unregister_all_DSA, e, &dummy_nid, 1, 0);
+				engine_unregister_all_DSA, e, &dummy_nid, 1, 0);
 	return 1;
 	}
 

crypto/engine/tb_rand.c

@@ -78,7 +78,7 @@ int ENGINE_register_RAND(ENGINE *e)
 	{
 	if(e->rand_meth)
 		return engine_table_register(&rand_table,
-				&engine_unregister_all_RAND, e, &dummy_nid, 1, 0);
+				engine_unregister_all_RAND, e, &dummy_nid, 1, 0);
 	return 1;
 	}
 
@@ -94,7 +94,7 @@ int ENGINE_set_default_RAND(ENGINE *e)
 	{
 	if(e->rand_meth)
 		return engine_table_register(&rand_table,
-				&engine_unregister_all_RAND, e, &dummy_nid, 1, 1);
+				engine_unregister_all_RAND, e, &dummy_nid, 1, 1);
 	return 1;
 	}
 

crypto/engine/tb_rsa.c

@@ -78,7 +78,7 @@ int ENGINE_register_RSA(ENGINE *e)
 	{
 	if(e->rsa_meth)
 		return engine_table_register(&rsa_table,
-				&engine_unregister_all_RSA, e, &dummy_nid, 1, 0);
+				engine_unregister_all_RSA, e, &dummy_nid, 1, 0);
 	return 1;
 	}
 
@@ -94,7 +94,7 @@ int ENGINE_set_default_RSA(ENGINE *e)
 	{
 	if(e->rsa_meth)
 		return engine_table_register(&rsa_table,
-				&engine_unregister_all_RSA, e, &dummy_nid, 1, 1);
+				engine_unregister_all_RSA, e, &dummy_nid, 1, 1);
 	return 1;
 	}
 

crypto/evp/evp_test.c

@@ -126,6 +126,7 @@ static unsigned char *ustrsep(char **p,const char *sep)
 static int test1_exit(int ec)
 	{
 	EXIT(ec);
+	return(0);		/* To keep some compilers quiet */
 	}
 
 static void test1(const EVP_CIPHER *c,const unsigned char *key,int kn,

crypto/md32_common.h

@@ -198,7 +198,7 @@
    *
    * 					<appro@fy.chalmers.se>
    */
-#  if defined(__i386) || defined(__i386__)
+#  if defined(__i386) || defined(__i386__) || defined(__x86_64) || defined(__x86_64__)
 #   define ROTATE(a,n)	({ register unsigned int ret;	\
 				asm (			\
 				"roll %1,%0"		\
@@ -224,7 +224,7 @@
  */
 # if defined(__GNUC__) && __GNUC__>=2 && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM)
   /* some GNU C inline assembler templates by <appro@fy.chalmers.se> */
-#  if (defined(__i386) || defined(__i386__)) && !defined(I386_ONLY)
+#  if (defined(__i386) || defined(__i386__) || defined(__x86_64) || defined(__x86_64__)) && !defined(I386_ONLY)
 #   define BE_FETCH32(a)	({ register unsigned int l=(a);\
 				asm (			\
 				"bswapl %0"		\
@@ -610,3 +610,28 @@ int HASH_FINAL (unsigned char *md, HASH_CTX *c)
 	 */
 	return 1;
 	}
+
+#ifndef MD32_REG_T
+#define MD32_REG_T long
+/*
+ * This comment was originaly written for MD5, which is why it
+ * discusses A-D. But it basically applies to all 32-bit digests,
+ * which is why it was moved to common header file.
+ *
+ * In case you wonder why A-D are declared as long and not
+ * as MD5_LONG. Doing so results in slight performance
+ * boost on LP64 architectures. The catch is we don't
+ * really care if 32 MSBs of a 64-bit register get polluted
+ * with eventual overflows as we *save* only 32 LSBs in
+ * *either* case. Now declaring 'em long excuses the compiler
+ * from keeping 32 MSBs zeroed resulting in 13% performance
+ * improvement under SPARC Solaris7/64 and 5% under AlphaLinux.
+ * Well, to be honest it should say that this *prevents* 
+ * performance degradation.
+ *				<appro@fy.chalmers.se>
+ * Apparently there're LP64 compilers that generate better
+ * code if A-D are declared int. Most notably GCC-x86_64
+ * generates better code.
+ *				<appro@fy.chalmers.se>
+ */
+#endif

crypto/md4/md4_dgst.c

@@ -86,21 +86,7 @@ int MD4_Init(MD4_CTX *c)
 void md4_block_host_order (MD4_CTX *c, const void *data, int num)
 	{
 	const MD4_LONG *X=data;
-	register unsigned long A,B,C,D;
-	/*
-	 * In case you wonder why A-D are declared as long and not
-	 * as MD4_LONG. Doing so results in slight performance
-	 * boost on LP64 architectures. The catch is we don't
-	 * really care if 32 MSBs of a 64-bit register get polluted
-	 * with eventual overflows as we *save* only 32 LSBs in
-	 * *either* case. Now declaring 'em long excuses the compiler
-	 * from keeping 32 MSBs zeroed resulting in 13% performance
-	 * improvement under SPARC Solaris7/64 and 5% under AlphaLinux.
-	 * Well, to be honest it should say that this *prevents* 
-	 * performance degradation.
-	 *
-	 *				<appro@fy.chalmers.se>
-	 */
+	register unsigned MD32_REG_T A,B,C,D;
 
 	A=c->A;
 	B=c->B;
@@ -176,25 +162,11 @@ void md4_block_host_order (MD4_CTX *c, const void *data, int num)
 void md4_block_data_order (MD4_CTX *c, const void *data_, int num)
 	{
 	const unsigned char *data=data_;
-	register unsigned long A,B,C,D,l;
-	/*
-	 * In case you wonder why A-D are declared as long and not
-	 * as MD4_LONG. Doing so results in slight performance
-	 * boost on LP64 architectures. The catch is we don't
-	 * really care if 32 MSBs of a 64-bit register get polluted
-	 * with eventual overflows as we *save* only 32 LSBs in
-	 * *either* case. Now declaring 'em long excuses the compiler
-	 * from keeping 32 MSBs zeroed resulting in 13% performance
-	 * improvement under SPARC Solaris7/64 and 5% under AlphaLinux.
-	 * Well, to be honest it should say that this *prevents* 
-	 * performance degradation.
-	 *
-	 *				<appro@fy.chalmers.se>
-	 */
+	register unsigned MD32_REG_T A,B,C,D,l;
 #ifndef MD32_XARRAY
 	/* See comment in crypto/sha/sha_locl.h for details. */
-	unsigned long	XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
-			XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
+	unsigned MD32_REG_T	XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
+				XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
 # define X(i)	XX##i
 #else
 	MD4_LONG XX[MD4_LBLOCK];

crypto/md5/md5_dgst.c

@@ -86,21 +86,7 @@ int MD5_Init(MD5_CTX *c)
 void md5_block_host_order (MD5_CTX *c, const void *data, int num)
 	{
 	const MD5_LONG *X=data;
-	register unsigned long A,B,C,D;
-	/*
-	 * In case you wonder why A-D are declared as long and not
-	 * as MD5_LONG. Doing so results in slight performance
-	 * boost on LP64 architectures. The catch is we don't
-	 * really care if 32 MSBs of a 64-bit register get polluted
-	 * with eventual overflows as we *save* only 32 LSBs in
-	 * *either* case. Now declaring 'em long excuses the compiler
-	 * from keeping 32 MSBs zeroed resulting in 13% performance
-	 * improvement under SPARC Solaris7/64 and 5% under AlphaLinux.
-	 * Well, to be honest it should say that this *prevents* 
-	 * performance degradation.
-	 *
-	 *				<appro@fy.chalmers.se>
-	 */
+	register unsigned MD32_REG_T A,B,C,D;
 
 	A=c->A;
 	B=c->B;
@@ -193,25 +179,11 @@ void md5_block_host_order (MD5_CTX *c, const void *data, int num)
 void md5_block_data_order (MD5_CTX *c, const void *data_, int num)
 	{
 	const unsigned char *data=data_;
-	register unsigned long A,B,C,D,l;
-	/*
-	 * In case you wonder why A-D are declared as long and not
-	 * as MD5_LONG. Doing so results in slight performance
-	 * boost on LP64 architectures. The catch is we don't
-	 * really care if 32 MSBs of a 64-bit register get polluted
-	 * with eventual overflows as we *save* only 32 LSBs in
-	 * *either* case. Now declaring 'em long excuses the compiler
-	 * from keeping 32 MSBs zeroed resulting in 13% performance
-	 * improvement under SPARC Solaris7/64 and 5% under AlphaLinux.
-	 * Well, to be honest it should say that this *prevents* 
-	 * performance degradation.
-	 *
-	 *				<appro@fy.chalmers.se>
-	 */
+	register unsigned MD32_REG_T A,B,C,D,l;
 #ifndef MD32_XARRAY
 	/* See comment in crypto/sha/sha_locl.h for details. */
-	unsigned long	XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
-			XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
+	unsigned MD32_REG_T	XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
+				XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
 # define X(i)	XX##i
 #else
 	MD5_LONG XX[MD5_LBLOCK];

crypto/opensslconf.h.in

@@ -47,7 +47,7 @@
 #endif
 #endif
 
-#if (defined(HEADER_DES_H) || defined(HEADER_DES_OLD_H)) && !defined(DES_LONG)
+#if (defined(HEADER_NEW_DES_H) || defined(HEADER_DES_H)) && !defined(DES_LONG)
 /* If this is set to 'unsigned int' on a DEC Alpha, this gives about a
  * %20 speed up (longs are 8 bytes, int's are 4). */
 #ifndef DES_LONG

crypto/opensslv.h

@@ -25,8 +25,8 @@
  * (Prior to 0.9.5a beta1, a different scheme was used: MMNNFFRBB for
  *  major minor fix final patch/beta)
  */
-#define OPENSSL_VERSION_NUMBER	0x00907005L
-#define OPENSSL_VERSION_TEXT	"OpenSSL 0.9.7-beta5  5 Dec 2002"
+#define OPENSSL_VERSION_NUMBER	0x00907006L
+#define OPENSSL_VERSION_TEXT	"OpenSSL 0.9.7-beta6 17 Dec 2002"
 #define OPENSSL_VERSION_PTEXT	" part of " OPENSSL_VERSION_TEXT
 
 

crypto/perlasm/cbc.pl

@@ -146,9 +146,15 @@ sub cbc
 	&mov($count,	&wparam(2));	# length
 	&and($count,	7);
 	&jz(&label("finish"));
+	&call(&label("PIC_point"));
+&set_label("PIC_point");
+	&blindpop("edx");
+	&lea("ecx",&DWP(&label("cbc_enc_jmp_table")."-".&label("PIC_point"),"edx"));
+	&mov($count,&DWP(0,"ecx",$count,4))
+	&add($count,"edx");
 	&xor("ecx","ecx");
 	&xor("edx","edx");
-	&mov($count,&DWP(&label("cbc_enc_jmp_table"),"",$count,4));
+	#&mov($count,&DWP(&label("cbc_enc_jmp_table"),"",$count,4));
 	&jmp_ptr($count);
 
 &set_label("ej7");
@@ -318,22 +324,23 @@ sub cbc
 
 	&set_label("cbc_enc_jmp_table",1);
 	&data_word("0");
-	&data_word(&label("ej1"));
-	&data_word(&label("ej2"));
-	&data_word(&label("ej3"));
-	&data_word(&label("ej4"));
-	&data_word(&label("ej5"));
-	&data_word(&label("ej6"));
-	&data_word(&label("ej7"));
-	&set_label("cbc_dec_jmp_table",1);
-	&data_word("0");
-	&data_word(&label("dj1"));
-	&data_word(&label("dj2"));
-	&data_word(&label("dj3"));
-	&data_word(&label("dj4"));
-	&data_word(&label("dj5"));
-	&data_word(&label("dj6"));
-	&data_word(&label("dj7"));
+	&data_word(&label("ej1")."-".&label("PIC_point"));
+	&data_word(&label("ej2")."-".&label("PIC_point"));
+	&data_word(&label("ej3")."-".&label("PIC_point"));
+	&data_word(&label("ej4")."-".&label("PIC_point"));
+	&data_word(&label("ej5")."-".&label("PIC_point"));
+	&data_word(&label("ej6")."-".&label("PIC_point"));
+	&data_word(&label("ej7")."-".&label("PIC_point"));
+	# not used
+	#&set_label("cbc_dec_jmp_table",1);
+	#&data_word("0");
+	#&data_word(&label("dj1")."-".&label("PIC_point"));
+	#&data_word(&label("dj2")."-".&label("PIC_point"));
+	#&data_word(&label("dj3")."-".&label("PIC_point"));
+	#&data_word(&label("dj4")."-".&label("PIC_point"));
+	#&data_word(&label("dj5")."-".&label("PIC_point"));
+	#&data_word(&label("dj6")."-".&label("PIC_point"));
+	#&data_word(&label("dj7")."-".&label("PIC_point"));
 
 	&function_end_B($name);
 	

crypto/perlasm/x86ms.pl

@@ -92,6 +92,8 @@ sub get_mem
 		$addr="_$addr";
 		}
 
+	if ($addr =~ /^.+\-.+$/) { $addr="($addr)"; }
+
 	$reg1="$regs{$reg1}" if defined($regs{$reg1});
 	$reg2="$regs{$reg2}" if defined($regs{$reg2});
 	if (($addr ne "") && ($addr ne 0))
@@ -111,6 +113,7 @@ sub get_mem
 		{
 		$ret.="[$reg1$post]"
 		}
+	$ret =~ s/\[\]//;	# in case $addr was the only argument
 	return($ret);
 	}
 
@@ -151,7 +154,7 @@ sub main'push	{ &out1("push",@_); $stack+=4; }
 sub main'pop	{ &out1("pop",@_); $stack-=4; }
 sub main'bswap	{ &out1("bswap",@_); &using486(); }
 sub main'not	{ &out1("not",@_); }
-sub main'call	{ &out1("call",'_'.$_[0]); }
+sub main'call	{ &out1("call",($_[0]=~/^\$L/?'':'_').$_[0]); }
 sub main'ret	{ &out0("ret"); }
 sub main'nop	{ &out0("nop"); }
 
@@ -338,7 +341,7 @@ sub main'set_label
 	{
 	if (!defined($label{$_[0]}))
 		{
-		$label{$_[0]}="${label}${_[0]}";
+		$label{$_[0]}="\$${label}${_[0]}";
 		$label++;
 		}
 	if((defined $_[2]) && ($_[2] == 1))
@@ -363,3 +366,5 @@ sub out1p
 
 	push(@out,"\t$name\t ".&conv($p1)."\n");
 	}
+
+sub main'blindpop { &out1("pop",@_); }

crypto/perlasm/x86nasm.pl

@@ -98,6 +98,8 @@ sub get_mem
 		$addr="_$addr";
 		}
 
+	if ($addr =~ /^.+\-.+$/) { $addr="($addr)"; }
+
 	$reg1="$regs{$reg1}" if defined($regs{$reg1});
 	$reg2="$regs{$reg2}" if defined($regs{$reg2});
 	if (($addr ne "") && ($addr ne 0))
@@ -117,6 +119,7 @@ sub get_mem
 		{
 		$ret.="$reg1$post]"
 		}
+	$ret =~ s/\+\]/]/; # in case $addr was the only argument
 	return($ret);
 	}
 
@@ -160,7 +163,7 @@ sub main'push	{ &out1("push",@_); $stack+=4; }
 sub main'pop	{ &out1("pop",@_); $stack-=4; }
 sub main'bswap	{ &out1("bswap",@_); &using486(); }
 sub main'not	{ &out1("not",@_); }
-sub main'call	{ &out1("call",'_'.$_[0]); }
+sub main'call	{ &out1("call",($_[0]=~/^\$L/?'':'_').$_[0]); }
 sub main'ret	{ &out0("ret"); }
 sub main'nop	{ &out0("nop"); }
 
@@ -322,7 +325,7 @@ sub main'set_label
 	{
 	if (!defined($label{$_[0]}))
 		{
-		$label{$_[0]}="${label}${_[0]}";
+		$label{$_[0]}="\$${label}${_[0]}";
 		$label++;
 		}
 	push(@out,"$label{$_[0]}:\n");
@@ -340,3 +343,5 @@ sub out1p
 
 	push(@out,"\t$name\t ".&conv($p1)."\n");
 	}
+
+sub main'blindpop { &out1("pop",@_); }

crypto/perlasm/x86unix.pl

@@ -87,12 +87,12 @@ sub main'DWP
 	$ret.=$addr if ($addr ne "") && ($addr ne 0);
 	if ($reg2 ne "")
 		{
-		if($idx ne "")
+		if($idx ne "" && $idx != 0)
 		    { $ret.="($reg1,$reg2,$idx)"; }
 		else
 		    { $ret.="($reg1,$reg2)"; }
 	        }
-	else
+	elsif ($reg1 ne "")
 		{ $ret.="($reg1)" }
 	return($ret);
 	}
@@ -167,7 +167,7 @@ sub main'pop	{ &out1("popl",@_); $stack-=4; }
 sub main'pushf	{ &out0("pushf"); $stack+=4; }
 sub main'popf	{ &out0("popf"); $stack-=4; }
 sub main'not	{ &out1("notl",@_); }
-sub main'call	{ &out1("call",$under.$_[0]); }
+sub main'call	{ &out1("call",($_[0]=~/^\.L/?'':$under).$_[0]); }
 sub main'ret	{ &out0("ret"); }
 sub main'nop	{ &out0("nop"); }
 
@@ -542,3 +542,22 @@ sub popvars
 	&main'pop("edx");
 	&main'popf();
 	}
+
+sub main'picmeup
+	{
+	local($dst,$sym)=@_;
+	local($tmp)=<<___;
+#if (defined(ELF) || defined(SOL)) && defined(PIC)
+	.align	8
+	call	1f
+1:	popl	$regs{$dst}
+	addl	\$_GLOBAL_OFFSET_TABLE_+[.-1b],$regs{$dst}
+	movl	$sym\@GOT($regs{$dst}),$regs{$dst}
+#else
+	leal	$sym,$regs{$dst}
+#endif
+___
+	push(@out,$tmp);
+	}
+
+sub main'blindpop { &out1("popl",@_); }

crypto/rc5/rc5_locl.h

@@ -149,8 +149,28 @@
 #if defined(OPENSSL_SYS_WIN32) && defined(_MSC_VER)
 #define ROTATE_l32(a,n)     _lrotl(a,n)
 #define ROTATE_r32(a,n)     _lrotr(a,n)
-#else
+#elif defined(__GNUC__) && __GNUC__>=2 && !defined(__STRICT_ANSI__) && !defined(NO_ASM) && !defined(NO_INLINE_ASM)
+# if defined(__i386) || defined(__i386__) || defined(__x86_64) || defined(__x86_64__)
+#  define ROTATE_l32(a,n)	({ register unsigned int ret;	\
+					asm ("roll %%cl,%0"	\
+						: "=r"(ret)	\
+						: "c"(n),"0"(a)	\
+						: "cc");	\
+					ret;			\
+				})
+#  define ROTATE_r32(a,n)	({ register unsigned int ret;	\
+					asm ("rorl %%cl,%0"	\
+						: "=r"(ret)	\
+						: "c"(n),"0"(a)	\
+						: "cc");	\
+					ret;			\
+				})
+# endif
+#endif
+#ifndef ROTATE_l32
 #define ROTATE_l32(a,n)     (((a)<<(n&0x1f))|(((a)&0xffffffff)>>(32-(n&0x1f))))
+#endif
+#ifndef ROTATE_r32
 #define ROTATE_r32(a,n)     (((a)<<(32-(n&0x1f)))|(((a)&0xffffffff)>>(n&0x1f)))
 #endif
 

crypto/ripemd/rmd_dgst.c

@@ -90,8 +90,8 @@ int RIPEMD160_Init(RIPEMD160_CTX *c)
 void ripemd160_block_host_order (RIPEMD160_CTX *ctx, const void *p, int num)
 	{
 	const RIPEMD160_LONG *XX=p;
-	register unsigned long A,B,C,D,E;
-	register unsigned long a,b,c,d,e;
+	register unsigned MD32_REG_T A,B,C,D,E;
+	register unsigned MD32_REG_T a,b,c,d,e;
 
 	for (;num--;XX+=HASH_LBLOCK)
 		{
@@ -290,12 +290,12 @@ void ripemd160_block_host_order (RIPEMD160_CTX *ctx, const void *p, int num)
 void ripemd160_block_data_order (RIPEMD160_CTX *ctx, const void *p, int num)
 	{
 	const unsigned char *data=p;
-	register unsigned long A,B,C,D,E;
-	unsigned long a,b,c,d,e,l;
+	register unsigned MD32_REG_T A,B,C,D,E;
+	unsigned MD32_REG_T a,b,c,d,e,l;
 #ifndef MD32_XARRAY
 	/* See comment in crypto/sha/sha_locl.h for details. */
-	unsigned long	XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
-			XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
+	unsigned MD32_REG_T	XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
+				XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
 # define X(i)	XX##i
 #else
 	RIPEMD160_LONG	XX[16];

crypto/sha/sha_locl.h

@@ -224,10 +224,10 @@ int HASH_INIT (SHA_CTX *c)
 void HASH_BLOCK_HOST_ORDER (SHA_CTX *c, const void *d, int num)
 	{
 	const SHA_LONG *W=d;
-	register unsigned long A,B,C,D,E,T;
+	register unsigned MD32_REG_T A,B,C,D,E,T;
 #ifndef MD32_XARRAY
-	unsigned long	XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
-			XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
+	unsigned MD32_REG_T	XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
+				XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
 #else
 	SHA_LONG	XX[16];
 #endif
@@ -349,10 +349,10 @@ void HASH_BLOCK_HOST_ORDER (SHA_CTX *c, const void *d, int num)
 void HASH_BLOCK_DATA_ORDER (SHA_CTX *c, const void *p, int num)
 	{
 	const unsigned char *data=p;
-	register unsigned long A,B,C,D,E,T,l;
+	register unsigned MD32_REG_T A,B,C,D,E,T,l;
 #ifndef MD32_XARRAY
-	unsigned long	XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
-			XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
+	unsigned MD32_REG_T	XX0, XX1, XX2, XX3, XX4, XX5, XX6, XX7,
+				XX8, XX9,XX10,XX11,XX12,XX13,XX14,XX15;
 #else
 	SHA_LONG	XX[16];
 #endif

crypto/x509/x509_vfy.c

@@ -1158,7 +1158,7 @@ void X509_STORE_CTX_cleanup(X509_STORE_CTX *ctx)
 		ctx->chain=NULL;
 		}
 	CRYPTO_free_ex_data(CRYPTO_EX_INDEX_X509_STORE_CTX, ctx, &(ctx->ex_data));
-	OPENSSL_cleanse(&ctx->ex_data,sizeof(CRYPTO_EX_DATA));
+	memset(&ctx->ex_data,0,sizeof(CRYPTO_EX_DATA));
 	}
 
 void X509_STORE_CTX_set_flags(X509_STORE_CTX *ctx, long flags)

doc/crypto/BIO_new_bio_pair.pod

@@ -1,103 +0,0 @@
-=pod
-
-=head1 NAME
-
-BIO_new_bio_pair - create a new BIO pair
-
-=head1 SYNOPSIS
-
- #include <openssl/bio.h>
-
- int BIO_new_bio_pair(BIO **bio1, size_t writebuf1, BIO **bio2, size_t writebuf2);
-
-=head1 DESCRIPTION
-
-BIO_new_bio_pair() creates a buffering BIO pair based on the
-L<SSL_set_bio(3)|SSL_set_bio(3)> method. The BIO pair has two endpoints between which
-data can be buffered. Its typical use is to connect one endpoint as underlying
-input/output BIO to an SSL and access the other one controlled by the program
-instead of accessing the network connection directly.
-
-The two new BIOs B<bio1> and B<bio2> are symmetric with respect to their
-functionality. The size of their buffers is determined by B<writebuf1> and
-B<writebuf2>. If the size give is 0, the default size is used.
-
-BIO_new_bio_pair() does not check whether B<bio1> or B<bio2> do point to
-some other BIO, the values are overwritten, BIO_free() is not called.
-
-The two BIOs, even though forming a BIO pair and must be BIO_free()'ed
-separately. This can be of importance, as some SSL-functions like SSL_set_bio()
-or SSL_free() call BIO_free() implicitly, so that the peer-BIO is left
-untouched and must also be BIO_free()'ed.
-
-=head1 EXAMPLE
-
-The BIO pair can be used to have full control over the network access of an
-application. The application can call select() on the socket as required
-without having to go through the SSL-interface.
-
- BIO *internal_bio, *network_bio;
- ...
- BIO_new_bio_pair(internal_bio, 0, network_bio, 0);
- SSL_set_bio(ssl, internal_bio, internal_bio);
- SSL_operations();
- ...
-
- application |   TLS-engine
-    |        |
-    +----------> SSL_operations()
-             |     /\    ||
-             |     ||    \/
-             |   BIO-pair (internal_bio)
-    +----------< BIO-pair (network_bio)
-    |        |
-  socket     |
-
-  ...
-  SSL_free(ssl);		/* implicitly frees internal_bio */
-  BIO_free(network_bio);
-  ...
-
-As the BIO pair will only buffer the data and never directly access the
-connection, it behaves non-blocking and will return as soon as the write
-buffer is full or the read buffer is drained. Then the application has to
-flush the write buffer and/or fill the read buffer.
-
-Use the BIO_ctrl_pending(), to find out whether data is buffered in the BIO
-and must be transfered to the network. Use BIO_ctrl_get_read_request() to
-find out, how many bytes must be written into the buffer before the
-SSL_operation() can successfully be continued.
-
-=head1 WARNING
-
-As the data is buffered, SSL_operation() may return with a ERROR_SSL_WANT_READ
-condition, but there is still data in the write buffer. An application must
-not rely on the error value of SSL_operation() but must assure that the
-write buffer is always flushed first. Otherwise a deadlock may occur as
-the peer might be waiting for the data before being able to continue.
-
-=head1 RETURN VALUES
-
-The following return values can occur:
-
-=over 4
-
-=item 1
-
-The BIO pair was created successfully. The new BIOs are available in
-B<bio1> and B<bio2>.
-
-=item 0
-
-The operation failed. The NULL pointer is stored into the locations for
-B<bio1> and B<bio2>. Check the error stack for more information.
-
-=back
-
-=head1 SEE ALSO
-
-L<SSL_set_bio(3)|SSL_set_bio(3)>, L<ssl(3)|ssl(3)>, L<bio(3)|bio(3)>,
-L<BIO_ctrl_pending(3)|BIO_ctrl_pending(3)>,
-L<BIO_ctrl_get_read_request(3)|BIO_ctrl_get_read_request(3)>
-
-=cut

doc/crypto/BIO_s_accept.pod

@@ -2,7 +2,7 @@
 
 =head1 NAME
 
-BIO_s_accept, BIO_set_nbio, BIO_set_accept_port, BIO_get_accept_port,
+BIO_s_accept, BIO_set_accept_port, BIO_get_accept_port,
 BIO_set_nbio_accept, BIO_set_accept_bios, BIO_set_bind_mode,
 BIO_get_bind_mode, BIO_do_accept - accept BIO
 

doc/crypto/BIO_s_bio.pod

@@ -76,7 +76,9 @@ BIO_get_write_buf_size() returns the size of the write buffer.
 BIO_new_bio_pair() combines the calls to BIO_new(), BIO_make_bio_pair() and
 BIO_set_write_buf_size() to create a connected pair of BIOs B<bio1>, B<bio2>
 with write buffer sizes B<writebuf1> and B<writebuf2>. If either size is
-zero then the default size is used.
+zero then the default size is used.  BIO_new_bio_pair() does not check whether
+B<bio1> or B<bio2> do point to some other BIO, the values are overwritten,
+BIO_free() is not called.
 
 BIO_get_write_guarantee() and BIO_ctrl_get_write_guarantee() return the maximum
 length of data that can be currently written to the BIO. Writes larger than this
@@ -118,9 +120,59 @@ the application then waits for data to be available on the underlying transport
 before flushing the write buffer it will never succeed because the request was
 never sent!
 
+=head1 RETURN VALUES
+
+BIO_new_bio_pair() returns 1 on success, with the new BIOs available in
+B<bio1> and B<bio2>, or 0 on failure, with NULL pointers stored into the
+locations for B<bio1> and B<bio2>. Check the error stack for more information.
+
+[XXXXX: More return values need to be added here]
+
 =head1 EXAMPLE
 
-TBA
+The BIO pair can be used to have full control over the network access of an
+application. The application can call select() on the socket as required
+without having to go through the SSL-interface.
+
+ BIO *internal_bio, *network_bio;
+ ...
+ BIO_new_bio_pair(internal_bio, 0, network_bio, 0);
+ SSL_set_bio(ssl, internal_bio, internal_bio);
+ SSL_operations();
+ ...
+
+ application |   TLS-engine
+    |        |
+    +----------> SSL_operations()
+             |     /\    ||
+             |     ||    \/
+             |   BIO-pair (internal_bio)
+    +----------< BIO-pair (network_bio)
+    |        |
+  socket     |
+
+  ...
+  SSL_free(ssl);		/* implicitly frees internal_bio */
+  BIO_free(network_bio);
+  ...
+
+As the BIO pair will only buffer the data and never directly access the
+connection, it behaves non-blocking and will return as soon as the write
+buffer is full or the read buffer is drained. Then the application has to
+flush the write buffer and/or fill the read buffer.
+
+Use the BIO_ctrl_pending(), to find out whether data is buffered in the BIO
+and must be transfered to the network. Use BIO_ctrl_get_read_request() to
+find out, how many bytes must be written into the buffer before the
+SSL_operation() can successfully be continued.
+
+=head1 WARNING
+
+As the data is buffered, SSL_operation() may return with a ERROR_SSL_WANT_READ
+condition, but there is still data in the write buffer. An application must
+not rely on the error value of SSL_operation() but must assure that the
+write buffer is always flushed first. Otherwise a deadlock may occur as
+the peer might be waiting for the data before being able to continue.
 
 =head1 SEE ALSO
 

doc/crypto/engine.pod

@@ -187,7 +187,7 @@ tell which one you are dealing with at any given point in time (after all
 they are both simply (ENGINE *) pointers, the difference is in the way they
 are used).
 
-=head3 Structural references
+I<Structural references>
 
 This basic type of reference is typically used for creating new ENGINEs
 dynamically, iterating across OpenSSL's internal linked-list of loaded
@@ -224,7 +224,7 @@ To clarify a particular function's handling of references, one should
 always consult that function's documentation "man" page, or failing that
 the openssl/engine.h header file includes some hints.
 
-=head3 Functional references
+I<Functional references>
 
 As mentioned, functional references exist when the cryptographic
 functionality of an ENGINE is required to be available. A functional
@@ -386,7 +386,7 @@ things, so we will simply illustrate the consequences as they apply to a
 couple of simple cases and leave developers to consider these and the
 source code to openssl's builtin utilities as guides.
 
-=head3 Using a specific ENGINE implementation
+I<Using a specific ENGINE implementation>
 
 Here we'll assume an application has been configured by its user or admin
 to want to use the "ACME" ENGINE if it is available in the version of
@@ -418,7 +418,7 @@ illustrates how to approach this;
  /* Release the structural reference from ENGINE_by_id() */
  ENGINE_free(e);
 
-=head3 Automatically using builtin ENGINE implementations
+I<Automatically using builtin ENGINE implementations>
 
 Here we'll assume we want to load and register all ENGINE implementations
 bundled with OpenSSL, such that for any cryptographic algorithm required by
@@ -469,7 +469,7 @@ in same cases both. ENGINE implementations should provide indications of
 this in the descriptions attached to builtin control commands and/or in
 external product documentation.
 
-=head3 Issuing control commands to an ENGINE
+I<Issuing control commands to an ENGINE>
 
 Let's illustrate by example; a function for which the caller supplies the
 name of the ENGINE it wishes to use, a table of string-pairs for use before
@@ -526,7 +526,7 @@ return success without doing anything. In this case we assume the user is
 only supplying commands specific to the given ENGINE so we set this to
 FALSE.
 
-=head3 Discovering supported control commands
+I<Discovering supported control commands>
 
 It is possible to discover at run-time the names, numerical-ids, descriptions
 and input parameters of the control commands supported from a structural

ssl/ssl_ciph.c

@@ -751,7 +751,12 @@ STACK_OF(SSL_CIPHER) *ssl_create_cipher_list(const SSL_METHOD *ssl_method,
 	 */
 	if (rule_str == NULL) return(NULL);
 
-	if (init_ciphers) load_ciphers();
+	if (init_ciphers)
+		{
+		CRYPTO_w_lock(CRYPTO_LOCK_SSL);
+		if (init_ciphers) load_ciphers();
+		CRYPTO_w_unlock(CRYPTO_LOCK_SSL);
+		}
 
 	/*
 	 * To reduce the work to do we only want to process the compiled

test/Makefile.ssl

@@ -119,7 +119,7 @@ apps:
 	@(cd ..; $(MAKE) DIRS=apps all)
 
 SET_SO_PATHS=LIBPATH="`cd ..; pwd`"; LD_LIBRARY_PATH="$$LIBPATH"; DYLD_LIBRARY_PATH="$$LIBPATH"; SHLIB_PATH="$$LIBPATH"; \
-		if [ "$(PLATFORM)" == "DJGPP" ]; then PATH="$$LIBPATH\;$$PATH";  \
+		if [ "$(PLATFORM)" = "DJGPP" ]; then PATH="$$LIBPATH\;$$PATH";  \
 		elif [ "$(PLATFORM)" != "Cygwin" ];  then PATH="$$LIBPATH:$$PATH"; fi; \
 		export LD_LIBRARY_PATH DYLD_LIBRARY_PATH SHLIB_PATH LIBPATH PATH
 

test/testssl

@@ -121,8 +121,12 @@ $ssltest -bio_pair -server_auth -client_auth -app_verify $CA $extra || exit 1
 
 #############################################################################
 
-echo test tls1 with 1024bit anonymous DH, multiple handshakes
-$ssltest -v -bio_pair -tls1 -cipher ADH -dhe1024dsa -num 10 -f -time $extra || exit 1
+if ../apps/openssl no-dh; then
+  echo skipping anonymous DH tests
+else
+  echo test tls1 with 1024bit anonymous DH, multiple handshakes
+  $ssltest -v -bio_pair -tls1 -cipher ADH -dhe1024dsa -num 10 -f -time $extra || exit 1
+fi
 
 if ../apps/openssl no-rsa; then
   echo skipping RSA tests
@@ -130,8 +134,12 @@ else
   echo test tls1 with 1024bit RSA, no DHE, multiple handshakes
   ./ssltest -v -bio_pair -tls1 -cert ../apps/server2.pem -no_dhe -num 10 -f -time $extra || exit 1
 
-  echo test tls1 with 1024bit RSA, 1024bit DHE, multiple handshakes
-  ./ssltest -v -bio_pair -tls1 -cert ../apps/server2.pem -dhe1024dsa -num 10 -f -time $extra || exit 1
+  if ../apps/openssl no-dh; then
+    echo skipping RSA+DHE tests
+  else
+    echo test tls1 with 1024bit RSA, 1024bit DHE, multiple handshakes
+    ./ssltest -v -bio_pair -tls1 -cert ../apps/server2.pem -dhe1024dsa -num 10 -f -time $extra || exit 1
+  fi
 fi
 
 exit 0

test/testssl.com

@@ -159,16 +159,25 @@ $	if $severity .ne. 1 then goto exit3
 $
 $!###########################################################################
 $
-$	write sys$output "test tls1 with 1024bit anonymous DH, multiple handshakes"
-$	'ssltest' -v -bio_pair -tls1 -cipher "ADH" -dhe1024dsa -num 10 -f -time
-$	if $severity .ne. 1 then goto exit3
-$
 $	set noon
 $	define/user sys$output nla0:
 $	mcr 'exe_dir'openssl no-rsa
-$	save_severity=$SEVERITY
+$	no_rsa=$SEVERITY
+$	define/user sys$output nla0:
+$	mcr 'exe_dir'openssl no-dh
+$	no_dh=$SEVERITY
 $	set on
-$	if save_severity
+$
+$	if no_dh
+$	then
+$	    write sys$output "skipping anonymous DH tests"
+$	else
+$	    write sys$output "test tls1 with 1024bit anonymous DH, multiple handshakes"
+$	    'ssltest' -v -bio_pair -tls1 -cipher "ADH" -dhe1024dsa -num 10 -f -time
+$	    if $severity .ne. 1 then goto exit3
+$	endif
+$
+$	if no_rsa
 $	then
 $	    write sys$output "skipping RSA tests"
 $	else
@@ -176,9 +185,14 @@ $	    write sys$output "test tls1 with 1024bit RSA, no DHE, multiple handshakes"
 $	    mcr 'texe_dir'ssltest -v -bio_pair -tls1 -cert [-.apps]server2.pem -no_dhe -num 10 -f -time
 $	    if $severity .ne. 1 then goto exit3
 $
-$	    write sys$output "test tls1 with 1024bit RSA, 1024bit DHE, multiple handshakes"
-$	    mcr 'texe_dir'ssltest -v -bio_pair -tls1 -cert [-.apps]server2.pem -dhe1024dsa -num 10 -f -time
-$	    if $severity .ne. 1 then goto exit3
+$	    if no_dh
+$	    then
+$		write sys$output "skipping RSA+DHE tests"
+$	    else
+$		write sys$output "test tls1 with 1024bit RSA, 1024bit DHE, multiple handshakes"
+$		mcr 'texe_dir'ssltest -v -bio_pair -tls1 -cert [-.apps]server2.pem -dhe1024dsa -num 10 -f -time
+$		if $severity .ne. 1 then goto exit3
+$	    endif
 $	endif
 $
 $	RET = 1

tools/Makefile.ssl

@@ -27,13 +27,15 @@ all:
 install:
 	@for i in $(APPS) ; \
 	do  \
-	(cp $$i $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$i; \
-	chmod 755 $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$i ); \
+	(cp $$i $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$i.new; \
+	chmod 755 $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$i.new; \
+	mv -f $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$i.new $(INSTALL_PREFIX)$(INSTALLTOP)/bin/$$i ); \
 	done;
 	@for i in $(MISC_APPS) ; \
 	do  \
-	(cp $$i $(INSTALL_PREFIX)$(OPENSSLDIR)/misc/$$i; \
-	chmod 755 $(INSTALL_PREFIX)$(OPENSSLDIR)/misc/$$i ); \
+	(cp $$i $(INSTALL_PREFIX)$(OPENSSLDIR)/misc/$$i.new; \
+	chmod 755 $(INSTALL_PREFIX)$(OPENSSLDIR)/misc/$$i.new; \
+	mv -f $(INSTALL_PREFIX)$(INSTALLTOP)/misc/$$i.new $(INSTALL_PREFIX)$(INSTALLTOP)/misc/$$i ); \
 	done;
 
 files: