Add crossbuild patch to 1.2 FIPS branch.

1_2-stable'.

CHANGES

@@ -4,6 +4,37 @@
 
  Changes between 0.9.8e and 0.9.8f-fips  [xx XXX xxxx]
 
+
+  *) Add option --with-fipslibdir to specify location of fipscanister.lib
+     and friends. When combined with fips build option fipscanister.lib is
+     not built but linked from the supplied directory. Always link fips
+     utilities against fiscanister.lib only except in fipsdso builds.
+     [Steve Henson]
+
+  *) Add SSE2 instruction support to WIN32 build. These will be compiled
+     by default and used if an appopriate CPU is detected. Some older versions
+     of NASM or MASM which don't support SSE2 will need to be updated. 
+     [Steve Henson]
+
+  *) Tolerate DigestInfo structure with absent parameters in FIPS mode
+     (as required by several standards).
+     [Steve Henson]
+
+  *) Enhance mkfipsscr.pl to cope with different directory layouts. It now
+     relies on the filename and makes no assumptions about the pathname.
+     In the case of PSS it scans the file to determine the salt length.
+     Paths can be filtered. Also reports duplicate and missing files.
+     [Steve Henson]
+
+  *) Updates to WIN32 build system. Make use of AES assembly language routines.
+     Use assembly language routines in FIPS compilation.
+     [Steve Henson]
+
+  *) Use standard implementations of SHAx, DES, AES under crypto/ in FIPS
+     mode to avoid having to maintain two versions. This will also make use
+     of appropriate assembly language optimizations.
+     [Steve Henson]
+
   *) Check for selftest status in all crypto operations and exit with a
      fatal error if selftest failed.
      [Steve Henson]

Configure

@@ -116,15 +116,17 @@ my $tlib="-lnsl -lsocket";
 my $bits1="THIRTY_TWO_BIT ";
 my $bits2="SIXTY_FOUR_BIT ";
 
-my $x86_elf_asm="x86cpuid-elf.o:bn86-elf.o co86-elf.o:dx86-elf.o yx86-elf.o:ax86-elf.o:bx86-elf.o:mx86-elf.o:sx86-elf.o s512sse2-elf.o:cx86-elf.o:rx86-elf.o:rm86-elf.o:r586-elf.o";
+my $x86_elf_asm="x86cpuid-elf.o:bn86-elf.o co86-elf.o mo86-elf.o:dx86-elf.o yx86-elf.o:ax86-elf.o:bx86-elf.o:mx86-elf.o:sx86-elf.o s512sse2-elf.o:cx86-elf.o:rx86-elf.o:rm86-elf.o:r586-elf.o";
-my $x86_coff_asm="x86cpuid-cof.o:bn86-cof.o co86-cof.o:dx86-cof.o yx86-cof.o:ax86-cof.o:bx86-cof.o:mx86-cof.o:sx86-cof.o s512sse2-cof.o:cx86-cof.o:rx86-cof.o:rm86-cof.o:r586-cof.o";
+my $x86_coff_asm="x86cpuid-cof.o:bn86-cof.o co86-cof.o mo86-cof.o:dx86-cof.o yx86-cof.o:ax86-cof.o:bx86-cof.o:mx86-cof.o:sx86-cof.o s512sse2-cof.o:cx86-cof.o:rx86-cof.o:rm86-cof.o:r586-cof.o";
-my $x86_out_asm="x86cpuid-out.o:bn86-out.o co86-out.o:dx86-out.o yx86-out.o:ax86-out.o:bx86-out.o:mx86-out.o:sx86-out.o s512sse2-out.o:cx86-out.o:rx86-out.o:rm86-out.o:r586-out.o";
+my $x86_out_asm="x86cpuid-out.o:bn86-out.o co86-out.o mo86-out.o:dx86-out.o yx86-out.o:ax86-out.o:bx86-out.o:mx86-out.o:sx86-out.o s512sse2-out.o:cx86-out.o:rx86-out.o:rm86-out.o:r586-out.o";
 
-my $x86_64_asm="x86_64cpuid.o:x86_64-gcc.o::::md5-x86_64.o:::rc4-x86_64.o::";
+my $x86_64_asm="x86_64cpuid.o:x86_64-gcc.o x86_64-mont.o::aes-x86_64.o::md5-x86_64.o:sha1-x86_64.o sha256-x86_64.o sha512-x86_64.o::rc4-x86_64.o::";
-my $ia64_asm=":bn-ia64.o::aes_core.o aes_cbc.o aes-ia64.o:::sha1-ia64.o sha256-ia64.o sha512-ia64.o::rc4-ia64.o::";
 
 my $no_asm="::::::::::";
 
+my $ia64_asm=$no_asm;
+my $s390x_asm=$no_asm;
+
 # As for $BSDthreads. Idea is to maintain "collective" set of flags,
 # which would cover all BSD flavors. -pthread applies to them all, 
 # but is treated differently. OpenBSD expands is as -D_POSIX_THREAD
@@ -201,25 +203,25 @@ my %table=(
 
 #### 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:${no_asm}:dlfcn:solaris-shared:-fPIC:-shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"solaris-sparcv8-gcc","gcc:-mv8 -O3 -fomit-frame-pointer -Wall -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR::sparcv8.o:des_enc-sparc.o fcrypt_b.o:::::::::dlfcn:solaris-shared:-fPIC:-shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"solaris-sparcv8-gcc","gcc:-mv8 -O3 -fomit-frame-pointer -Wall -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:${no_asm}: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::sparcv8plus.o:des_enc-sparc.o fcrypt_b.o:::md5-sparcv8plus.o::::::dlfcn:solaris-shared:-fPIC:-shared:.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:${no_asm}:dlfcn:solaris-shared:-fPIC:-shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"solaris64-sparcv9-gcc","gcc:-m64 -mcpu=ultrasparc -O3 -Wall -DB_ENDIAN::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_INT DES_PTR DES_RISC1 DES_UNROLL BF_PTR:::des_enc-sparc.o fcrypt_b.o:::md5-sparcv9.o::::::dlfcn:solaris-shared:-fPIC:-m64 -shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"solaris64-sparcv9-gcc","gcc:-m64 -mcpu=ultrasparc -O3 -Wall -DB_ENDIAN::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_INT DES_PTR DES_RISC1 DES_UNROLL BF_PTR:${no_asm}:dlfcn:solaris-shared:-fPIC:-m64 -shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 ####
-"debug-solaris-sparcv8-gcc","gcc:-DBN_DEBUG -DREF_CHECK -DCONF_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::sparcv8.o::::::::::dlfcn:solaris-shared:-fPIC:-shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"debug-solaris-sparcv8-gcc","gcc:-DBN_DEBUG -DREF_CHECK -DCONF_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:${no_asm}:dlfcn:solaris-shared:-fPIC:-shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"debug-solaris-sparcv9-gcc","gcc:-DBN_DEBUG -DREF_CHECK -DCONF_DEBUG -DCRYPTO_MDEBUG_ALL -DPEDANTIC -O -g -mcpu=ultrasparc -pedantic -ansi -Wall -Wshadow -Wno-long-long -D__EXTENSIONS__ -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR::sparcv8plus.o:des_enc-sparc.o fcrypt_b.o:::md5-sparcv8plus.o::::::dlfcn:solaris-shared:-fPIC:-shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"debug-solaris-sparcv9-gcc","gcc:-DBN_DEBUG -DREF_CHECK -DCONF_DEBUG -DCRYPTO_MDEBUG_ALL -DPEDANTIC -O -g -mcpu=ultrasparc -pedantic -ansi -Wall -Wshadow -Wno-long-long -D__EXTENSIONS__ -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:${no_asm}:dlfcn:solaris-shared:-fPIC:-shared:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 
 #### SPARC Solaris with Sun C setups
 # SC4.0 doesn't pass 'make test', upgrade to SC5.0 or SC4.2.
 # SC4.2 is ok, better than gcc even on bn as long as you tell it -xarch=v8
 # SC5.0 note: Compiler common patch 107357-01 or later is required!
 "solaris-sparcv7-cc","cc:-xO5 -xstrconst -xdepend -Xa -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_PTR DES_RISC1 DES_UNROLL BF_PTR:${no_asm}:dlfcn:solaris-shared:-KPIC:-G -dy -z text:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"solaris-sparcv8-cc","cc:-xarch=v8 -xO5 -xstrconst -xdepend -Xa -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_PTR DES_RISC1 DES_UNROLL BF_PTR::sparcv8.o:des_enc-sparc.o fcrypt_b.o:::::::::dlfcn:solaris-shared:-KPIC:-G -dy -z text:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"solaris-sparcv8-cc","cc:-xarch=v8 -xO5 -xstrconst -xdepend -Xa -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_PTR DES_RISC1 DES_UNROLL BF_PTR:${no_asm}:dlfcn:solaris-shared:-KPIC:-G -dy -z text:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"solaris-sparcv9-cc","cc:-xtarget=ultra -xarch=v8plus -xO5 -xstrconst -xdepend -Xa -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK_LL DES_PTR DES_RISC1 DES_UNROLL BF_PTR::sparcv8plus.o:des_enc-sparc.o fcrypt_b.o:::md5-sparcv8plus.o::::::dlfcn:solaris-shared:-KPIC:-G -dy -z text:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"solaris-sparcv9-cc","cc:-xtarget=ultra -xarch=v8plus -xO5 -xstrconst -xdepend -Xa -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK_LL DES_PTR DES_RISC1 DES_UNROLL BF_PTR:${no_asm}:dlfcn:solaris-shared:-KPIC:-G -dy -z text:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"solaris64-sparcv9-cc","cc:-xtarget=ultra -xarch=v9 -xO5 -xstrconst -xdepend -Xa -DB_ENDIAN::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_INT DES_PTR DES_RISC1 DES_UNROLL BF_PTR:::des_enc-sparc.o fcrypt_b.o:::md5-sparcv9.o::::::dlfcn:solaris-shared:-KPIC:-xarch=v9 -G -dy -z text:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR):/usr/ccs/bin/ar rs",
+"solaris64-sparcv9-cc","cc:-xtarget=ultra -xarch=v9 -xO5 -xstrconst -xdepend -Xa -DB_ENDIAN::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_INT DES_PTR DES_RISC1 DES_UNROLL BF_PTR:${no_asm}:dlfcn:solaris-shared:-KPIC:-xarch=v9 -G -dy -z text:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR):/usr/ccs/bin/ar rs",
 ####
-"debug-solaris-sparcv8-cc","cc:-DBN_DEBUG -DREF_CHECK -DCONF_DEBUG -DCRYPTO_MDEBUG_ALL -xarch=v8 -g -O -xstrconst -Xa -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_PTR DES_RISC1 DES_UNROLL BF_PTR::sparcv8.o::::::::::dlfcn:solaris-shared:-KPIC:-G -dy -z text:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"debug-solaris-sparcv8-cc","cc:-DBN_DEBUG -DREF_CHECK -DCONF_DEBUG -DCRYPTO_MDEBUG_ALL -xarch=v8 -g -O -xstrconst -Xa -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT::-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_PTR DES_RISC1 DES_UNROLL BF_PTR:${no_asm}:dlfcn:solaris-shared:-KPIC:-G -dy -z text:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"debug-solaris-sparcv9-cc","cc:-DBN_DEBUG -DREF_CHECK -DCONF_DEBUG -DCRYPTO_MDEBUG_ALL -xtarget=ultra -xarch=v8plus -g -O -xstrconst -Xa -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK_LL DES_PTR DES_RISC1 DES_UNROLL BF_PTR::sparcv8plus.o::::md5-sparcv8plus.o::::::dlfcn:solaris-shared:-KPIC:-G -dy -z text:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)", 
+"debug-solaris-sparcv9-cc","cc:-DBN_DEBUG -DREF_CHECK -DCONF_DEBUG -DCRYPTO_MDEBUG_ALL -xtarget=ultra -xarch=v8plus -g -O -xstrconst -Xa -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT:ULTRASPARC:-lsocket -lnsl -ldl:BN_LLONG RC4_CHAR RC4_CHUNK_LL DES_PTR DES_RISC1 DES_UNROLL BF_PTR:${no_asm}:dlfcn:solaris-shared:-KPIC:-G -dy -z text:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)", 
 
 #### SunOS configs, assuming sparc for the gcc one.
 #"sunos-cc", "cc:-O4 -DNOPROTO -DNOCONST::(unknown):SUNOS::DES_UNROLL:${no_asm}::",
@@ -232,11 +234,11 @@ my %table=(
 #### IRIX 6.x configs
 # Only N32 and N64 ABIs are supported. If you need O32 ABI build, invoke
 # './Configure irix-cc -o32' manually.
-"irix-mips3-gcc","gcc:-mabi=n32 -O3 -DTERMIOS -DB_ENDIAN -DBN_DIV3W::-D_SGI_MP_SOURCE:::MD2_CHAR RC4_INDEX RC4_CHAR RC4_CHUNK_LL DES_UNROLL DES_RISC2 DES_PTR BF_PTR SIXTY_FOUR_BIT::bn-mips3.o::::::::::dlfcn:irix-shared::-mabi=n32:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"irix-mips3-gcc","gcc:-mabi=n32 -O3 -DTERMIOS -DB_ENDIAN -DBN_DIV3W::-D_SGI_MP_SOURCE:::MD2_CHAR RC4_INDEX RC4_CHAR RC4_CHUNK_LL DES_UNROLL DES_RISC2 DES_PTR BF_PTR SIXTY_FOUR_BIT:${no_asm}:dlfcn:irix-shared::-mabi=n32:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"irix-mips3-cc", "cc:-n32 -mips3 -O2 -use_readonly_const -DTERMIOS -DB_ENDIAN -DBN_DIV3W::-D_SGI_MP_SOURCE:::DES_PTR RC4_CHAR RC4_CHUNK_LL DES_RISC2 DES_UNROLL BF_PTR SIXTY_FOUR_BIT::bn-mips3.o::::::::::dlfcn:irix-shared::-n32:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"irix-mips3-cc", "cc:-n32 -mips3 -O2 -use_readonly_const -G0 -rdata_shared -DTERMIOS -DB_ENDIAN -DBN_DIV3W::-D_SGI_MP_SOURCE:::DES_PTR RC4_CHAR RC4_CHUNK_LL DES_RISC2 DES_UNROLL BF_PTR SIXTY_FOUR_BIT:${no_asm}:dlfcn:irix-shared::-n32:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 # N64 ABI builds.
-"irix64-mips4-gcc","gcc:-mabi=64 -mips4 -O3 -DTERMIOS -DB_ENDIAN -DBN_DIV3W::-D_SGI_MP_SOURCE:::RC4_CHAR RC4_CHUNK DES_RISC2 DES_UNROLL SIXTY_FOUR_BIT_LONG::bn-mips3.o::::::::::dlfcn:irix-shared::-mabi=64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"irix64-mips4-gcc","gcc:-mabi=64 -mips4 -O3 -DTERMIOS -DB_ENDIAN -DBN_DIV3W::-D_SGI_MP_SOURCE:::RC4_CHAR RC4_CHUNK DES_RISC2 DES_UNROLL SIXTY_FOUR_BIT_LONG:${no_asm}:dlfcn:irix-shared::-mabi=64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"irix64-mips4-cc", "cc:-64 -mips4 -O2 -use_readonly_const -DTERMIOS -DB_ENDIAN -DBN_DIV3W::-D_SGI_MP_SOURCE:::RC4_CHAR RC4_CHUNK DES_RISC2 DES_UNROLL SIXTY_FOUR_BIT_LONG::bn-mips3.o::::::::::dlfcn:irix-shared::-64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"irix64-mips4-cc", "cc:-64 -mips4 -O2 -use_readonly_const -G0 -rdata_shared -DTERMIOS -DB_ENDIAN -DBN_DIV3W::-D_SGI_MP_SOURCE:::RC4_CHAR RC4_CHUNK DES_RISC2 DES_UNROLL SIXTY_FOUR_BIT_LONG:${no_asm}:dlfcn:irix-shared::-64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 
 #### Unified HP-UX ANSI C configs.
 # Special notes:
@@ -269,8 +271,8 @@ my %table=(
 # Since there is mention of this in shlib/hpux10-cc.sh
 "hpux-parisc-cc-o4","cc:-Ae +O4 +ESlit -z -DB_ENDIAN -DBN_DIV2W -DMD32_XARRAY::-D_REENTRANT::-ldld:BN_LLONG DES_PTR DES_UNROLL DES_RISC1:${no_asm}:dl:hpux-shared:+Z:-b:.sl.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "hpux-parisc-gcc","gcc:-O3 -DB_ENDIAN -DBN_DIV2W::-D_REENTRANT::-Wl,+s -ldld:BN_LLONG DES_PTR DES_UNROLL DES_RISC1:${no_asm}:dl:hpux-shared:-fPIC:-shared:.sl.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"hpux-parisc2-gcc","gcc:-march=2.0 -O3 -DB_ENDIAN -D_REENTRANT::::-Wl,+s -ldld:SIXTY_FOUR_BIT RC4_CHAR RC4_CHUNK DES_PTR DES_UNROLL DES_RISC1::pa-risc2.o::::::::::dl:hpux-shared:-fPIC:-shared:.sl.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"hpux-parisc2-gcc","gcc:-march=2.0 -O3 -DB_ENDIAN -D_REENTRANT::::-Wl,+s -ldld:SIXTY_FOUR_BIT RC4_CHAR RC4_CHUNK DES_PTR DES_UNROLL DES_RISC1:${no_asm}:dl:hpux-shared:-fPIC:-shared:.sl.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"hpux64-parisc2-gcc","gcc:-O3 -DB_ENDIAN -D_REENTRANT::::-ldl:SIXTY_FOUR_BIT_LONG MD2_CHAR RC4_INDEX RC4_CHAR DES_UNROLL DES_RISC1 DES_INT::pa-risc2W.o::::::::::dlfcn:hpux-shared:-fpic:-shared:.sl.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"hpux64-parisc2-gcc","gcc:-O3 -DB_ENDIAN -D_REENTRANT::::-ldl:SIXTY_FOUR_BIT_LONG MD2_CHAR RC4_INDEX RC4_CHAR DES_UNROLL DES_RISC1 DES_INT:${no_asm}:dlfcn:hpux-shared:-fpic:-shared:.sl.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 
 # More attempts at unified 10.X and 11.X targets for HP C compiler.
 #
@@ -278,8 +280,8 @@ my %table=(
 # Kevin Steves <ks@hp.se>
 "hpux-parisc-cc","cc:+O3 +Optrs_strongly_typed -Ae +ESlit -DB_ENDIAN -DBN_DIV2W -DMD32_XARRAY::-D_REENTRANT::-Wl,+s -ldld:MD2_CHAR RC4_INDEX RC4_CHAR DES_UNROLL DES_RISC1 DES_INT:${no_asm}:dl:hpux-shared:+Z:-b:.sl.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "hpux-parisc1_0-cc","cc:+DAportable +O3 +Optrs_strongly_typed -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY::-D_REENTRANT::-Wl,+s -ldld:MD2_CHAR RC4_INDEX RC4_CHAR DES_UNROLL DES_RISC1 DES_INT:${no_asm}:dl:hpux-shared:+Z:-b:.sl.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"hpux-parisc2-cc","cc:+DA2.0 +DS2.0 +O3 +Optrs_strongly_typed -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY -D_REENTRANT::::-Wl,+s -ldld:SIXTY_FOUR_BIT MD2_CHAR RC4_INDEX RC4_CHAR DES_UNROLL DES_RISC1 DES_INT::pa-risc2.o::::::::::dl:hpux-shared:+Z:-b:.sl.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"hpux-parisc2-cc","cc:+DA2.0 +DS2.0 +O3 +Optrs_strongly_typed -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY -D_REENTRANT::::-Wl,+s -ldld:SIXTY_FOUR_BIT MD2_CHAR RC4_INDEX RC4_CHAR DES_UNROLL DES_RISC1 DES_INT:${no_asm}:dl:hpux-shared:+Z:-b:.sl.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"hpux64-parisc2-cc","cc:+DD64 +O3 +Optrs_strongly_typed -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY -D_REENTRANT::::-ldl:SIXTY_FOUR_BIT_LONG MD2_CHAR RC4_INDEX RC4_CHAR DES_UNROLL DES_RISC1 DES_INT::pa-risc2W.o::::::::::dlfcn:hpux-shared:+Z:+DD64 -b:.sl.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"hpux64-parisc2-cc","cc:+DD64 +O3 +Optrs_strongly_typed -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY -D_REENTRANT::::-ldl:SIXTY_FOUR_BIT_LONG MD2_CHAR RC4_INDEX RC4_CHAR DES_UNROLL DES_RISC1 DES_INT:${no_asm}:dlfcn:hpux-shared:+Z:+DD64 -b:.sl.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 
 # HP/UX IA-64 targets
 "hpux-ia64-cc","cc:-Ae +DD32 +O2 +Olit=all -z -DB_ENDIAN -D_REENTRANT::::-ldl:SIXTY_FOUR_BIT MD2_CHAR RC4_INDEX DES_UNROLL DES_RISC1 DES_INT:${ia64_asm}:dlfcn:hpux-shared:+Z:+DD32 -b:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
@@ -316,27 +318,28 @@ my %table=(
 # *-generic* is endian-neutral target, but ./config is free to
 # throw in -D[BL]_ENDIAN, whichever appropriate...
 "linux-generic32","gcc:-DTERMIO -O3 -fomit-frame-pointer -Wall::-D_REENTRANT::-ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_INT DES_UNROLL BF_PTR:${no_asm}:dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"linux-ppc",	"gcc:-DB_ENDIAN -DTERMIO -O3 -Wall::-D_REENTRANT::-ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_RISC1 DES_UNROLL::linux_ppc32.o::::::::::dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"linux-ppc",	"gcc:-DB_ENDIAN -DTERMIO -O3 -Wall::-D_REENTRANT::-ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_RISC1 DES_UNROLL:${no_asm}:dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 #### IA-32 targets...
 "linux-ia32-icc",	"icc:-DL_ENDIAN -DTERMIO -O2 -no_cpprt::-D_REENTRANT::-ldl:BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}:${x86_elf_asm}:dlfcn:linux-shared:-KPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "linux-elf",	"gcc:-DL_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -Wall::-D_REENTRANT::-ldl:BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}:${x86_elf_asm}:dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "linux-aout",	"gcc:-DL_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -march=i486 -Wall::(unknown):::BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}:${x86_out_asm}",
 ####
 "linux-generic64","gcc:-DTERMIO -O3 -Wall::-D_REENTRANT::-ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_INT DES_UNROLL BF_PTR:${no_asm}:dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"linux-ppc64",	"gcc:-m64 -DB_ENDIAN -DTERMIO -O3 -Wall::-D_REENTRANT::-ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_RISC1 DES_UNROLL::linux_ppc64.o::::::::::dlfcn:linux-shared:-fPIC:-m64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"linux-ppc64",	"gcc:-m64 -DB_ENDIAN -DTERMIO -O3 -Wall::-D_REENTRANT::-ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_RISC1 DES_UNROLL:${no_asm}:dlfcn:linux-shared:-fPIC:-m64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "linux-ia64",	"gcc:-DL_ENDIAN -DTERMIO -O3 -Wall::-D_REENTRANT::-ldl:SIXTY_FOUR_BIT_LONG RC4_CHUNK:${ia64_asm}:dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "linux-ia64-ecc","ecc:-DL_ENDIAN -DTERMIO -O2 -Wall -no_cpprt::-D_REENTRANT::-ldl:SIXTY_FOUR_BIT_LONG RC4_CHUNK:${ia64_asm}:dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "linux-ia64-icc","icc:-DL_ENDIAN -DTERMIO -O2 -Wall -no_cpprt::-D_REENTRANT::-ldl:SIXTY_FOUR_BIT_LONG RC4_CHUNK:${ia64_asm}:dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "linux-x86_64",	"gcc:-m64 -DL_ENDIAN -DTERMIO -O3 -Wall -DMD32_REG_T=int::-D_REENTRANT::-ldl:SIXTY_FOUR_BIT_LONG RC4_CHUNK BF_PTR2 DES_INT DES_UNROLL:${x86_64_asm}:dlfcn:linux-shared:-fPIC:-m64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"linux-s390x",  "gcc:-m64 -DB_ENDIAN -DTERMIO -O3 -Wall::-D_REENTRANT::-ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_INT DES_UNROLL:${s390x_asm}:dlfcn:linux-shared:-fPIC:-m64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 #### SPARC Linux setups
 # Ray Miller <ray.miller@computing-services.oxford.ac.uk> has patiently
 # assisted with debugging of following two configs.
-"linux-sparcv8","gcc:-mv8 -DB_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -Wall -DBN_DIV2W::-D_REENTRANT::-ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR::sparcv8.o:des_enc-sparc.o fcrypt_b.o:::::::::dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"linux-sparcv8","gcc:-mv8 -DB_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -Wall -DBN_DIV2W::-D_REENTRANT::-ldl:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:${no_asm}:dlfcn:linux-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 # it's a real mess with -mcpu=ultrasparc option under Linux, but
 # -Wa,-Av8plus should do the trick no matter what.
-"linux-sparcv9","gcc:-m32 -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::sparcv8plus.o:des_enc-sparc.o fcrypt_b.o:::md5-sparcv8plus.o::::::dlfcn:linux-shared:-fPIC:-m32:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"linux-sparcv9","gcc:-m32 -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:${no_asm}:dlfcn:linux-shared:-fPIC:-m32:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 # GCC 3.1 is a requirement
-"linux64-sparcv9","gcc:-m64 -mcpu=ultrasparc -DB_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -Wall::-D_REENTRANT:ULTRASPARC:-ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:::des_enc-sparc.o fcrypt_b.o:::md5-sparcv9.o::::::dlfcn:linux-shared:-fPIC:-m64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"linux64-sparcv9","gcc:-m64 -mcpu=ultrasparc -DB_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -Wall::-D_REENTRANT:ULTRASPARC:-ldl:SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:${no_asm}:dlfcn:linux-shared:-fPIC:-m64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 #### Alpha Linux with GNU C and Compaq C setups
 # Special notes:
 # - linux-alpha+bwx-gcc is ment to be used from ./config only. If you
@@ -360,13 +363,13 @@ my %table=(
 "BSD-x86",	"gcc:-DL_ENDIAN -DTERMIOS -O3 -fomit-frame-pointer -Wall::${BSDthreads}:::BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}:${x86_out_asm}:dlfcn:bsd-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "BSD-x86-elf",	"gcc:-DL_ENDIAN -DTERMIOS -O3 -fomit-frame-pointer -Wall::${BSDthreads}:::BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}:${x86_elf_asm}:dlfcn:bsd-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "debug-BSD-x86-elf",	"gcc:-DL_ENDIAN -DTERMIOS -O3 -Wall -g::${BSDthreads}:::BN_LLONG ${x86_gcc_des} ${x86_gcc_opts}:${x86_elf_asm}:dlfcn:bsd-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
-"BSD-sparcv8",	"gcc:-DB_ENDIAN -DTERMIOS -O3 -mv8 -Wall::${BSDthreads}:::BN_LLONG RC2_CHAR RC4_INDEX DES_INT DES_UNROLL::sparcv8.o:des_enc-sparc.o fcrypt_b.o:::::::::dlfcn:bsd-gcc-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"BSD-sparcv8",	"gcc:-DB_ENDIAN -DTERMIOS -O3 -mv8 -Wall::${BSDthreads}:::BN_LLONG RC2_CHAR RC4_INDEX DES_INT DES_UNROLL:${no_asm}:dlfcn:bsd-gcc-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 
 "BSD-generic64","gcc:-DTERMIOS -O3 -Wall::${BSDthreads}:::SIXTY_FOUR_BIT_LONG RC4_CHUNK DES_INT DES_UNROLL:${no_asm}:dlfcn:bsd-gcc-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 # -DMD32_REG_T=int doesn't actually belong in sparc64 target, it
 # simply *happens* to work around a compiler bug in gcc 3.3.3,
 # triggered by RIPEMD160 code.
-"BSD-sparc64",	"gcc:-DB_ENDIAN -DTERMIOS -O3 -DMD32_REG_T=int -Wall::${BSDthreads}:::SIXTY_FOUR_BIT_LONG RC2_CHAR RC4_CHUNK DES_INT DES_PTR DES_RISC2 BF_PTR:::des_enc-sparc.o fcrypt_b.o:::md5-sparcv9.o::::::dlfcn:bsd-gcc-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
+"BSD-sparc64",	"gcc:-DB_ENDIAN -DTERMIOS -O3 -DMD32_REG_T=int -Wall::${BSDthreads}:::SIXTY_FOUR_BIT_LONG RC2_CHAR RC4_CHUNK DES_INT DES_PTR DES_RISC2 BF_PTR:${no_asm}:dlfcn:bsd-gcc-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "BSD-ia64",	"gcc:-DL_ENDIAN -DTERMIOS -O3 -Wall::${BSDthreads}:::SIXTY_FOUR_BIT_LONG RC4_CHUNK:${ia64_asm}:dlfcn:bsd-gcc-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 "BSD-x86_64",	"gcc:-DL_ENDIAN -DTERMIOS -O3 -DMD32_REG_T=int -Wall::${BSDthreads}:::SIXTY_FOUR_BIT_LONG RC4_CHUNK DES_INT DES_UNROLL:${x86_64_asm}:dlfcn:bsd-gcc-shared:-fPIC::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)",
 
@@ -404,12 +407,12 @@ my %table=(
 
 #### IBM's AIX.
 "aix3-cc",  "cc:-O -DB_ENDIAN -qmaxmem=16384::(unknown):AIX::BN_LLONG RC4_CHAR:::",
-"aix-gcc",  "gcc:-O -DB_ENDIAN::-D_THREAD_SAFE:AIX::BN_LLONG RC4_CHAR::aix_ppc32.o::::::::::dlfcn:aix-shared:::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)::-X 32",
+"aix-gcc",  "gcc:-O -DB_ENDIAN::-D_THREAD_SAFE:AIX::BN_LLONG RC4_CHAR:${no_asm}:dlfcn:aix-shared:::.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)::-X 32",
-"aix64-gcc","gcc:-maix64 -O -DB_ENDIAN::-D_THREAD_SAFE:AIX::SIXTY_FOUR_BIT_LONG RC4_CHAR::aix_ppc64.o::::::::::dlfcn:aix-shared::-maix64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)::-X64",
+"aix64-gcc","gcc:-maix64 -O -DB_ENDIAN::-D_THREAD_SAFE:AIX::SIXTY_FOUR_BIT_LONG RC4_CHAR:${no_asm}:dlfcn:aix-shared::-maix64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)::-X64",
 # Below targets assume AIX 5. Idea is to effectively disregard $OBJECT_MODE
 # at build time. $OBJECT_MODE is respected at ./config stage!
-"aix-cc",   "cc:-q32 -O -DB_ENDIAN -qmaxmem=16384 -qro -qroconst::-qthreaded:AIX::BN_LLONG RC4_CHAR::aix_ppc32.o::::::::::dlfcn:aix-shared::-q32:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)::-X 32",
+"aix-cc",   "cc:-q32 -O -DB_ENDIAN -qmaxmem=16384 -qro -qroconst::-qthreaded:AIX::BN_LLONG RC4_CHAR:${no_asm}:dlfcn:aix-shared::-q32:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)::-X 32",
-"aix64-cc", "cc:-q64 -O -DB_ENDIAN -qmaxmem=16384 -qro -qroconst::-qthreaded:AIX::SIXTY_FOUR_BIT_LONG RC4_CHAR::aix_ppc64.o::::::::::dlfcn:aix-shared::-q64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)::-X 64",
+"aix64-cc", "cc:-q64 -O -DB_ENDIAN -qmaxmem=16384 -qro -qroconst::-qthreaded:AIX::SIXTY_FOUR_BIT_LONG RC4_CHAR:${no_asm}:dlfcn:aix-shared::-q64:.so.\$(SHLIB_MAJOR).\$(SHLIB_MINOR)::-X 64",
 
 #
 # Cray T90 and similar (SDSC)
@@ -501,9 +504,9 @@ my %table=(
 
 ##### MacOS X (a.k.a. Rhapsody or Darwin) setup
 "rhapsody-ppc-cc","cc:-O3 -DB_ENDIAN::(unknown):MACOSX_RHAPSODY::BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:${no_asm}::",
-"darwin-ppc-cc","cc:-O3 -DB_ENDIAN::-D_REENTRANT:MACOSX:-Wl,-search_paths_first:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR::osx_ppc32.o::::::::::dlfcn:darwin-shared:-fPIC -fno-common:-dynamiclib:.\$(SHLIB_MAJOR).\$(SHLIB_MINOR).dylib",
+"darwin-ppc-cc","cc:-O3 -DB_ENDIAN::-D_REENTRANT:MACOSX:-Wl,-search_paths_first%:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:${no_asm}:dlfcn:darwin-shared:-fPIC -fno-common:-dynamiclib:.\$(SHLIB_MAJOR).\$(SHLIB_MINOR).dylib",
-"darwin-i386-cc","cc:-O3 -fomit-frame-pointer -fno-common::-D_REENTRANT:MACOSX::BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:${no_asm}:dlfcn:darwin-shared:-fPIC -fno-common:-dynamiclib:.\$(SHLIB_MAJOR).\$(SHLIB_MINOR).dylib",
+"darwin-i386-cc","cc:-O3 -fomit-frame-pointer -fno-common::-D_REENTRANT:MACOSX:-Wl,-search_paths_first%:BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:${no_asm}:dlfcn:darwin-shared:-fPIC -fno-common:-dynamiclib:.\$(SHLIB_MAJOR).\$(SHLIB_MINOR).dylib",
-"debug-darwin-ppc-cc","cc:-DBN_DEBUG -DREF_CHECK -DCONF_DEBUG -DCRYPTO_MDEBUG -DB_ENDIAN -g -Wall -O::-D_REENTRANT:MACOSX::BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR::osx_ppc32.o::::::::::dlfcn:darwin-shared:-fPIC -fno-common:-dynamiclib:.\$(SHLIB_MAJOR).\$(SHLIB_MINOR).dylib",
+"debug-darwin-ppc-cc","cc:-DBN_DEBUG -DREF_CHECK -DCONF_DEBUG -DCRYPTO_MDEBUG -DB_ENDIAN -g -Wall -O::-D_REENTRANT:MACOSX::BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:${no_asm}:dlfcn:darwin-shared:-fPIC -fno-common:-dynamiclib:.\$(SHLIB_MAJOR).\$(SHLIB_MINOR).dylib",
 
 ##### A/UX
 "aux3-gcc","gcc:-O2 -DTERMIO::(unknown):AUX:-lbsd:RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR:::",
@@ -563,8 +566,9 @@ my $idx_arflags = $idx++;
 my $prefix="";
 my $openssldir="";
 my $exe_ext="";
-my $install_prefix="";
+my $install_prefix="$ENV{'INSTALL_PREFIX'}";
-my $fipslibdir="/usr/local/ssl/lib/";
+my $cross_compile_prefix="$ENV{'CROSS_COMPILE'}";
+my $fipslibdir="/usr/local/ssl/lib/fips-1.0/";
 my $nofipscanistercheck=0;
 my $fipsdso=0;
 my $fipscanisterinternal="n";
@@ -792,6 +796,14 @@ PROCESS_ARGS:
 				{
 				$withargs{"zlib-include"}="-I$1";
 				}
+			elsif (/^--with-fipslibdir=(.*)$/)
+				{
+				$fipslibdir="$1/";
+				}
+			elsif (/^--with-baseaddr=(.*)$/)
+				{
+				$baseaddr="$1";
+				}
 			else
 				{
 				print STDERR $usage;
@@ -895,6 +907,41 @@ print "Configuring for $target\n";
 
 &usage if (!defined($table{$target}));
 
+my @fields = split(/\s*:\s*/,$table{$target} . ":" x 30 , -1);
+my $cc = $fields[$idx_cc];
+my $cflags = $fields[$idx_cflags];
+my $unistd = $fields[$idx_unistd];
+my $thread_cflag = $fields[$idx_thread_cflag];
+my $sys_id = $fields[$idx_sys_id];
+my $lflags = $fields[$idx_lflags];
+my $bn_ops = $fields[$idx_bn_ops];
+my $cpuid_obj = $fields[$idx_cpuid_obj];
+my $bn_obj = $fields[$idx_bn_obj];
+my $des_obj = $fields[$idx_des_obj];
+my $aes_obj = $fields[$idx_aes_obj];
+my $bf_obj = $fields[$idx_bf_obj];
+my $md5_obj = $fields[$idx_md5_obj];
+my $sha1_obj = $fields[$idx_sha1_obj];
+my $cast_obj = $fields[$idx_cast_obj];
+my $rc4_obj = $fields[$idx_rc4_obj];
+my $rmd160_obj = $fields[$idx_rmd160_obj];
+my $rc5_obj = $fields[$idx_rc5_obj];
+my $dso_scheme = $fields[$idx_dso_scheme];
+my $shared_target = $fields[$idx_shared_target];
+my $shared_cflag = $fields[$idx_shared_cflag];
+my $shared_ldflag = $fields[$idx_shared_ldflag];
+my $shared_extension = $fields[$idx_shared_extension];
+my $ranlib = $fields[$idx_ranlib];
+my $arflags = $fields[$idx_arflags];
+
+if ($fips)
+	{
+	delete $disabled{"shared"} if ($disabled{"shared"} eq "default");
+	$disabled{"asm"}="forced"
+		if ($target !~ "VC\-.*" &&
+		    "$cpuid_obj:$bn_obj:$aes_obj:$des_obj:$sha1_obj" eq "::::");
+	}
+
 
 foreach (sort (keys %disabled))
 	{
@@ -960,7 +1007,17 @@ $no_shared = 0 if ($fipsdso && !$IsMK1MF);
 
 $exe_ext=".exe" if ($target eq "Cygwin" || $target eq "DJGPP" || $target eq "mingw");
 $exe_ext=".pm"  if ($target =~ /vos/);
-$openssldir="/usr/local/ssl" if ($openssldir eq "" and $prefix eq "");
+if ($openssldir eq "" and $prefix eq "")
+	{
+	if ($fips)
+		{
+		$openssldir="/usr/local/ssl/fips-1.0";
+		}
+	else
+		{
+		$openssldir="/usr/local/ssl";
+		}
+	}
 $prefix=$openssldir if $prefix eq "";
 
 $default_ranlib= &which("ranlib") or $default_ranlib="true";
@@ -976,32 +1033,10 @@ $openssldir=$prefix . "/" . $openssldir if $openssldir !~ /(^\/|^[a-zA-Z]:[\\\/]
 
 print "IsMK1MF=$IsMK1MF\n";
 
-my @fields = split(/\s*:\s*/,$table{$target} . ":" x 30 , -1);
+# '%' in $lflags is used to split flags to "pre-" and post-flags
-my $cc = $fields[$idx_cc];
+my ($prelflags,$postlflags)=split('%',$lflags);
-my $cflags = $fields[$idx_cflags];
+if (defined($postlflags))	{ $lflags=$postlflags;  }
-my $unistd = $fields[$idx_unistd];
+else				{ $lflags=$prelflags; undef $prelflags; }
-my $thread_cflag = $fields[$idx_thread_cflag];
-my $sys_id = $fields[$idx_sys_id];
-my $lflags = $fields[$idx_lflags];
-my $bn_ops = $fields[$idx_bn_ops];
-my $cpuid_obj = $fields[$idx_cpuid_obj];
-my $bn_obj = $fields[$idx_bn_obj];
-my $des_obj = $fields[$idx_des_obj];
-my $aes_obj = $fields[$idx_aes_obj];
-my $bf_obj = $fields[$idx_bf_obj];
-my $md5_obj = $fields[$idx_md5_obj];
-my $sha1_obj = $fields[$idx_sha1_obj];
-my $cast_obj = $fields[$idx_cast_obj];
-my $rc4_obj = $fields[$idx_rc4_obj];
-my $rmd160_obj = $fields[$idx_rmd160_obj];
-my $rc5_obj = $fields[$idx_rc5_obj];
-my $dso_scheme = $fields[$idx_dso_scheme];
-my $shared_target = $fields[$idx_shared_target];
-my $shared_cflag = $fields[$idx_shared_cflag];
-my $shared_ldflag = $fields[$idx_shared_ldflag];
-my $shared_extension = $fields[$idx_shared_extension];
-my $ranlib = $fields[$idx_ranlib];
-my $arflags = $fields[$idx_arflags];
 
 my $no_shared_warn=0;
 my $no_user_cflags=0;
@@ -1131,6 +1166,8 @@ if ($no_asm)
 	{
 	$cpuid_obj=$bn_obj=$des_obj=$aes_obj=$bf_obj=$cast_obj=$rc4_obj=$rc5_obj="";
 	$sha1_obj=$md5_obj=$rmd160_obj="";
+	$cflags=~s/\-D[BL]_ENDIAN//		if ($fips);
+	$thread_cflags=~s/\-D[BL]_ENDIAN//	if ($fips);
 	}
 
 if (!$no_shared)
@@ -1161,7 +1198,7 @@ if ($zlib)
 my $shared_mark = "";
 if ($shared_target eq "")
 	{
-	$no_shared_warn = 1 if !$no_shared;
+	$no_shared_warn = 1 if !$no_shared && !$fips;
 	$no_shared = 1;
 	}
 if (!$no_shared)
@@ -1185,12 +1222,19 @@ if (!$IsMK1MF)
 	}
 
 $cpuid_obj.=" uplink.o uplink-cof.o" if ($cflags =~ /\-DOPENSSL_USE_APPLINK/);
-# Compiler fix-ups
+
-if ($target =~ /icc$/)
+#
+# Platform fix-ups
+#
+if ($target =~ /\-icc$/)	# Intel C compiler
 	{
-	my($iccver)=`$cc -V 2>&1`;
+	my $iccver=0;
-	if ($iccver =~ /Version ([0-9]+)\./)	{ $iccver=$1; }
+	if (open(FD,"$cc -V 2>&1 |"))
-	else					{ $iccver=0;  }
+		{
+		while(<FD>) { $iccver=$1 if (/Version ([0-9]+)\./); }
+		close(FD);
+		}
+
 	if ($iccver>=8)
 		{
 		# Eliminate unnecessary dependency from libirc.a. This is
@@ -1198,6 +1242,28 @@ if ($target =~ /icc$/)
 		# apps/openssl can end up in endless loop upon startup...
 		$cflags.=" -Dmemcpy=__builtin_memcpy -Dmemset=__builtin_memset";
 		}
+	if ($iccver>=9)
+		{
+		$cflags.=" -i-static";
+		$cflags=~s/\-no_cpprt/-no-cpprt/;
+		}
+	if ($iccver>=10)
+		{
+		$cflags=~s/\-i\-static/-static-intel/;
+		}
+	}
+
+# Unlike other OSes (like Solaris, Linux, Tru64, IRIX) BSD run-time
+# linkers (tested OpenBSD, NetBSD and FreeBSD) "demand" RPATH set on
+# .so objects. Apparently application RPATH is not global and does
+# not apply to .so linked with other .so. Problem manifests itself
+# when libssl.so fails to load libcrypto.so. One can argue that we
+# should engrave this into Makefile.shared rules or into BSD-* config
+# lines above. Meanwhile let's try to be cautious and pass -rpath to
+# linker only when --prefix is not /usr.
+if ($target =~ /^BSD\-/)
+	{
+	$shared_ldflag.=" -Wl,-rpath,\$(LIBRPATH)" if ($prefix !~ m|^/usr[/]*$|);
 	}
 
 if ($sys_id ne "")
@@ -1223,24 +1289,10 @@ $bn_obj = $bn_asm unless $bn_obj ne "";
 $cflags.=" -DOPENSSL_BN_ASM_PART_WORDS" if ($bn_obj =~ /bn86/);
 $cflags.=" -DOPENSSL_IA32_SSE2" if (!$no_sse2 && $bn_obj =~ /bn86/);
 
-my $fips_des_obj;
+$cflags.=" -DOPENSSL_BN_ASM_MONT" if ($bn_obj =~ /\-mont|mo86\-/);
-my $fips_aes_obj;
+
-my $fips_sha1_obj;
 if ($fips)
 	{
-	if ($des_obj =~ /\-elf\.o$/)
-		{
-		$fips_des_obj='asm/fips-dx86-elf.o';
-		$openssl_other_defines.="#define OPENSSL_FIPS_DES_ASM\n";
-		$fips_aes_obj='asm/fips-ax86-elf.o';
-		$openssl_other_defines.="#define OPENSSL_FIPS_AES_ASM\n";
-		}
-	else	{
-		$fips_des_obj=$fips_des_enc;
-		$fips_aes_obj='fips_aes_core.o';
-		}
-	$fips_sha1_obj='asm/fips-sx86-elf.o' if ($sha1_obj =~ /\-elf\.o$/);
-	$des_obj=$sha1_obj=$aes_obj="";
 	$openssl_other_defines.="#define OPENSSL_FIPS\n";
 	}
 
@@ -1255,7 +1307,7 @@ if ($sha1_obj =~ /\.o$/)
 	$cflags.=" -DSHA1_ASM"   if ($sha1_obj =~ /sx86/ || $sha1_obj =~ /sha1/);
 	$cflags.=" -DSHA256_ASM" if ($sha1_obj =~ /sha256/);
 	$cflags.=" -DSHA512_ASM" if ($sha1_obj =~ /sha512/);
-	if ($sha1_obj =~ /x86/)
+	if ($sha1_obj =~ /sse2/)
 	    {	if ($no_sse2)
 		{   $sha1_obj =~ s/\S*sse2\S+//;        }
 		elsif ($cflags !~ /OPENSSL_IA32_SSE2/)
@@ -1346,16 +1398,24 @@ while (<IN>)
 	s/^PLATFORM=.*$/PLATFORM=$target/;
 	s/^OPTIONS=.*$/OPTIONS=$options/;
 	s/^CONFIGURE_ARGS=.*$/CONFIGURE_ARGS=$argvstring/;
+	if ($cross_compile_prefix)
+		{
+		s/^CC=.*$/CROSS_COMPILE= $cross_compile_prefix\nCC= \$\(CROSS_COMPILE\)$cc/;
+		s/^AR=\s*/AR= \$\(CROSS_COMPILE\)/;
+		s/^RANLIB=\s*/RANLIB= \$\(CROSS_COMPILE\)/;
+		}
+	else	{
 		s/^CC=.*$/CC= $cc/;
+		s/^RANLIB=.*/RANLIB= $ranlib/;
+		}
 	s/^MAKEDEPPROG=.*$/MAKEDEPPROG= $cc/ if $cc eq "gcc";
 	s/^CFLAG=.*$/CFLAG= $cflags/;
 	s/^DEPFLAG=.*$/DEPFLAG= $depflags/;
+	s/^PEX_LIBS=.*$/PEX_LIBS= $prelflags/;
 	s/^EX_LIBS=.*$/EX_LIBS= $lflags/;
 	s/^EXE_EXT=.*$/EXE_EXT= $exe_ext/;
 	s/^CPUID_OBJ=.*$/CPUID_OBJ= $cpuid_obj/;
 	s/^BN_ASM=.*$/BN_ASM= $bn_obj/;
-	s/^FIPS_DES_ENC=.*$/FIPS_DES_ENC= $fips_des_obj/;
-	s/^FIPS_AES_ENC=.*$/FIPS_AES_ENC= $fips_aes_obj/;
 	s/^DES_ENC=.*$/DES_ENC= $des_obj/;
 	s/^AES_ASM_OBJ=.*$/AES_ASM_OBJ= $aes_obj/;
 	s/^BF_ENC=.*$/BF_ENC= $bf_obj/;
@@ -1364,10 +1424,8 @@ while (<IN>)
 	s/^RC5_ENC=.*$/RC5_ENC= $rc5_obj/;
 	s/^MD5_ASM_OBJ=.*$/MD5_ASM_OBJ= $md5_obj/;
 	s/^SHA1_ASM_OBJ=.*$/SHA1_ASM_OBJ= $sha1_obj/;
-	s/^FIPS_SHA1_ASM_OBJ=.*$/FIPS_SHA1_ASM_OBJ= $fips_sha1_obj/;
 	s/^RMD160_ASM_OBJ=.*$/RMD160_ASM_OBJ= $rmd160_obj/;
 	s/^PROCESSOR=.*/PROCESSOR= $processor/;
-	s/^RANLIB=.*/RANLIB= $ranlib/;
 	s/^ARFLAGS=.*/ARFLAGS= $arflags/;
 	s/^PERL=.*/PERL= $perl/;
 	s/^KRB5_INCLUDES=.*/KRB5_INCLUDES=$withargs{"krb5-include"}/;
@@ -1391,7 +1449,7 @@ while (<IN>)
 	s/^BASEADDR=.*/BASEADDR=$baseaddr/;
 	s/^SHLIB_TARGET=.*/SHLIB_TARGET=$shared_target/;
 	s/^SHLIB_MARK=.*/SHLIB_MARK=$shared_mark/;
-	s/^SHARED_LIBS=.*/SHARED_LIBS=\$(SHARED_CRYPTO) \$(SHARED_SSL) \$(SHARED_FIPS)/ if (!$no_shared);
+	s/^SHARED_LIBS=.*/SHARED_LIBS=\$(SHARED_FIPS) \$(SHARED_CRYPTO) \$(SHARED_SSL)/ if (!$no_shared);
 	if ($shared_extension ne "" && $shared_extension =~ /^\.s([ol])\.[^\.]*$/)
 		{
 		my $sotmp = $1;

Makefile.org

@@ -87,8 +87,6 @@ PROCESSOR=
 # CPUID module collects small commonly used assembler snippets
 CPUID_OBJ= 
 BN_ASM= bn_asm.o
-FIPS_DES_ENC= des_enc.o fcrypt_b.o
-FIPS_AES_ENC= fips_aes_core.o
 DES_ENC= des_enc.o fcrypt_b.o
 AES_ASM_OBJ=aes_core.o aes_cbc.o
 BF_ENC= bf_enc.o
@@ -96,7 +94,6 @@ CAST_ENC= c_enc.o
 RC4_ENC= rc4_enc.o
 RC5_ENC= rc5_enc.o
 MD5_ASM_OBJ= 
-FIPS_SHA1_ASM_OBJ= 
 SHA1_ASM_OBJ= 
 RMD160_ASM_OBJ= 
 
@@ -134,7 +131,7 @@ FIPSCANLIB=
 
 BASEADDR=
 
-DIRS=   crypto fips-1.0 ssl engines apps test tools
+DIRS=   crypto fips ssl engines apps test tools
 SHLIBDIRS= crypto ssl fips
 
 # dirs in crypto to build
@@ -222,9 +219,6 @@ BUILDENV=	PLATFORM='${PLATFORM}' PROCESSOR='${PROCESSOR}' \
 		SHA1_ASM_OBJ='${SHA1_ASM_OBJ}'			\
 		MD5_ASM_OBJ='${MD5_ASM_OBJ}'			\
 		RMD160_ASM_OBJ='${RMD160_ASM_OBJ}'		\
-		FIPS_SHA1_ASM_OBJ='${FIPS_SHA1_ASM_OBJ}'	\
-		FIPS_DES_ENC='${FIPS_DES_ENC}'		 	\
-		FIPS_AES_ENC='${FIPS_AES_ENC}'		 	\
 		FIPSLIBDIR='${FIPSLIBDIR}' FIPSCANLIB='${FIPSCANLIB}' \
 		FIPSCANISTERINTERNAL='${FIPSCANISTERINTERNAL}'	\
 		FIPS_EX_OBJ='${FIPS_EX_OBJ}'	\
@@ -259,8 +253,7 @@ BUILD_ONE_CMD=\
 reflect:
 	@[ -n "$(THIS)" ] && $(CLEARENV) && $(MAKE) $(THIS) -e $(BUILDENV)
 
-FIPS_EX_OBJ= ../crypto/aes/aes_cbc.o \
+FIPS_EX_OBJ= ../crypto/aes/aes_cfb.o \
-	../crypto/aes/aes_cfb.o \
 	../crypto/aes/aes_ecb.o \
 	../crypto/aes/aes_ofb.o \
 	../crypto/bn/bn_add.o \
@@ -286,13 +279,12 @@ FIPS_EX_OBJ= ../crypto/aes/aes_cbc.o \
 	../crypto/des/cfb64ede.o \
 	../crypto/des/cfb64enc.o \
 	../crypto/des/cfb_enc.o \
-	../crypto/des/des_enc.o \
 	../crypto/des/ecb3_enc.o \
 	../crypto/des/ecb_enc.o \
 	../crypto/des/ofb64ede.o \
 	../crypto/des/ofb64enc.o \
-	../crypto/des/fcrypt_b.o \
 	../crypto/des/fcrypt.o \
+	../crypto/des/set_key.o \
 	../crypto/dsa/dsa_utl.o \
 	../crypto/dsa/dsa_sign.o \
 	../crypto/dsa/dsa_vrf.o \
@@ -319,23 +311,26 @@ FIPS_EX_OBJ= ../crypto/aes/aes_cbc.o \
 	../crypto/rsa/rsa_pss.o \
 	../crypto/rsa/rsa_ssl.o \
 	../crypto/rsa/rsa_x931.o \
+	../crypto/sha/sha1dgst.o \
+	../crypto/sha/sha256.o \
+	../crypto/sha/sha512.o \
 	../crypto/uid.o
 
 sub_all: build_all
 build_all: build_libs build_apps build_tests build_tools
 
-build_libs: build_crypto build_fips build_ssl build_engines
+build_libs: build_crypto build_fips build_ssl build_shared build_engines
 
 build_crypto:
 	if [ -n "$(FIPSCANLIB)" ]; then \
-		EXCL_OBJ='$(BN_ASM) $(CPUID_OBJ) $(FIPS_EX_OBJ)' ; export EXCL_OBJ ; \
+		EXCL_OBJ='$(AES_ASM_OBJ) $(BN_ASM) $(DES_ENC) $(CPUID_OBJ) $(SHA1_ASM_OBJ) $(FIPS_EX_OBJ)' ; export EXCL_OBJ ; \
 		ARX='$(PERL) $${TOP}/util/arx.pl $(AR)' ; \
 	else \
 		ARX='${AR}' ; \
 	fi ; export ARX ; \
 		dir=crypto; target=all; $(BUILD_ONE_CMD)
 build_fips:
-	@dir=fips-1.0; target=all; $(BUILD_ONE_CMD)
+	@dir=fips; target=all; [ -z "$(FIPSCANLIB)" ] || $(BUILD_ONE_CMD)
 build_ssl:
 	@dir=ssl; target=all; $(BUILD_ONE_CMD)
 build_engines:
@@ -351,42 +346,52 @@ all_testapps: build_libs build_testapps
 build_testapps:
 	@dir=crypto; target=testapps; $(BUILD_ONE_CMD)
 
+build_shared:	$(SHARED_LIBS)
 libcrypto$(SHLIB_EXT): libcrypto.a $(SHARED_FIPS)
 	@if [ "$(SHLIB_TARGET)" != "" ]; then \
 		if [ "$(FIPSCANLIB)" = "libfips" ]; then \
-			( dir=fips-1.0; target=all; $(BUILD_ONE_CMD) ) ; \
 			$(ARD) libcrypto.a fipscanister.o ; \
 			$(MAKE) SHLIBDIRS='crypto' SHLIBDEPS='-lfips' build-shared; \
-			$(AR) libcrypto.a fips-1.0/fipscanister.o ; \
+			$(AR) libcrypto.a fips/fipscanister.o ; \
 		else \
-			$(MAKE) SHLIBDIRS='crypto' build-shared; \
+			if [ "$(FIPSCANLIB)" = "libcrypto" ]; then \
+				FIPSLD_CC=$(CC); CC=fips/fipsld; \
+				export CC FIPSLD_CC; \
+			fi; \
+			$(MAKE) -e SHLIBDIRS='crypto' build-shared; \
 		fi \
 	else \
 		echo "There's no support for shared libraries on this platform" >&2; \
+		exit 1; \
 	fi
 
 libssl$(SHLIB_EXT): libcrypto$(SHLIB_EXT) libssl.a
 	@if [ "$(SHLIB_TARGET)" != "" ]; then \
-		$(MAKE) SHLIBDIRS=ssl SHLIBDEPS='-lcrypto' build-shared; \
+		shlibdeps=-lcrypto; \
+		[ "$(FIPSCANLIB)" = "libfips" ] && shlibdeps="$$shlibdeps -lfips"; \
+		$(MAKE) SHLIBDIRS=ssl SHLIBDEPS="$$shlibdeps" build-shared; \
 	else \
 		echo "There's no support for shared libraries on this platform" >&2 ; \
 		exit 1; \
 	fi
 
-libfips$(SHLIB_EXT): libfips.a
+fips/fipscanister.o:	build_fips
+libfips$(SHLIB_EXT):		fips/fipscanister.o
 	@if [ "$(SHLIB_TARGET)" != "" ]; then \
-		if [ "$(FIPSCANLIB)" = "libfips" ]; then \
+		FIPSLD_CC=$(CC); CC=fips/fipsld; export CC FIPSLD_CC; \
-			FIPSLD_CC=$(CC); CC=fips-1.0/fipsld; FIPSLD_NPT="y"; \
+		$(MAKE) -f Makefile.shared -e $(BUILDENV) \
-			FIPSLD_LIBFIPS=y; \
+			CC=$${CC} LIBNAME=fips THIS=$@ \
-			export CC FIPSLD_CC FIPSLD_NPT FIPSLD_LIBFIPS; \
+			LIBEXTRAS=fips/fipscanister.o \
-		fi; \
+			LIBDEPS="$(EX_LIBS)" \
-		$(MAKE) -e SHLIBDIRS=fips build-shared; \
+			LIBVERSION=${SHLIB_MAJOR}.${SHLIB_MINOR} \
+			link_o.$(SHLIB_TARGET) || { rm -f $@; exit 1; } \
 	else \
 		echo "There's no support for shared libraries on this platform" >&2; \
+		exit 1; \
 	fi
 
 libfips.a:
-	dir=fips-1.0; target=all; $(BUILD_ONE_CMD)
+	dir=fips; target=all; $(BUILD_ONE_CMD)
 
 clean-shared:
 	@set -e; for i in $(SHLIBDIRS); do \
@@ -507,12 +512,14 @@ dclean:
 	@set -e; target=dclean; $(RECURSIVE_BUILD_CMD)
 
 rehash: rehash.time
-rehash.time: certs
+rehash.time: certs apps
-	@(OPENSSL="`pwd`/util/opensslwrap.sh"; \
+	@if [ -z "$(CROSS_COMPILE)" ]; then \
+		(OPENSSL="`pwd`/util/opensslwrap.sh"; \
 		OPENSSL_DEBUG_MEMORY=on; \
 		export OPENSSL OPENSSL_DEBUG_MEMORY; \
-	  $(PERL) tools/c_rehash certs)
+		$(PERL) tools/c_rehash certs) && \
-	touch rehash.time
+		touch rehash.time; \
+	else :; fi
 
 test:   tests
 
@@ -601,7 +608,7 @@ dist:
 dist_pem_h:
 	(cd crypto/pem; $(MAKE) -e $(BUILDENV) pem.h; $(MAKE) clean)
 
-install: all install_docs install_sw
+install: all install_sw
 
 install_sw:
 	@$(PERL) $(TOP)/util/mkdir-p.pl $(INSTALL_PREFIX)$(INSTALLTOP)/bin \

Makefile.shared

@@ -101,15 +101,13 @@ LINK_SO=	\
     LIBDEPS="$${LIBDEPS:-$(LIBDEPS)}"; \
     SHAREDCMD="$${SHAREDCMD:-$(CC)}"; \
     SHAREDFLAGS="$${SHAREDFLAGS:-$(CFLAGS) $(SHARED_LDFLAGS)}"; \
-    nm -Pg $$SHOBJECTS | grep ' [BDT] ' | cut -f1 -d' ' > lib$(LIBNAME).exp; \
     LIBPATH=`for x in $$LIBDEPS; do if echo $$x | grep '^ *-L' > /dev/null 2>&1; then echo $$x | sed -e 's/^ *-L//'; fi; done | uniq`; \
     LIBPATH=`echo $$LIBPATH | sed -e 's/ /:/g'`; \
     LD_LIBRARY_PATH=$$LIBPATH:$$LD_LIBRARY_PATH \
     $${SHAREDCMD} $${SHAREDFLAGS} \
 	-o $$SHLIB$$SHLIB_SOVER$$SHLIB_SUFFIX \
 	$$ALLSYMSFLAGS $$SHOBJECTS $$NOALLSYMSFLAGS $$LIBDEPS \
-  ) && $(SYMLINK_SO); \
+  ) && $(SYMLINK_SO)
-  ( $(SET_X); rm -f lib$(LIBNAME).exp )
 
 SYMLINK_SO=	\
 	if [ -n "$$INHIBIT_SYMLINKS" ]; then :; else \
@@ -202,8 +200,10 @@ link_app.bsd:
 # to use native NSModule(3) API and refers to dlfcn as termporary hack.
 link_o.darwin:
 	@ $(CALC_VERSIONS); \
-	SHLIB=lib$(LIBNAME); \
+	SHLIB=`expr "$$THIS" : '.*/\([^/\.]*\)\.'`; \
-	SHLIB_SUFFIX=.so; \
+	SHLIB=$${SHLIB:-lib$(LIBNAME)}; \
+	SHLIB_SUFFIX=`expr "$$THIS" : '.*\(\.[^\.]*\)$$'`; \
+	SHLIB_SUFFIX=$${SHLIB_SUFFIX:-.so}; \
 	ALLSYMSFLAGS='-all_load'; \
 	NOALLSYMSFLAGS=''; \
 	SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS)"; \
@@ -284,7 +284,7 @@ link_o.alpha-osf1:
 		SHLIB_SOVER=; \
 		ALLSYMSFLAGS='-all'; \
 		NOALLSYMSFLAGS='-none'; \
-		SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared"; \
+		SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared -Wl,-B,symbolic"; \
 		if [ -n "$$SHLIB_HIST" ]; then \
 			SHAREDFLAGS="$$SHAREDFLAGS -set_version $$SHLIB_HIST"; \
 		fi; \
@@ -305,7 +305,7 @@ link_a.alpha-osf1:
 		SHLIB_SOVER=; \
 		ALLSYMSFLAGS='-all'; \
 		NOALLSYMSFLAGS='-none'; \
-		SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared"; \
+		SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared -Wl,-B,symbolic"; \
 		if [ -n "$$SHLIB_HIST" ]; then \
 			SHAREDFLAGS="$$SHAREDFLAGS -set_version $$SHLIB_HIST"; \
 		fi; \
@@ -428,7 +428,7 @@ link_o.irix:
 		($(CC) -v 2>&1 | grep gcc) > /dev/null && MINUSWL="-Wl,"; \
 		ALLSYMSFLAGS="$${MINUSWL}-all"; \
 		NOALLSYMSFLAGS="$${MINUSWL}-none"; \
-		SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared -Wl,-soname,$$SHLIB$$SHLIB_SOVER$$SHLIB_SUFFIX"; \
+		SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared -Wl,-soname,$$SHLIB$$SHLIB_SOVER$$SHLIB_SUFFIX,-B,symbolic"; \
 	fi; \
 	$(LINK_SO_O)
 link_a.irix:
@@ -442,7 +442,7 @@ link_a.irix:
 		($(CC) -v 2>&1 | grep gcc) > /dev/null && MINUSWL="-Wl,"; \
 		ALLSYMSFLAGS="$${MINUSWL}-all"; \
 		NOALLSYMSFLAGS="$${MINUSWL}-none"; \
-		SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared -Wl,-soname,$$SHLIB$$SHLIB_SOVER$$SHLIB_SUFFIX"; \
+		SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -shared -Wl,-soname,$$SHLIB$$SHLIB_SOVER$$SHLIB_SUFFIX,-B,symbolic"; \
 	fi; \
 	$(LINK_SO_A)
 link_app.irix:
@@ -466,7 +466,7 @@ link_o.hpux:
 	ALLSYMSFLAGS='-Wl,-Fl'; \
 	NOALLSYMSFLAGS=''; \
 	expr $(PLATFORM) : 'hpux64' > /dev/null && ALLSYMSFLAGS='-Wl,+forceload'; \
-	SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -Wl,-B,symbolic,+vnocompatwarnings,-z,+s,+h,$$SHLIB$$SHLIB_SOVER$$SHLIB_SUFFIX"; \
+	SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -Wl,-B,symbolic,+vnocompatwarnings,-z,+s,+h,$$SHLIB$$SHLIB_SOVER$$SHLIB_SUFFIX,+cdp,../:,+cdp,./:"; \
 	fi; \
 	rm -f $$SHLIB$$SHLIB_SOVER$$SHLIB_SUFFIX || :; \
 	$(LINK_SO_O) && chmod a=rx $$SHLIB$$SHLIB_SOVER$$SHLIB_SUFFIX
@@ -479,7 +479,7 @@ link_a.hpux:
 	ALLSYMSFLAGS='-Wl,-Fl'; \
 	NOALLSYMSFLAGS=''; \
 	expr $(PLATFORM) : 'hpux64' > /dev/null && ALLSYMSFLAGS='-Wl,+forceload'; \
-	SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -Wl,-B,symbolic,+vnocompatwarnings,-z,+s,+h,$$SHLIB$$SHLIB_SOVER$$SHLIB_SUFFIX"; \
+	SHAREDFLAGS="$(CFLAGS) $(SHARED_LDFLAGS) -Wl,-B,symbolic,+vnocompatwarnings,-z,+s,+h,$$SHLIB$$SHLIB_SOVER$$SHLIB_SUFFIX,+cdp,../:,+cdp,./:"; \
 	fi; \
 	rm -f $$SHLIB$$SHLIB_SOVER$$SHLIB_SUFFIX || :; \
 	$(LINK_SO_A) && chmod a=rx $$SHLIB$$SHLIB_SOVER$$SHLIB_SUFFIX

TABLE

@@ -117,14 +117,14 @@ $sys_id       =
 $lflags       = 
 $bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHUNK
 $cpuid_obj    = 
-$bn_obj       = bn-ia64.o
+$bn_obj       = 
 $des_obj      = 
-$aes_obj      = aes_core.o aes_cbc.o aes-ia64.o
+$aes_obj      = 
 $bf_obj       = 
 $md5_obj      = 
-$sha1_obj     = sha1-ia64.o sha256-ia64.o sha512-ia64.o
+$sha1_obj     = 
 $cast_obj     = 
-$rc4_obj      = rc4-ia64.o
+$rc4_obj      = 
 $rmd160_obj   = 
 $rc5_obj      = 
 $dso_scheme   = dlfcn
@@ -145,10 +145,10 @@ $lflags       =
 $bn_ops       = SIXTY_FOUR_BIT_LONG RC2_CHAR RC4_CHUNK DES_INT DES_PTR DES_RISC2 BF_PTR
 $cpuid_obj    = 
 $bn_obj       = 
-$des_obj      = des_enc-sparc.o fcrypt_b.o
+$des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
-$md5_obj      = md5-sparcv9.o
+$md5_obj      = 
 $sha1_obj     = 
 $cast_obj     = 
 $rc4_obj      = 
@@ -171,8 +171,8 @@ $sys_id       =
 $lflags       = 
 $bn_ops       = BN_LLONG RC2_CHAR RC4_INDEX DES_INT DES_UNROLL
 $cpuid_obj    = 
-$bn_obj       = sparcv8.o
+$bn_obj       = 
-$des_obj      = des_enc-sparc.o fcrypt_b.o
+$des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
 $md5_obj      = 
@@ -198,7 +198,7 @@ $sys_id       =
 $lflags       = 
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-out.o
-$bn_obj       = bn86-out.o co86-out.o
+$bn_obj       = bn86-out.o co86-out.o mo86-out.o
 $des_obj      = dx86-out.o yx86-out.o
 $aes_obj      = ax86-out.o
 $bf_obj       = bx86-out.o
@@ -225,7 +225,7 @@ $sys_id       =
 $lflags       = 
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -252,12 +252,12 @@ $sys_id       =
 $lflags       = 
 $bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHUNK DES_INT DES_UNROLL
 $cpuid_obj    = x86_64cpuid.o
-$bn_obj       = x86_64-gcc.o
+$bn_obj       = x86_64-gcc.o x86_64-mont.o
 $des_obj      = 
-$aes_obj      = 
+$aes_obj      = aes-x86_64.o
 $bf_obj       = 
 $md5_obj      = md5-x86_64.o
-$sha1_obj     = 
+$sha1_obj     = sha1-x86_64.o sha256-x86_64.o sha512-x86_64.o
 $cast_obj     = 
 $rc4_obj      = rc4-x86_64.o
 $rmd160_obj   = 
@@ -279,7 +279,7 @@ $sys_id       = CYGWIN32
 $lflags       = 
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-cof.o
-$bn_obj       = bn86-cof.o co86-cof.o
+$bn_obj       = bn86-cof.o co86-cof.o mo86-cof.o
 $des_obj      = dx86-cof.o yx86-cof.o
 $aes_obj      = ax86-cof.o
 $bf_obj       = bx86-cof.o
@@ -333,7 +333,7 @@ $sys_id       = MSDOS
 $lflags       = -L/dev/env/WATT_ROOT/lib -lwatt
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-out.o
-$bn_obj       = bn86-out.o co86-out.o
+$bn_obj       = bn86-out.o co86-out.o mo86-out.o
 $des_obj      = dx86-out.o yx86-out.o
 $aes_obj      = ax86-out.o
 $bf_obj       = bx86-out.o
@@ -684,7 +684,7 @@ $sys_id       = AIX
 $lflags       = 
 $bn_ops       = BN_LLONG RC4_CHAR
 $cpuid_obj    = 
-$bn_obj       = aix_ppc32.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -711,7 +711,7 @@ $sys_id       = AIX
 $lflags       = 
 $bn_ops       = BN_LLONG RC4_CHAR
 $cpuid_obj    = 
-$bn_obj       = aix_ppc32.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -765,7 +765,7 @@ $sys_id       = AIX
 $lflags       = 
 $bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHAR
 $cpuid_obj    = 
-$bn_obj       = aix_ppc64.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -792,7 +792,7 @@ $sys_id       = AIX
 $lflags       = 
 $bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHAR
 $cpuid_obj    = 
-$bn_obj       = aix_ppc64.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -846,7 +846,7 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -951,7 +951,7 @@ $cflags       = -O3 -fomit-frame-pointer -fno-common
 $unistd       = 
 $thread_cflag = -D_REENTRANT
 $sys_id       = MACOSX
-$lflags       = 
+$lflags       = -Wl,-search_paths_first%
 $bn_ops       = BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR
 $cpuid_obj    = 
 $bn_obj       = 
@@ -978,10 +978,10 @@ $cflags       = -O3 -DB_ENDIAN
 $unistd       = 
 $thread_cflag = -D_REENTRANT
 $sys_id       = MACOSX
-$lflags       = -Wl,-search_paths_first
+$lflags       = -Wl,-search_paths_first%
 $bn_ops       = BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR
 $cpuid_obj    = 
-$bn_obj       = osx_ppc32.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -1035,7 +1035,7 @@ $sys_id       =
 $lflags       = 
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -1224,7 +1224,7 @@ $sys_id       =
 $lflags       = 
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -1251,7 +1251,7 @@ $sys_id       = MACOSX
 $lflags       = 
 $bn_ops       = BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR
 $cpuid_obj    = 
-$bn_obj       = osx_ppc32.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -1305,7 +1305,7 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -1332,7 +1332,7 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -1413,7 +1413,7 @@ $sys_id       =
 $lflags       = -lefence -ldl
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -1440,7 +1440,7 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -1467,7 +1467,7 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -1494,7 +1494,7 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -1521,7 +1521,7 @@ $sys_id       =
 $lflags       = 
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -1548,7 +1548,7 @@ $sys_id       =
 $lflags       = -lsocket -lnsl -ldl
 $bn_ops       = BN_LLONG RC4_CHAR RC4_CHUNK DES_PTR DES_RISC1 DES_UNROLL BF_PTR
 $cpuid_obj    = 
-$bn_obj       = sparcv8.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -1575,7 +1575,7 @@ $sys_id       =
 $lflags       = -lsocket -lnsl -ldl
 $bn_ops       = BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR
 $cpuid_obj    = 
-$bn_obj       = sparcv8.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -1602,11 +1602,11 @@ $sys_id       = ULTRASPARC
 $lflags       = -lsocket -lnsl -ldl
 $bn_ops       = BN_LLONG RC4_CHAR RC4_CHUNK_LL DES_PTR DES_RISC1 DES_UNROLL BF_PTR
 $cpuid_obj    = 
-$bn_obj       = sparcv8plus.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
-$md5_obj      = md5-sparcv8plus.o
+$md5_obj      = 
 $sha1_obj     = 
 $cast_obj     = 
 $rc4_obj      = 
@@ -1629,11 +1629,11 @@ $sys_id       = ULTRASPARC
 $lflags       = -lsocket -lnsl -ldl
 $bn_ops       = BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR
 $cpuid_obj    = 
-$bn_obj       = sparcv8plus.o
+$bn_obj       = 
-$des_obj      = des_enc-sparc.o fcrypt_b.o
+$des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
-$md5_obj      = md5-sparcv8plus.o
+$md5_obj      = 
 $sha1_obj     = 
 $cast_obj     = 
 $rc4_obj      = 
@@ -1649,14 +1649,14 @@ $arflags      =
 
 *** debug-steve
 $cc           = gcc
-$cflags       = -DL_ENDIAN -DREF_CHECK -DCONF_DEBUG -DDEBUG_SAFESTACK -DCRYPTO_MDEBUG_ALL -DPEDANTIC -g -march=i486 -pedantic -Wno-long-long -Wall -Werror -Wshadow -pipe
+$cflags       = -DL_ENDIAN -DREF_CHECK -DCONF_DEBUG -DDEBUG_SAFESTACK -DCRYPTO_MDEBUG_ALL -DPEDANTIC -g -m32 -pedantic -Wno-long-long -Wall -Werror -Wshadow -pipe
 $unistd       = 
 $thread_cflag = -D_REENTRANT
 $sys_id       = 
 $lflags       = -rdynamic -ldl
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -1818,7 +1818,7 @@ $sys_id       =
 $lflags       = -lnsl -lsocket
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -1953,14 +1953,14 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = SIXTY_FOUR_BIT MD2_CHAR RC4_INDEX DES_UNROLL DES_RISC1 DES_INT
 $cpuid_obj    = 
-$bn_obj       = bn-ia64.o
+$bn_obj       = 
 $des_obj      = 
-$aes_obj      = aes_core.o aes_cbc.o aes-ia64.o
+$aes_obj      = 
 $bf_obj       = 
 $md5_obj      = 
-$sha1_obj     = sha1-ia64.o sha256-ia64.o sha512-ia64.o
+$sha1_obj     = 
 $cast_obj     = 
-$rc4_obj      = rc4-ia64.o
+$rc4_obj      = 
 $rmd160_obj   = 
 $rc5_obj      = 
 $dso_scheme   = dlfcn
@@ -1980,14 +1980,14 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = SIXTY_FOUR_BIT MD2_CHAR RC4_INDEX DES_UNROLL DES_RISC1 DES_INT
 $cpuid_obj    = 
-$bn_obj       = bn-ia64.o
+$bn_obj       = 
 $des_obj      = 
-$aes_obj      = aes_core.o aes_cbc.o aes-ia64.o
+$aes_obj      = 
 $bf_obj       = 
 $md5_obj      = 
-$sha1_obj     = sha1-ia64.o sha256-ia64.o sha512-ia64.o
+$sha1_obj     = 
 $cast_obj     = 
-$rc4_obj      = rc4-ia64.o
+$rc4_obj      = 
 $rmd160_obj   = 
 $rc5_obj      = 
 $dso_scheme   = dlfcn
@@ -2115,7 +2115,7 @@ $sys_id       =
 $lflags       = -Wl,+s -ldld
 $bn_ops       = SIXTY_FOUR_BIT MD2_CHAR RC4_INDEX RC4_CHAR DES_UNROLL DES_RISC1 DES_INT
 $cpuid_obj    = 
-$bn_obj       = pa-risc2.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -2142,7 +2142,7 @@ $sys_id       =
 $lflags       = -Wl,+s -ldld
 $bn_ops       = SIXTY_FOUR_BIT RC4_CHAR RC4_CHUNK DES_PTR DES_UNROLL DES_RISC1
 $cpuid_obj    = 
-$bn_obj       = pa-risc2.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -2169,14 +2169,14 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = SIXTY_FOUR_BIT_LONG MD2_CHAR RC4_INDEX DES_UNROLL DES_RISC1 DES_INT
 $cpuid_obj    = 
-$bn_obj       = bn-ia64.o
+$bn_obj       = 
 $des_obj      = 
-$aes_obj      = aes_core.o aes_cbc.o aes-ia64.o
+$aes_obj      = 
 $bf_obj       = 
 $md5_obj      = 
-$sha1_obj     = sha1-ia64.o sha256-ia64.o sha512-ia64.o
+$sha1_obj     = 
 $cast_obj     = 
-$rc4_obj      = rc4-ia64.o
+$rc4_obj      = 
 $rmd160_obj   = 
 $rc5_obj      = 
 $dso_scheme   = dlfcn
@@ -2196,14 +2196,14 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = SIXTY_FOUR_BIT_LONG MD2_CHAR RC4_INDEX DES_UNROLL DES_RISC1 DES_INT
 $cpuid_obj    = 
-$bn_obj       = bn-ia64.o
+$bn_obj       = 
 $des_obj      = 
-$aes_obj      = aes_core.o aes_cbc.o aes-ia64.o
+$aes_obj      = 
 $bf_obj       = 
 $md5_obj      = 
-$sha1_obj     = sha1-ia64.o sha256-ia64.o sha512-ia64.o
+$sha1_obj     = 
 $cast_obj     = 
-$rc4_obj      = rc4-ia64.o
+$rc4_obj      = 
 $rmd160_obj   = 
 $rc5_obj      = 
 $dso_scheme   = dlfcn
@@ -2223,7 +2223,7 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = SIXTY_FOUR_BIT_LONG MD2_CHAR RC4_INDEX RC4_CHAR DES_UNROLL DES_RISC1 DES_INT
 $cpuid_obj    = 
-$bn_obj       = pa-risc2W.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -2250,7 +2250,7 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = SIXTY_FOUR_BIT_LONG MD2_CHAR RC4_INDEX RC4_CHAR DES_UNROLL DES_RISC1 DES_INT
 $cpuid_obj    = 
-$bn_obj       = pa-risc2W.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -2277,7 +2277,7 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -2351,14 +2351,14 @@ $arflags      =
 
 *** irix-mips3-cc
 $cc           = cc
-$cflags       = -n32 -mips3 -O2 -use_readonly_const -DTERMIOS -DB_ENDIAN -DBN_DIV3W
+$cflags       = -n32 -mips3 -O2 -use_readonly_const -G0 -rdata_shared -DTERMIOS -DB_ENDIAN -DBN_DIV3W
 $unistd       = 
 $thread_cflag = -D_SGI_MP_SOURCE
 $sys_id       = 
 $lflags       = 
 $bn_ops       = DES_PTR RC4_CHAR RC4_CHUNK_LL DES_RISC2 DES_UNROLL BF_PTR SIXTY_FOUR_BIT
 $cpuid_obj    = 
-$bn_obj       = bn-mips3.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -2385,7 +2385,7 @@ $sys_id       =
 $lflags       = 
 $bn_ops       = MD2_CHAR RC4_INDEX RC4_CHAR RC4_CHUNK_LL DES_UNROLL DES_RISC2 DES_PTR BF_PTR SIXTY_FOUR_BIT
 $cpuid_obj    = 
-$bn_obj       = bn-mips3.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -2405,14 +2405,14 @@ $arflags      =
 
 *** irix64-mips4-cc
 $cc           = cc
-$cflags       = -64 -mips4 -O2 -use_readonly_const -DTERMIOS -DB_ENDIAN -DBN_DIV3W
+$cflags       = -64 -mips4 -O2 -use_readonly_const -G0 -rdata_shared -DTERMIOS -DB_ENDIAN -DBN_DIV3W
 $unistd       = 
 $thread_cflag = -D_SGI_MP_SOURCE
 $sys_id       = 
 $lflags       = 
 $bn_ops       = RC4_CHAR RC4_CHUNK DES_RISC2 DES_UNROLL SIXTY_FOUR_BIT_LONG
 $cpuid_obj    = 
-$bn_obj       = bn-mips3.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -2439,7 +2439,7 @@ $sys_id       =
 $lflags       = 
 $bn_ops       = RC4_CHAR RC4_CHUNK DES_RISC2 DES_UNROLL SIXTY_FOUR_BIT_LONG
 $cpuid_obj    = 
-$bn_obj       = bn-mips3.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -2574,7 +2574,7 @@ $sys_id       =
 $lflags       = 
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-out.o
-$bn_obj       = bn86-out.o co86-out.o
+$bn_obj       = bn86-out.o co86-out.o mo86-out.o
 $des_obj      = dx86-out.o yx86-out.o
 $aes_obj      = ax86-out.o
 $bf_obj       = bx86-out.o
@@ -2601,7 +2601,7 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -2682,7 +2682,7 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -2709,14 +2709,14 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHUNK
 $cpuid_obj    = 
-$bn_obj       = bn-ia64.o
+$bn_obj       = 
 $des_obj      = 
-$aes_obj      = aes_core.o aes_cbc.o aes-ia64.o
+$aes_obj      = 
 $bf_obj       = 
 $md5_obj      = 
-$sha1_obj     = sha1-ia64.o sha256-ia64.o sha512-ia64.o
+$sha1_obj     = 
 $cast_obj     = 
-$rc4_obj      = rc4-ia64.o
+$rc4_obj      = 
 $rmd160_obj   = 
 $rc5_obj      = 
 $dso_scheme   = dlfcn
@@ -2736,14 +2736,14 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHUNK
 $cpuid_obj    = 
-$bn_obj       = bn-ia64.o
+$bn_obj       = 
 $des_obj      = 
-$aes_obj      = aes_core.o aes_cbc.o aes-ia64.o
+$aes_obj      = 
 $bf_obj       = 
 $md5_obj      = 
-$sha1_obj     = sha1-ia64.o sha256-ia64.o sha512-ia64.o
+$sha1_obj     = 
 $cast_obj     = 
-$rc4_obj      = rc4-ia64.o
+$rc4_obj      = 
 $rmd160_obj   = 
 $rc5_obj      = 
 $dso_scheme   = dlfcn
@@ -2763,14 +2763,14 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHUNK
 $cpuid_obj    = 
-$bn_obj       = bn-ia64.o
+$bn_obj       = 
 $des_obj      = 
-$aes_obj      = aes_core.o aes_cbc.o aes-ia64.o
+$aes_obj      = 
 $bf_obj       = 
 $md5_obj      = 
-$sha1_obj     = sha1-ia64.o sha256-ia64.o sha512-ia64.o
+$sha1_obj     = 
 $cast_obj     = 
-$rc4_obj      = rc4-ia64.o
+$rc4_obj      = 
 $rmd160_obj   = 
 $rc5_obj      = 
 $dso_scheme   = dlfcn
@@ -2790,7 +2790,7 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = BN_LLONG RC4_CHAR RC4_CHUNK DES_RISC1 DES_UNROLL
 $cpuid_obj    = 
-$bn_obj       = linux_ppc32.o
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -2817,7 +2817,34 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_RISC1 DES_UNROLL
 $cpuid_obj    = 
-$bn_obj       = linux_ppc64.o
+$bn_obj       = 
+$des_obj      = 
+$aes_obj      = 
+$bf_obj       = 
+$md5_obj      = 
+$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       = 
+$arflags      = 
+
+*** linux-s390x
+$cc           = gcc
+$cflags       = -m64 -DB_ENDIAN -DTERMIO -O3 -Wall
+$unistd       = 
+$thread_cflag = -D_REENTRANT
+$sys_id       = 
+$lflags       = -ldl
+$bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_INT DES_UNROLL
+$cpuid_obj    = 
+$bn_obj       = 
 $des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
@@ -2844,8 +2871,8 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR
 $cpuid_obj    = 
-$bn_obj       = sparcv8.o
+$bn_obj       = 
-$des_obj      = des_enc-sparc.o fcrypt_b.o
+$des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
 $md5_obj      = 
@@ -2871,11 +2898,11 @@ $sys_id       = ULTRASPARC
 $lflags       = -ldl
 $bn_ops       = BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR
 $cpuid_obj    = 
-$bn_obj       = sparcv8plus.o
+$bn_obj       = 
-$des_obj      = des_enc-sparc.o fcrypt_b.o
+$des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
-$md5_obj      = md5-sparcv8plus.o
+$md5_obj      = 
 $sha1_obj     = 
 $cast_obj     = 
 $rc4_obj      = 
@@ -2898,12 +2925,12 @@ $sys_id       =
 $lflags       = -ldl
 $bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHUNK BF_PTR2 DES_INT DES_UNROLL
 $cpuid_obj    = x86_64cpuid.o
-$bn_obj       = x86_64-gcc.o
+$bn_obj       = x86_64-gcc.o x86_64-mont.o
 $des_obj      = 
-$aes_obj      = 
+$aes_obj      = aes-x86_64.o
 $bf_obj       = 
 $md5_obj      = md5-x86_64.o
-$sha1_obj     = 
+$sha1_obj     = sha1-x86_64.o sha256-x86_64.o sha512-x86_64.o
 $cast_obj     = 
 $rc4_obj      = rc4-x86_64.o
 $rmd160_obj   = 
@@ -2926,10 +2953,10 @@ $lflags       = -ldl
 $bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR
 $cpuid_obj    = 
 $bn_obj       = 
-$des_obj      = des_enc-sparc.o fcrypt_b.o
+$des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
-$md5_obj      = md5-sparcv9.o
+$md5_obj      = 
 $sha1_obj     = 
 $cast_obj     = 
 $rc4_obj      = 
@@ -2952,7 +2979,7 @@ $sys_id       = MINGW32
 $lflags       = -lwsock32 -lgdi32
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT EXPORT_VAR_AS_FN
 $cpuid_obj    = x86cpuid-cof.o
-$bn_obj       = bn86-cof.o co86-cof.o
+$bn_obj       = bn86-cof.o co86-cof.o mo86-cof.o
 $des_obj      = dx86-cof.o yx86-cof.o
 $aes_obj      = ax86-cof.o
 $bf_obj       = bx86-cof.o
@@ -3357,7 +3384,7 @@ $sys_id       =
 $lflags       = -lsocket -lnsl
 $bn_ops       = DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -3384,7 +3411,7 @@ $sys_id       =
 $lflags       = -lsocket -lnsl
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -3465,8 +3492,8 @@ $sys_id       =
 $lflags       = -lsocket -lnsl -ldl
 $bn_ops       = BN_LLONG RC4_CHAR RC4_CHUNK DES_PTR DES_RISC1 DES_UNROLL BF_PTR
 $cpuid_obj    = 
-$bn_obj       = sparcv8.o
+$bn_obj       = 
-$des_obj      = des_enc-sparc.o fcrypt_b.o
+$des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
 $md5_obj      = 
@@ -3492,8 +3519,8 @@ $sys_id       =
 $lflags       = -lsocket -lnsl -ldl
 $bn_ops       = BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR
 $cpuid_obj    = 
-$bn_obj       = sparcv8.o
+$bn_obj       = 
-$des_obj      = des_enc-sparc.o fcrypt_b.o
+$des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
 $md5_obj      = 
@@ -3519,11 +3546,11 @@ $sys_id       = ULTRASPARC
 $lflags       = -lsocket -lnsl -ldl
 $bn_ops       = BN_LLONG RC4_CHAR RC4_CHUNK_LL DES_PTR DES_RISC1 DES_UNROLL BF_PTR
 $cpuid_obj    = 
-$bn_obj       = sparcv8plus.o
+$bn_obj       = 
-$des_obj      = des_enc-sparc.o fcrypt_b.o
+$des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
-$md5_obj      = md5-sparcv8plus.o
+$md5_obj      = 
 $sha1_obj     = 
 $cast_obj     = 
 $rc4_obj      = 
@@ -3546,11 +3573,11 @@ $sys_id       = ULTRASPARC
 $lflags       = -lsocket -lnsl -ldl
 $bn_ops       = BN_LLONG RC4_CHAR RC4_CHUNK DES_UNROLL BF_PTR
 $cpuid_obj    = 
-$bn_obj       = sparcv8plus.o
+$bn_obj       = 
-$des_obj      = des_enc-sparc.o fcrypt_b.o
+$des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
-$md5_obj      = md5-sparcv8plus.o
+$md5_obj      = 
 $sha1_obj     = 
 $cast_obj     = 
 $rc4_obj      = 
@@ -3600,7 +3627,7 @@ $sys_id       =
 $lflags       = -lsocket -lnsl -ldl
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -3628,10 +3655,10 @@ $lflags       = -lsocket -lnsl -ldl
 $bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_INT DES_PTR DES_RISC1 DES_UNROLL BF_PTR
 $cpuid_obj    = 
 $bn_obj       = 
-$des_obj      = des_enc-sparc.o fcrypt_b.o
+$des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
-$md5_obj      = md5-sparcv9.o
+$md5_obj      = 
 $sha1_obj     = 
 $cast_obj     = 
 $rc4_obj      = 
@@ -3655,10 +3682,10 @@ $lflags       = -lsocket -lnsl -ldl
 $bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHAR RC4_CHUNK DES_INT DES_PTR DES_RISC1 DES_UNROLL BF_PTR
 $cpuid_obj    = 
 $bn_obj       = 
-$des_obj      = des_enc-sparc.o fcrypt_b.o
+$des_obj      = 
 $aes_obj      = 
 $bf_obj       = 
-$md5_obj      = md5-sparcv9.o
+$md5_obj      = 
 $sha1_obj     = 
 $cast_obj     = 
 $rc4_obj      = 
@@ -3681,12 +3708,12 @@ $sys_id       =
 $lflags       = -lsocket -lnsl -ldl
 $bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHUNK BF_PTR2 DES_INT DES_UNROLL
 $cpuid_obj    = x86_64cpuid.o
-$bn_obj       = x86_64-gcc.o
+$bn_obj       = x86_64-gcc.o x86_64-mont.o
 $des_obj      = 
-$aes_obj      = 
+$aes_obj      = aes-x86_64.o
 $bf_obj       = 
 $md5_obj      = md5-x86_64.o
-$sha1_obj     = 
+$sha1_obj     = sha1-x86_64.o sha256-x86_64.o sha512-x86_64.o
 $cast_obj     = 
 $rc4_obj      = rc4-x86_64.o
 $rmd160_obj   = 
@@ -3708,12 +3735,12 @@ $sys_id       =
 $lflags       = -lsocket -lnsl -ldl
 $bn_ops       = SIXTY_FOUR_BIT_LONG RC4_CHUNK BF_PTR2 DES_INT DES_UNROLL
 $cpuid_obj    = x86_64cpuid.o
-$bn_obj       = x86_64-gcc.o
+$bn_obj       = x86_64-gcc.o x86_64-mont.o
 $des_obj      = 
-$aes_obj      = 
+$aes_obj      = aes-x86_64.o
 $bf_obj       = 
 $md5_obj      = md5-x86_64.o
-$sha1_obj     = 
+$sha1_obj     = sha1-x86_64.o sha256-x86_64.o sha512-x86_64.o
 $cast_obj     = 
 $rc4_obj      = rc4-x86_64.o
 $rmd160_obj   = 
@@ -3924,7 +3951,7 @@ $sys_id       =
 $lflags       = -lsocket -lnsl
 $bn_ops       = BN_LLONG MD2_CHAR RC4_INDEX DES_PTR DES_RISC1 DES_UNROLL
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o
@@ -3951,7 +3978,7 @@ $sys_id       =
 $lflags       = -lsocket -lnsl
 $bn_ops       = BN_LLONG DES_PTR DES_RISC1 DES_UNROLL RC4_INDEX MD2_INT
 $cpuid_obj    = x86cpuid-elf.o
-$bn_obj       = bn86-elf.o co86-elf.o
+$bn_obj       = bn86-elf.o co86-elf.o mo86-elf.o
 $des_obj      = dx86-elf.o yx86-elf.o
 $aes_obj      = ax86-elf.o
 $bf_obj       = bx86-elf.o

apps/Makefile

@@ -153,20 +153,14 @@ $(EXE): progs.h $(E_OBJ) $(PROGRAM).o $(DLIBCRYPTO) $(DLIBSSL)
 	shlib_target=; if [ -n "$(SHARED_LIBS)" ]; then \
 		shlib_target="$(SHLIB_TARGET)"; \
 	elif [ -n "$(FIPSCANLIB)" ]; then \
-	  FIPSLD_CC=$(CC); CC=$(TOP)/fips-1.0/fipsld; export CC FIPSLD_CC; \
+	  FIPSLD_CC="$(CC)"; CC=$(TOP)/fips/fipsld; export CC FIPSLD_CC; \
 	fi; \
-	if [ "$${shlib_target}" = "darwin-shared" ] ; then \
-	  LIBRARIES="$(DLIBSSL) $(LIBKRB5) $(DLIBCRYPTO)" ; \
-	else \
 	LIBRARIES="$(LIBSSL) $(LIBKRB5) $(LIBCRYPTO)" ; \
-	fi; \
+	[ "x$(FIPSCANLIB)" = "xlibfips" ] && LIBRARIES="$$LIBRARIES -lfips"; \
 	$(MAKE) -f $(TOP)/Makefile.shared -e \
-		CC=$${CC} APPNAME=$(EXE) OBJECTS="$(PROGRAM).o $(E_OBJ)" \
+		CC="$${CC}" APPNAME=$(EXE) OBJECTS="$(PROGRAM).o $(E_OBJ)" \
 		LIBDEPS="$(PEX_LIBS) $$LIBRARIES $(EX_LIBS)" \
 		link_app.$${shlib_target}
-	-(cd ..; \
-	  OPENSSL="`pwd`/util/opensslwrap.sh"; export OPENSSL; \
-	  $(PERL) tools/c_rehash certs)
 
 progs.h: progs.pl
 	$(PERL) progs.pl $(E_EXE) >progs.h

apps/dgst.c

@@ -256,29 +256,33 @@ ERR_load_crypto_strings();
 		BIO_printf(bio_err,"-engine e       use engine e, possibly a hardware device.\n");
 #endif
 
-		BIO_printf(bio_err,"-%3s to use the %s message digest algorithm (default)\n",
+		BIO_printf(bio_err,"-%-14s to use the %s message digest algorithm (default)\n",
 			LN_md5,LN_md5);
-		BIO_printf(bio_err,"-%3s to use the %s message digest algorithm\n",
+		BIO_printf(bio_err,"-%-14s to use the %s message digest algorithm\n",
 			LN_md4,LN_md4);
-		BIO_printf(bio_err,"-%3s to use the %s message digest algorithm\n",
+		BIO_printf(bio_err,"-%-14s to use the %s message digest algorithm\n",
 			LN_md2,LN_md2);
 #ifndef OPENSSL_NO_SHA
-		BIO_printf(bio_err,"-%3s to use the %s message digest algorithm\n",
+		BIO_printf(bio_err,"-%-14s to use the %s message digest algorithm\n",
 			LN_sha1,LN_sha1);
-		BIO_printf(bio_err,"-%3s to use the %s message digest algorithm\n",
+		BIO_printf(bio_err,"-%-14s to use the %s message digest algorithm\n",
 			LN_sha,LN_sha);
 #ifndef OPENSSL_NO_SHA256
-		BIO_printf(bio_err,"-%3s to use the %s message digest algorithm\n",
+		BIO_printf(bio_err,"-%-14s to use the %s message digest algorithm\n",
+			LN_sha224,LN_sha224);
+		BIO_printf(bio_err,"-%-14s to use the %s message digest algorithm\n",
 			LN_sha256,LN_sha256);
 #endif
 #ifndef OPENSSL_NO_SHA512
-		BIO_printf(bio_err,"-%3s to use the %s message digest algorithm\n",
+		BIO_printf(bio_err,"-%-14s to use the %s message digest algorithm\n",
+			LN_sha384,LN_sha384);
+		BIO_printf(bio_err,"-%-14s to use the %s message digest algorithm\n",
 			LN_sha512,LN_sha512);
 #endif
 #endif
-		BIO_printf(bio_err,"-%3s to use the %s message digest algorithm\n",
+		BIO_printf(bio_err,"-%-14s to use the %s message digest algorithm\n",
 			LN_mdc2,LN_mdc2);
-		BIO_printf(bio_err,"-%3s to use the %s message digest algorithm\n",
+		BIO_printf(bio_err,"-%-14s to use the %s message digest algorithm\n",
 			LN_ripemd160,LN_ripemd160);
 		err=1;
 		goto end;

apps/openssl-vms.cnf

@@ -8,9 +8,8 @@
 HOME			= .
 RANDFILE		= $ENV::HOME/.rnd
 
-# Extra OBJECT IDENTIFIER info:
+# Uncomment out to enable OpenSSL configuration see config(3)
-#oid_file		= $ENV::HOME/.oid
+# openssl_conf = openssl_init
-oid_section		= new_oids
 
 # To use this configuration file with the "-extfile" option of the
 # "openssl x509" utility, name here the section containing the
@@ -19,13 +18,22 @@ oid_section		= new_oids
 # (Alternatively, use a configuration file that has only
 # X.509v3 extensions in its main [= default] section.)
 
+[openssl_init]
+# Extra OBJECT IDENTIFIER info:
+oid_section = new_oids
+alg_section = algs
+
 [ new_oids ]
 
-# We can add new OIDs in here for use by 'ca' and 'req'.
+# We can add new OIDs in here for use by any config aware application
 # Add a simple OID like this:
-# testoid1=1.2.3.4
+# shortname=Long Object Identifier Name, 1.2.3.4
 # Or use config file substitution like this:
-# testoid2=${testoid1}.5.6
+# testoid2=OID2 LONG NAME, ${testoid1}.5.6, OTHER OID
+
+[ algs ]
+# Algorithm configuration options. Currently just fips_mode
+fips_mode = no
 
 ####################################################################
 [ ca ]

config

@@ -48,10 +48,10 @@ done
 
 # First get uname entries that we use below
 
-MACHINE=`(uname -m) 2>/dev/null` || MACHINE="unknown"
+[ "$MACHINE" ] || MACHINE=`(uname -m) 2>/dev/null` || MACHINE="unknown"
-RELEASE=`(uname -r) 2>/dev/null` || RELEASE="unknown"
+[ "$RELEASE" ] || RELEASE=`(uname -r) 2>/dev/null` || RELEASE="unknown"
-SYSTEM=`(uname -s) 2>/dev/null`  || SYSTEM="unknown"
+[ "$SYSTEM" ] || SYSTEM=`(uname -s) 2>/dev/null`  || SYSTEM="unknown"
-VERSION=`(uname -v) 2>/dev/null` || VERSION="unknown"
+[ "$BUILD" ] || VERSION=`(uname -v) 2>/dev/null` || VERSION="unknown"
 
 
 # Now test for ISC and SCO, since it is has a braindamaged uname.
@@ -501,20 +501,7 @@ case "$GUESSOS" in
 	OUT="irix-mips3-$CC"
 	;;
   mips4-sgi-irix64)
-	echo "WARNING! If you wish to build 64-bit library, then you have to"
+	OUT="irix64-mips4-$CC"
-	echo "         invoke './Configure irix64-mips4-$CC' *manually*."
-	if [ "$TEST" = "false" -a -t 1 ]; then
-	  echo "         You have about 5 seconds to press Ctrl-C to abort."
-	  (trap "stty `stty -g`" 2 0; stty -icanon min 0 time 50; read waste) <&1
-	fi
-        #CPU=`(hinv -t cpu) 2>/dev/null | head -1 | sed 's/^CPU:[^R]*R\([0-9]*\).*/\1/'`
-        #CPU=${CPU:-0}
-        #if [ $CPU -ge 5000 ]; then
-        #        options="$options -mips4"
-        #else
-        #        options="$options -mips3"
-        #fi
-	OUT="irix-mips3-$CC"
 	;;
   ppc-apple-rhapsody) OUT="rhapsody-ppc-cc" ;;
   ppc-apple-darwin*) OUT="darwin-ppc-cc" ;;
@@ -534,25 +521,12 @@ case "$GUESSOS" in
 	fi
 	;;
   ppc64-*-linux2)
-	echo "WARNING! If you wish to build 64-bit library, then you have to"
+	OUT="linux-ppc64"
-	echo "         invoke './Configure linux-ppc64' *manually*."
-	if [ "$TEST" = "false" -a -t 1 ]; then
-	    echo "         You have about 5 seconds to press Ctrl-C to abort."
-	    (trap "stty `stty -g`" 2 0; stty -icanon min 0 time 50; read waste) <&1
-	fi
-	OUT="linux-ppc"
 	;;
   ppc-*-linux2) OUT="linux-ppc" ;;
   ia64-*-linux?) OUT="linux-ia64" ;;
   sparc64-*-linux2)
-	echo "WARNING! If you *know* that your GNU C supports 64-bit/V9 ABI"
+	OUT="linux64-sparcv9" ;;
-	echo "         and wish to build 64-bit library, then you have to"
-	echo "         invoke './Configure linux64-sparcv9' *manually*."
-	if [ "$TEST" = "false" -a -t 1 ]; then
-	  echo "          You have about 5 seconds to press Ctrl-C to abort."
-	  (trap "stty `stty -g`" 2 0; stty -icanon min 0 time 50; read waste) <&1
-	fi
-	OUT="linux-sparcv9" ;;
   sparc-*-linux2)
 	KARCH=`awk '/^type/{print$3;exit(0);}' /proc/cpuinfo`
 	case ${KARCH:-sun4} in
@@ -589,7 +563,7 @@ case "$GUESSOS" in
   sh*-*-linux2)  OUT="linux-generic32"; options="$options -DL_ENDIAN" ;;
   m68k*-*-linux2) OUT="linux-generic32"; options="$options -DB_ENDIAN" ;;
   s390-*-linux2) OUT="linux-generic32"; options="$options -DB_ENDIAN -DNO_ASM" ;;
-  s390x-*-linux2) OUT="linux-generic64"; options="$options -DB_ENDIAN" ;;
+  s390x-*-linux2) OUT="linux-s390x" ;;
   x86_64-*-linux?) OUT="linux-x86_64" ;;
   *86-*-linux2) OUT="linux-elf"
 	if [ "$GCCVER" -gt 28 ]; then
@@ -610,32 +584,13 @@ case "$GUESSOS" in
 	ISA64=`(isalist) 2>/dev/null | grep sparcv9`
 	if [ "$ISA64" != "" ]; then
 	    if [ "$CC" = "cc" -a $CCVER -ge 50 ]; then
-		echo "WARNING! If you wish to build 64-bit library, then you have to"
+		OUT="solaris64-sparcv9-cc"
-		echo "         invoke './Configure solaris64-sparcv9-cc' *manually*."
-		if [ "$TEST" = "false" -a -t 1 ]; then
-		  echo "         You have about 5 seconds to press Ctrl-C to abort."
-		  (trap "stty `stty -g`" 2 0; stty -icanon min 0 time 50; read waste) <&1
-		fi
 	    elif [ "$CC" = "gcc" -a "$GCC_ARCH" = "-m64" ]; then
 		# $GCC_ARCH denotes default ABI chosen by compiler driver
 		# (first one found on the $PATH). I assume that user
 		# expects certain consistency with the rest of his builds
 		# and therefore switch over to 64-bit. <appro>
 		OUT="solaris64-sparcv9-gcc"
-		echo "WARNING! If you wish to build 32-bit library, then you have to"
-		echo "         invoke './Configure solaris-sparcv9-gcc' *manually*."
-		if [ "$TEST" = "false" -a -t 1 ]; then
-		  echo "         You have about 5 seconds to press Ctrl-C to abort."
-		  (trap "stty `stty -g`" 2 0; stty -icanon min 0 time 50; read waste) <&1
-		fi
-	    elif [ "$GCC_ARCH" = "-m32" ]; then
-		echo "NOTICE! If you *know* that your GNU C supports 64-bit/V9 ABI"
-		echo "        and wish to build 64-bit library, then you have to"
-		echo "        invoke './Configure solaris64-sparcv9-gcc' *manually*."
-		if [ "$TEST" = "false" -a -t 1 ]; then
-		  echo "         You have about 5 seconds to press Ctrl-C to abort."
-		  (trap "stty `stty -g`" 2 0; stty -icanon min 0 time 50; read waste) <&1
-		fi
 	    fi
 	fi
 	;;
@@ -656,8 +611,8 @@ case "$GUESSOS" in
   *-*-sunos4)		OUT="sunos-$CC" ;;
 
   *86*-*-bsdi4)		OUT="BSD-x86-elf"; options="$options no-sse2 -ldl" ;;
-  alpha*-*-*bsd*)	OUT="BSD-generic64; options="$options -DL_ENDIAN" ;;
+  alpha*-*-*bsd*)	OUT="BSD-generic64"; options="$options -DL_ENDIAN" ;;
-  powerpc64-*-*bsd*)	OUT="BSD-generic64; options="$options -DB_ENDIAN" ;;
+  powerpc64-*-*bsd*)	OUT="BSD-generic64"; options="$options -DB_ENDIAN" ;;
   sparc64-*-*bsd*)	OUT="BSD-sparc64" ;;
   ia64-*-*bsd*)		OUT="BSD-ia64" ;;
   amd64-*-*bsd*)	OUT="BSD-x86_64" ;;
@@ -702,23 +657,11 @@ case "$GUESSOS" in
 	CPU_VERSION=${CPU_VERSION:-0}
 	# See <sys/unistd.h> for further info on CPU_VERSION.
 	if   [ $CPU_VERSION -ge 768 ]; then	# IA-64 CPU
-	     echo "WARNING! 64-bit ABI is the default configured ABI on HP-UXi."
-	     echo "         If you wish to build 32-bit library, the you have to"
-	     echo "         invoke './Configure hpux-ia64-cc' *manually*."
-	     if [ "$TEST" = "false" -a -t 1 ]; then
-		echo "         You have about 5 seconds to press Ctrl-C to abort."
-		(trap "stty `stty -g`" 2 0; stty -icanon min 0 time 50; read waste) <&1
-	     fi
 	     OUT="hpux64-ia64-cc"
 	elif [ $CPU_VERSION -ge 532 ]; then	# PA-RISC 2.x CPU
 	     OUT=${OUT:-"hpux-parisc2-${CC}"}
 	     if [ $KERNEL_BITS -eq 64 -a "$CC" = "cc" ]; then
-		echo "WARNING! If you wish to build 64-bit library then you have to"
+		OUT="hpux64-parisc2-${CC}"
-		echo "         invoke './Configure hpux64-parisc2-cc' *manually*."
-		if [ "$TEST" = "false" -a -t 1 ]; then
-		  echo "         You have about 5 seconds to press Ctrl-C to abort."
-		  (trap "stty `stty -g`" 2 0; stty -icanon min 0 time 50; read waste) <&1
-		fi
 	     fi
 	elif [ $CPU_VERSION -ge 528 ]; then	# PA-RISC 1.1+ CPU
 	     OUT="hpux-parisc-${CC}"
@@ -741,12 +684,7 @@ case "$GUESSOS" in
 	else
 	    OUT="aix-cc"
 	    if [ $KERNEL_BITS -eq 64 ]; then
-		echo "WARNING! If you wish to build 64-bit kit, then you have to"
+		OUT="aix64-cc"
-		echo "         invoke './Configure aix64-cc' *manually*."
-		if [ "$TEST" = "false" -a -t 1 ]; then
-		    echo "         You have ~5 seconds to press Ctrl-C to abort."
-		    (trap "stty `stty -g`" 2 0; stty -icanon min 0 time 50; read waste) <&1
-		fi
 	    fi
 	fi
 	if (lsattr -E -O -l `lsdev -c processor|awk '{print$1;exit}'` | grep -i powerpc) >/dev/null 2>&1; then

crypto/Makefile

@@ -47,7 +47,7 @@ ALL=    $(GENERAL) $(SRC) $(HEADER)
 top:
 	@(cd ..; $(MAKE) DIRS=$(DIR) all)
 
-all: shared
+all: lib
 
 buildinf.h: ../Makefile
 	( echo "#ifndef MK1MF_BUILD"; \
@@ -95,7 +95,7 @@ links:
 	@target=links; $(RECURSIVE_MAKE)
 
 # lib: and $(LIB): are splitted to avoid end-less loop
-lib:	$(LIB)
+lib:	buildinf.h $(LIB) subdirs
 	@touch lib
 $(LIB):	$(LIBOBJ)
 	$(ARX) $(LIB) $(LIBOBJ)

crypto/aes/Makefile

@@ -57,6 +57,11 @@ ax86-cof.s: asm/aes-586.pl ../perlasm/x86asm.pl
 ax86-out.s: asm/aes-586.pl ../perlasm/x86asm.pl
 	(cd asm; $(PERL) aes-586.pl a.out $(CFLAGS) $(PROCESSOR) > ../$@)
 
+aes-x86_64.s: asm/aes-x86_64.pl
+	$(PERL) asm/aes-x86_64.pl $@
+# GNU make "catch all"
+aes-%.s:	asm/aes-%.pl;	$(PERL) $< $(CFLAGS) > $@
+
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
 

crypto/aes/aes_core.c

@@ -37,10 +37,9 @@
 
 #include <stdlib.h>
 #include <openssl/aes.h>
+#include <openssl/fips.h>
 #include "aes_locl.h"
 
-#ifndef OPENSSL_FIPS
-
 /*
 Te0[x] = S [x].[02, 01, 01, 03];
 Te1[x] = S [x].[03, 02, 01, 01];
@@ -633,6 +632,10 @@ int AES_set_encrypt_key(const unsigned char *userKey, const int bits,
    	int i = 0;
 	u32 temp;
 
+#ifdef OPENSSL_FIPS
+	FIPS_selftest_check();
+#endif
+
 	if (!userKey || !key)
 		return -1;
 	if (bits != 128 && bits != 192 && bits != 256)
@@ -1159,5 +1162,3 @@ void AES_decrypt(const unsigned char *in, unsigned char *out,
 }
 
 #endif /* AES_ASM */
-
-#endif

crypto/aes/aes_ige.c

@@ -62,9 +62,11 @@ typedef struct {
 /* XXX: probably some better way to do this */
 #if defined(__i386__) || defined(__x86_64__)
 #define UNALIGNED_MEMOPS_ARE_FAST 1
+#else
+#define UNALIGNED_MEMOPS_ARE_FAST 0
 #endif
 
-#ifdef UNALIGNED_MEMOPS_ARE_FAST
+#if UNALIGNED_MEMOPS_ARE_FAST
 #define load_block(d, s)        (d) = *(const aes_block_t *)(s)
 #define store_block(d, s)       *(aes_block_t *)(d) = (s)
 #else
@@ -89,7 +91,8 @@ void AES_ige_encrypt(const unsigned char *in, unsigned char *out,
 
 	if (AES_ENCRYPT == enc)
 		{
-		if (in != out)
+		if (in != out &&
+		    (UNALIGNED_MEMOPS_ARE_FAST || ((size_t)in|(size_t)out|(size_t)ivec)%sizeof(long)==0))
 			{
 			aes_block_t *ivp = (aes_block_t *)ivec;
 			aes_block_t *iv2p = (aes_block_t *)(ivec + AES_BLOCK_SIZE);
@@ -143,7 +146,8 @@ void AES_ige_encrypt(const unsigned char *in, unsigned char *out,
 		}
 	else
 		{
-		if(in != out)
+		if (in != out &&
+		    (UNALIGNED_MEMOPS_ARE_FAST || ((size_t)in|(size_t)out|(size_t)ivec)%sizeof(long)==0))
 			{
 			aes_block_t *ivp = (aes_block_t *)ivec;
 			aes_block_t *iv2p = (aes_block_t *)(ivec + AES_BLOCK_SIZE);

crypto/asn1/x_name.c

@@ -160,40 +160,40 @@ static int x509_name_ex_d2i(ASN1_VALUE **val, const unsigned char **in, long len
 					int tag, int aclass, char opt, ASN1_TLC *ctx)
 {
 	const unsigned char *p = *in, *q;
-	STACK *intname = NULL, **intname_pp = &intname;
+	union { STACK *s; ASN1_VALUE *a; } intname = {NULL};
+	union { X509_NAME *x; ASN1_VALUE *a; } nm = {NULL};
 	int i, j, ret;
-	X509_NAME *nm = NULL, **nm_pp = &nm;
 	STACK_OF(X509_NAME_ENTRY) *entries;
 	X509_NAME_ENTRY *entry;
 	q = p;
 
 	/* Get internal representation of Name */
-	ret = ASN1_item_ex_d2i((ASN1_VALUE **)intname_pp,
+	ret = ASN1_item_ex_d2i(&intname.a,
 			       &p, len, ASN1_ITEM_rptr(X509_NAME_INTERNAL),
 			       tag, aclass, opt, ctx);
 	
 	if(ret <= 0) return ret;
 
 	if(*val) x509_name_ex_free(val, NULL);
-	if(!x509_name_ex_new((ASN1_VALUE **)nm_pp, NULL)) goto err;
+	if(!x509_name_ex_new(&nm.a, NULL)) goto err;
 	/* We've decoded it: now cache encoding */
-	if(!BUF_MEM_grow(nm->bytes, p - q)) goto err;
+	if(!BUF_MEM_grow(nm.x->bytes, p - q)) goto err;
-	memcpy(nm->bytes->data, q, p - q);
+	memcpy(nm.x->bytes->data, q, p - q);
 
 	/* Convert internal representation to X509_NAME structure */
-	for(i = 0; i < sk_num(intname); i++) {
+	for(i = 0; i < sk_num(intname.s); i++) {
-		entries = (STACK_OF(X509_NAME_ENTRY) *)sk_value(intname, i);
+		entries = (STACK_OF(X509_NAME_ENTRY) *)sk_value(intname.s, i);
 		for(j = 0; j < sk_X509_NAME_ENTRY_num(entries); j++) {
 			entry = sk_X509_NAME_ENTRY_value(entries, j);
 			entry->set = i;
-			if(!sk_X509_NAME_ENTRY_push(nm->entries, entry))
+			if(!sk_X509_NAME_ENTRY_push(nm.x->entries, entry))
 				goto err;
 		}
 		sk_X509_NAME_ENTRY_free(entries);
 	}
-	sk_free(intname);
+	sk_free(intname.s);
-	nm->modified = 0;
+	nm.x->modified = 0;
-	*val = (ASN1_VALUE *)nm;
+	*val = nm.a;
 	*in = p;
 	return ret;
 	err:
@@ -219,35 +219,35 @@ static int x509_name_ex_i2d(ASN1_VALUE **val, unsigned char **out, const ASN1_IT
 
 static int x509_name_encode(X509_NAME *a)
 {
-	STACK *intname = NULL, **intname_pp = &intname;
+	union { STACK *s; ASN1_VALUE *a; } intname = {NULL};
 	int len;
 	unsigned char *p;
 	STACK_OF(X509_NAME_ENTRY) *entries = NULL;
 	X509_NAME_ENTRY *entry;
 	int i, set = -1;
-	intname = sk_new_null();
+	intname.s = sk_new_null();
-	if(!intname) goto memerr;
+	if(!intname.s) goto memerr;
 	for(i = 0; i < sk_X509_NAME_ENTRY_num(a->entries); i++) {
 		entry = sk_X509_NAME_ENTRY_value(a->entries, i);
 		if(entry->set != set) {
 			entries = sk_X509_NAME_ENTRY_new_null();
 			if(!entries) goto memerr;
-			if(!sk_push(intname, (char *)entries)) goto memerr;
+			if(!sk_push(intname.s, (char *)entries)) goto memerr;
 			set = entry->set;
 		}
 		if(!sk_X509_NAME_ENTRY_push(entries, entry)) goto memerr;
 	}
-	len = ASN1_item_ex_i2d((ASN1_VALUE **)intname_pp, NULL,
+	len = ASN1_item_ex_i2d(&intname.a, NULL,
 			       ASN1_ITEM_rptr(X509_NAME_INTERNAL), -1, -1);
 	if (!BUF_MEM_grow(a->bytes,len)) goto memerr;
 	p=(unsigned char *)a->bytes->data;
-	ASN1_item_ex_i2d((ASN1_VALUE **)intname_pp,
+	ASN1_item_ex_i2d(&intname.a,
 			 &p, ASN1_ITEM_rptr(X509_NAME_INTERNAL), -1, -1);
-	sk_pop_free(intname, sk_internal_free);
+	sk_pop_free(intname.s, sk_internal_free);
 	a->modified = 0;
 	return len;
 	memerr:
-	sk_pop_free(intname, sk_internal_free);
+	sk_pop_free(intname.s, sk_internal_free);
 	ASN1err(ASN1_F_X509_NAME_ENCODE, ERR_R_MALLOC_FAILURE);
 	return -1;
 }

crypto/bio/b_print.c

@@ -79,7 +79,7 @@
 #include <openssl/bn.h>         /* To get BN_LLONG properly defined */
 #include <openssl/bio.h>
 
-#ifdef BN_LLONG
+#if defined(BN_LLONG) || defined(SIXTY_FOUR_BIT)
 # ifndef HAVE_LONG_LONG
 #  define HAVE_LONG_LONG 1
 # endif
@@ -117,7 +117,7 @@
 
 #if HAVE_LONG_LONG
 # if defined(OPENSSL_SYS_WIN32) && !defined(__GNUC__)
-# define LLONG _int64
+# define LLONG __int64
 # else
 # define LLONG long long
 # endif

crypto/bio/b_sock.c

@@ -456,9 +456,6 @@ int BIO_sock_init(void)
 		{
 		int err;
 	  
-#ifdef SIGINT
-		signal(SIGINT,(void (*)(int))BIO_sock_cleanup);
-#endif
 		wsa_init_done=1;
 		memset(&wsa_state,0,sizeof(wsa_state));
 		if (WSAStartup(0x0101,&wsa_state)!=0)
@@ -484,11 +481,6 @@ int BIO_sock_init(void)
 
     if (!wsa_init_done)
     {
-   
-# ifdef SIGINT
-        signal(SIGINT,(void (*)(int))BIO_sock_cleanup);
-# endif
-
         wsa_init_done=1;
         wVerReq = MAKEWORD( 2, 0 );
         err = WSAStartup(wVerReq,&wsaData);
@@ -511,7 +503,7 @@ void BIO_sock_cleanup(void)
 		{
 		wsa_init_done=0;
 #ifndef OPENSSL_SYS_WINCE
-		WSACancelBlockingCall();
+		WSACancelBlockingCall();	/* Winsock 1.1 specific */
 #endif
 		WSACleanup();
 		}

crypto/bn/Makefile

@@ -67,16 +67,22 @@ bn86-elf.s:	asm/bn-586.pl ../perlasm/x86asm.pl
 	(cd asm; $(PERL) bn-586.pl elf $(CFLAGS) > ../$@)
 co86-elf.s:	asm/co-586.pl ../perlasm/x86asm.pl
 	(cd asm; $(PERL) co-586.pl elf $(CFLAGS) > ../$@)
+mo86-elf.s:	asm/x86-mont.pl ../perlasm/x86asm.pl
+	(cd asm; $(PERL) x86-mont.pl elf $(CFLAGS) > ../$@)
 # COFF
 bn86-cof.s: asm/bn-586.pl ../perlasm/x86asm.pl
 	(cd asm; $(PERL) bn-586.pl coff $(CFLAGS) > ../$@)
 co86-cof.s: asm/co-586.pl ../perlasm/x86asm.pl
 	(cd asm; $(PERL) co-586.pl coff $(CFLAGS) > ../$@)
+mo86-cof.s: asm/x86-mont.pl ../perlasm/x86asm.pl
+	(cd asm; $(PERL) x86-mont.pl coff $(CFLAGS) > ../$@)
 # a.out
 bn86-out.s: asm/bn-586.pl ../perlasm/x86asm.pl
 	(cd asm; $(PERL) bn-586.pl a.out $(CFLAGS) > ../$@)
 co86-out.s: asm/co-586.pl ../perlasm/x86asm.pl
 	(cd asm; $(PERL) co-586.pl a.out $(CFLAGS) > ../$@)
+mo86-out.s: asm/x86-mont.pl ../perlasm/x86asm.pl
+	(cd asm; $(PERL) x86-mont.pl a.out $(CFLAGS) > ../$@)
 
 sparcv8.o:	asm/sparcv8.S
 	$(CC) $(CFLAGS) -c asm/sparcv8.S
@@ -91,10 +97,15 @@ bn-mips3.o:	asm/mips3.s
 
 x86_64-gcc.o:	asm/x86_64-gcc.c
 	$(CC) $(CFLAGS) -c -o $@ asm/x86_64-gcc.c
+x86_64-mont.s:	asm/x86_64-mont.pl
+	$(PERL) asm/x86_64-mont.pl $@
 
 bn-ia64.s:	asm/ia64.S
 	$(CC) $(CFLAGS) -E asm/ia64.S > $@
 
+bn-s390x.s:	asm/s390x.S
+	$(CC) $(CFLAGS) -E asm/s390x.S > $@
+
 # GNU assembler fails to compile PA-RISC2 modules, insist on calling
 # vendor assembler...
 pa-risc2W.o: asm/pa-risc2W.s
@@ -109,6 +120,9 @@ aix_ppc32.s: asm/ppc.pl;	$(PERL) asm/ppc.pl $@
 aix_ppc64.s: asm/ppc.pl;	$(PERL) asm/ppc.pl $@
 osx_ppc32.s: asm/ppc.pl;	$(PERL) $< $@
 
+# GNU make "catch all"
+%-mont.s:	asm/%-mont.pl;  $(PERL) $< $(CFLAGS) > $@
+
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
 

crypto/bn/asm/s390x.S

@@ -0,0 +1,678 @@
+.ident "s390x.S, version 1.0"
+// ====================================================================
+// Written 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.
+// ====================================================================
+
+.text
+
+#define zero	%r0
+
+// BN_ULONG bn_mul_add_words(BN_ULONG *r2,BN_ULONG *r3,int r4,BN_ULONG r5);
+.globl	bn_mul_add_words
+.type	bn_mul_add_words,@function
+.align	4
+bn_mul_add_words:
+	lghi	zero,0		// zero = 0
+	la	%r1,0(%r2)	// put rp aside
+	lghi	%r2,0		// i=0;
+	ltgfr	%r4,%r4
+	bler	%r14		// if (len<=0) return 0;
+
+	stmg	%r6,%r10,48(%r15)
+	lghi	%r8,0		// carry = 0
+	srag	%r10,%r4,2	// cnt=len/4
+	jz	.Loop1_madd
+
+.Loop4_madd:
+	lg	%r7,0(%r2,%r3)	// ap[i]
+	mlgr	%r6,%r5		// *=w
+	algr	%r7,%r8		// +=carry
+	alcgr	%r6,zero
+	alg	%r7,0(%r2,%r1)	// +=rp[i]
+	alcgr	%r6,zero
+	stg	%r7,0(%r2,%r1)	// rp[i]=
+
+	lg	%r9,8(%r2,%r3)
+	mlgr	%r8,%r5
+	algr	%r9,%r6
+	alcgr	%r8,zero
+	alg	%r9,8(%r2,%r1)
+	alcgr	%r8,zero
+	stg	%r9,8(%r2,%r1)
+
+	lg	%r7,16(%r2,%r3)
+	mlgr	%r6,%r5
+	algr	%r7,%r8
+	alcgr	%r6,zero
+	alg	%r7,16(%r2,%r1)
+	alcgr	%r6,zero
+	stg	%r7,16(%r2,%r1)
+
+	lg	%r9,24(%r2,%r3)
+	mlgr	%r8,%r5
+	algr	%r9,%r6
+	alcgr	%r8,zero
+	alg	%r9,24(%r2,%r1)
+	alcgr	%r8,zero
+	stg	%r9,24(%r2,%r1)
+
+	la	%r2,32(%r2)	// i+=4
+	brct	%r10,.Loop4_madd
+
+	lghi	%r10,3
+	nr	%r4,%r10	// cnt=len%4
+	jz	.Lend_madd
+
+.Loop1_madd:
+	lg	%r7,0(%r2,%r3)	// ap[i]
+	mlgr	%r6,%r5		// *=w
+	algr	%r7,%r8		// +=carry
+	alcgr	%r6,zero
+	alg	%r7,0(%r2,%r1)	// +=rp[i]
+	alcgr	%r6,zero
+	stg	%r7,0(%r2,%r1)	// rp[i]=
+
+	lgr	%r8,%r6
+	la	%r2,8(%r2)	// i++
+	brct	%r4,.Loop1_madd
+
+.Lend_madd:
+	lgr	%r2,%r8
+	lmg	%r6,%r10,48(%r15)
+	br	%r14
+.size	bn_mul_add_words,.-bn_mul_add_words
+
+// BN_ULONG bn_mul_words(BN_ULONG *r2,BN_ULONG *r3,int r4,BN_ULONG r5);
+.globl	bn_mul_words
+.type	bn_mul_words,@function
+.align	4
+bn_mul_words:
+	lghi	zero,0		// zero = 0
+	la	%r1,0(%r2)	// put rp aside
+	lghi	%r2,0		// i=0;
+	ltgfr	%r4,%r4
+	bler	%r14		// if (len<=0) return 0;
+
+	stmg	%r6,%r10,48(%r15)
+	lghi	%r8,0		// carry = 0
+	srag	%r10,%r4,2	// cnt=len/4
+	jz	.Loop1_mul
+
+.Loop4_mul:
+	lg	%r7,0(%r2,%r3)	// ap[i]
+	mlgr	%r6,%r5		// *=w
+	algr	%r7,%r8		// +=carry
+	alcgr	%r6,zero
+	stg	%r7,0(%r2,%r1)	// rp[i]=
+
+	lg	%r9,8(%r2,%r3)
+	mlgr	%r8,%r5
+	algr	%r9,%r6
+	alcgr	%r8,zero
+	stg	%r9,8(%r2,%r1)
+
+	lg	%r7,16(%r2,%r3)
+	mlgr	%r6,%r5
+	algr	%r7,%r8
+	alcgr	%r6,zero
+	stg	%r7,16(%r2,%r1)
+
+	lg	%r9,24(%r2,%r3)
+	mlgr	%r8,%r5
+	algr	%r9,%r6
+	alcgr	%r8,zero
+	stg	%r9,24(%r2,%r1)
+
+	la	%r2,32(%r2)	// i+=4
+	brct	%r10,.Loop4_mul
+
+	lghi	%r10,3
+	nr	%r4,%r10	// cnt=len%4
+	jz	.Lend_mul
+
+.Loop1_mul:
+	lg	%r7,0(%r2,%r3)	// ap[i]
+	mlgr	%r6,%r5		// *=w
+	algr	%r7,%r8		// +=carry
+	alcgr	%r6,zero
+	stg	%r7,0(%r2,%r1)	// rp[i]=
+
+	lgr	%r8,%r6
+	la	%r2,8(%r2)	// i++
+	brct	%r4,.Loop1_mul
+
+.Lend_mul:
+	lgr	%r2,%r8
+	lmg	%r6,%r10,48(%r15)
+	br	%r14
+.size	bn_mul_words,.-bn_mul_words
+
+// void bn_sqr_words(BN_ULONG *r2,BN_ULONG *r2,int r4)
+.globl	bn_sqr_words
+.type	bn_sqr_words,@function
+.align	4
+bn_sqr_words:
+	ltgfr	%r4,%r4
+	bler	%r14
+
+	stmg	%r6,%r7,48(%r15)
+	srag	%r1,%r4,2	// cnt=len/4
+	jz	.Loop1_sqr
+
+.Loop4_sqr:
+	lg	%r7,0(%r3)
+	mlgr	%r6,%r7
+	stg	%r7,0(%r2)
+	stg	%r6,8(%r2)
+
+	lg	%r7,8(%r3)
+	mlgr	%r6,%r7
+	stg	%r7,16(%r2)
+	stg	%r6,24(%r2)
+
+	lg	%r7,16(%r3)
+	mlgr	%r6,%r7
+	stg	%r7,32(%r2)
+	stg	%r6,40(%r2)
+
+	lg	%r7,24(%r3)
+	mlgr	%r6,%r7
+	stg	%r7,48(%r2)
+	stg	%r6,56(%r2)
+
+	la	%r3,32(%r3)
+	la	%r2,64(%r2)
+	brct	%r1,.Loop4_sqr
+
+	lghi	%r1,3
+	nr	%r4,%r1		// cnt=len%4
+	jz	.Lend_sqr
+
+.Loop1_sqr:
+	lg	%r7,0(%r3)
+	mlgr	%r6,%r7
+	stg	%r7,0(%r2)
+	stg	%r6,8(%r2)
+
+	la	%r3,8(%r3)
+	la	%r2,16(%r2)
+	brct	%r4,.Loop1_sqr
+
+.Lend_sqr:
+	lmg	%r6,%r7,48(%r15)
+	br	%r14
+.size	bn_sqr_words,.-bn_sqr_words
+
+// BN_ULONG bn_div_words(BN_ULONG h,BN_ULONG l,BN_ULONG d);
+.globl	bn_div_words
+.type	bn_div_words,@function
+.align	4
+bn_div_words:
+	dlgr	%r2,%r4
+	lgr	%r2,%r3
+	br	%r14
+.size	bn_div_words,.-bn_div_words
+
+// BN_ULONG bn_add_words(BN_ULONG *r2,BN_ULONG *r3,BN_ULONG *r4,int r5);
+.globl	bn_add_words
+.type	bn_add_words,@function
+.align	4
+bn_add_words:
+	la	%r1,0(%r2)	// put rp aside
+	lghi	%r2,0		// i=0
+	ltgfr	%r5,%r5
+	bler	%r14		// if (len<=0) return 0;
+
+	stg	%r6,48(%r15)
+	lghi	%r6,3
+	nr	%r6,%r5		// len%4
+	sra	%r5,2		// len/4, use sra because it sets condition code
+	jz	.Loop1_add	// carry is incidentally cleared if branch taken
+	algr	%r2,%r2		// clear carry
+
+.Loop4_add:
+	lg	%r0,0(%r2,%r3)
+	alcg	%r0,0(%r2,%r4)
+	stg	%r0,0(%r2,%r1)
+	lg	%r0,8(%r2,%r3)
+	alcg	%r0,8(%r2,%r4)
+	stg	%r0,8(%r2,%r1)
+	lg	%r0,16(%r2,%r3)
+	alcg	%r0,16(%r2,%r4)
+	stg	%r0,16(%r2,%r1)
+	lg	%r0,24(%r2,%r3)
+	alcg	%r0,24(%r2,%r4)
+	stg	%r0,24(%r2,%r1)
+
+	la	%r2,32(%r2)	// i+=4
+	brct	%r5,.Loop4_add
+
+	la	%r6,1(%r6)	// see if len%4 is zero ...
+	brct	%r6,.Loop1_add	// without touching condition code:-)
+
+.Lexit_add:
+	lghi	%r2,0
+	alcgr	%r2,%r2
+	lg	%r6,48(%r15)
+	br	%r14
+
+.Loop1_add:
+	lg	%r0,0(%r2,%r3)
+	alcg	%r0,0(%r2,%r4)
+	stg	%r0,0(%r2,%r1)
+
+	la	%r2,8(%r2)	// i++
+	brct	%r6,.Loop1_add
+
+	j	.Lexit_add
+.size	bn_add_words,.-bn_add_words
+
+// BN_ULONG bn_sub_words(BN_ULONG *r2,BN_ULONG *r3,BN_ULONG *r4,int r5);
+.globl	bn_sub_words
+.type	bn_sub_words,@function
+.align	4
+bn_sub_words:
+	la	%r1,0(%r2)	// put rp aside
+	lghi	%r2,0		// i=0
+	ltgfr	%r5,%r5
+	bler	%r14		// if (len<=0) return 0;
+
+	stg	%r6,48(%r15)
+	lghi	%r6,3
+	nr	%r6,%r5		// len%4
+	sra	%r5,2		// len/4, use sra because it sets condition code
+	jnz	.Loop4_sub	// borrow is incidentally cleared if branch taken
+	slgr	%r2,%r2		// clear borrow
+
+.Loop1_sub:
+	lg	%r0,0(%r2,%r3)
+	slbg	%r0,0(%r2,%r4)
+	stg	%r0,0(%r2,%r1)
+
+	la	%r2,8(%r2)	// i++
+	brct	%r6,.Loop1_sub
+	j	.Lexit_sub
+
+.Loop4_sub:
+	lg	%r0,0(%r2,%r3)
+	slbg	%r0,0(%r2,%r4)
+	stg	%r0,0(%r2,%r1)
+	lg	%r0,8(%r2,%r3)
+	slbg	%r0,8(%r2,%r4)
+	stg	%r0,8(%r2,%r1)
+	lg	%r0,16(%r2,%r3)
+	slbg	%r0,16(%r2,%r4)
+	stg	%r0,16(%r2,%r1)
+	lg	%r0,24(%r2,%r3)
+	slbg	%r0,24(%r2,%r4)
+	stg	%r0,24(%r2,%r1)
+
+	la	%r2,32(%r2)	// i+=4
+	brct	%r5,.Loop4_sub
+
+	la	%r6,1(%r6)	// see if len%4 is zero ...
+	brct	%r6,.Loop1_sub	// without touching condition code:-)
+
+.Lexit_sub:
+	lghi	%r2,0
+	slbgr	%r2,%r2
+	lcgr	%r2,%r2
+	lg	%r6,48(%r15)
+	br	%r14
+.size	bn_sub_words,.-bn_sub_words
+
+#define c1	%r1
+#define c2	%r5
+#define c3	%r8
+
+#define mul_add_c(ai,bi,c1,c2,c3)	\
+	lg	%r7,ai*8(%r3);		\
+	mlg	%r6,bi*8(%r4);		\
+	algr	c1,%r7;			\
+	alcgr	c2,%r6;			\
+	alcgr	c3,zero
+
+// void bn_mul_comba8(BN_ULONG *r2,BN_ULONG *r3,BN_ULONG *r4);
+.globl	bn_mul_comba8
+.type	bn_mul_comba8,@function
+.align	4
+bn_mul_comba8:
+	stmg	%r6,%r8,48(%r15)
+
+	lghi	c1,0
+	lghi	c2,0
+	lghi	c3,0
+	lghi	zero,0
+
+	mul_add_c(0,0,c1,c2,c3);
+	stg	c1,0*8(%r2)
+	lghi	c1,0
+
+	mul_add_c(0,1,c2,c3,c1);
+	mul_add_c(1,0,c2,c3,c1);
+	stg	c2,1*8(%r2)
+	lghi	c2,0
+
+	mul_add_c(2,0,c3,c1,c2);
+	mul_add_c(1,1,c3,c1,c2);
+	mul_add_c(0,2,c3,c1,c2);
+	stg	c3,2*8(%r2)
+	lghi	c3,0
+
+	mul_add_c(0,3,c1,c2,c3);
+	mul_add_c(1,2,c1,c2,c3);
+	mul_add_c(2,1,c1,c2,c3);
+	mul_add_c(3,0,c1,c2,c3);
+	stg	c1,3*8(%r2)
+	lghi	c1,0
+
+	mul_add_c(4,0,c2,c3,c1);
+	mul_add_c(3,1,c2,c3,c1);
+	mul_add_c(2,2,c2,c3,c1);
+	mul_add_c(1,3,c2,c3,c1);
+	mul_add_c(0,4,c2,c3,c1);
+	stg	c2,4*8(%r2)
+	lghi	c2,0
+
+	mul_add_c(0,5,c3,c1,c2);
+	mul_add_c(1,4,c3,c1,c2);
+	mul_add_c(2,3,c3,c1,c2);
+	mul_add_c(3,2,c3,c1,c2);
+	mul_add_c(4,1,c3,c1,c2);
+	mul_add_c(5,0,c3,c1,c2);
+	stg	c3,5*8(%r2)
+	lghi	c3,0
+
+	mul_add_c(6,0,c1,c2,c3);
+	mul_add_c(5,1,c1,c2,c3);
+	mul_add_c(4,2,c1,c2,c3);
+	mul_add_c(3,3,c1,c2,c3);
+	mul_add_c(2,4,c1,c2,c3);
+	mul_add_c(1,5,c1,c2,c3);
+	mul_add_c(0,6,c1,c2,c3);
+	stg	c1,6*8(%r2)
+	lghi	c1,0
+
+	mul_add_c(0,7,c2,c3,c1);
+	mul_add_c(1,6,c2,c3,c1);
+	mul_add_c(2,5,c2,c3,c1);
+	mul_add_c(3,4,c2,c3,c1);
+	mul_add_c(4,3,c2,c3,c1);
+	mul_add_c(5,2,c2,c3,c1);
+	mul_add_c(6,1,c2,c3,c1);
+	mul_add_c(7,0,c2,c3,c1);
+	stg	c2,7*8(%r2)
+	lghi	c2,0
+
+	mul_add_c(7,1,c3,c1,c2);
+	mul_add_c(6,2,c3,c1,c2);
+	mul_add_c(5,3,c3,c1,c2);
+	mul_add_c(4,4,c3,c1,c2);
+	mul_add_c(3,5,c3,c1,c2);
+	mul_add_c(2,6,c3,c1,c2);
+	mul_add_c(1,7,c3,c1,c2);
+	stg	c3,8*8(%r2)
+	lghi	c3,0
+
+	mul_add_c(2,7,c1,c2,c3);
+	mul_add_c(3,6,c1,c2,c3);
+	mul_add_c(4,5,c1,c2,c3);
+	mul_add_c(5,4,c1,c2,c3);
+	mul_add_c(6,3,c1,c2,c3);
+	mul_add_c(7,2,c1,c2,c3);
+	stg	c1,9*8(%r2)
+	lghi	c1,0
+
+	mul_add_c(7,3,c2,c3,c1);
+	mul_add_c(6,4,c2,c3,c1);
+	mul_add_c(5,5,c2,c3,c1);
+	mul_add_c(4,6,c2,c3,c1);
+	mul_add_c(3,7,c2,c3,c1);
+	stg	c2,10*8(%r2)
+	lghi	c2,0
+
+	mul_add_c(4,7,c3,c1,c2);
+	mul_add_c(5,6,c3,c1,c2);
+	mul_add_c(6,5,c3,c1,c2);
+	mul_add_c(7,4,c3,c1,c2);
+	stg	c3,11*8(%r2)
+	lghi	c3,0
+
+	mul_add_c(7,5,c1,c2,c3);
+	mul_add_c(6,6,c1,c2,c3);
+	mul_add_c(5,7,c1,c2,c3);
+	stg	c1,12*8(%r2)
+	lghi	c1,0
+
+
+	mul_add_c(6,7,c2,c3,c1);
+	mul_add_c(7,6,c2,c3,c1);
+	stg	c2,13*8(%r2)
+	lghi	c2,0
+
+	mul_add_c(7,7,c3,c1,c2);
+	stg	c3,14*8(%r2)
+	stg	c1,15*8(%r2)
+
+	lmg	%r6,%r8,48(%r15)
+	br	%r14
+.size	bn_mul_comba8,.-bn_mul_comba8
+
+// void bn_mul_comba4(BN_ULONG *r2,BN_ULONG *r3,BN_ULONG *r4);
+.globl	bn_mul_comba4
+.type	bn_mul_comba4,@function
+.align	4
+bn_mul_comba4:
+	stmg	%r6,%r8,48(%r15)
+
+	lghi	c1,0
+	lghi	c2,0
+	lghi	c3,0
+	lghi	zero,0
+
+	mul_add_c(0,0,c1,c2,c3);
+	stg	c1,0*8(%r3)
+	lghi	c1,0
+
+	mul_add_c(0,1,c2,c3,c1);
+	mul_add_c(1,0,c2,c3,c1);
+	stg	c2,1*8(%r2)
+	lghi	c2,0
+
+	mul_add_c(2,0,c3,c1,c2);
+	mul_add_c(1,1,c3,c1,c2);
+	mul_add_c(0,2,c3,c1,c2);
+	stg	c3,2*8(%r2)
+	lghi	c3,0
+
+	mul_add_c(0,3,c1,c2,c3);
+	mul_add_c(1,2,c1,c2,c3);
+	mul_add_c(2,1,c1,c2,c3);
+	mul_add_c(3,0,c1,c2,c3);
+	stg	c1,3*8(%r2)
+	lghi	c1,0
+
+	mul_add_c(3,1,c2,c3,c1);
+	mul_add_c(2,2,c2,c3,c1);
+	mul_add_c(1,3,c2,c3,c1);
+	stg	c2,4*8(%r2)
+	lghi	c2,0
+
+	mul_add_c(2,3,c3,c1,c2);
+	mul_add_c(3,2,c3,c1,c2);
+	stg	c3,5*8(%r2)
+	lghi	c3,0
+
+	mul_add_c(3,3,c1,c2,c3);
+	stg	c1,6*8(%r2)
+	stg	c2,7*8(%r2)
+
+	stmg	%r6,%r8,48(%r15)
+	br	%r14
+.size	bn_mul_comba4,.-bn_mul_comba4
+
+#define sqr_add_c(ai,c1,c2,c3)		\
+	lg	%r7,ai*8(%r3);		\
+	mlgr	%r6,%r7;		\
+	algr	c1,%r7;			\
+	alcgr	c2,%r6;			\
+	alcgr	c3,zero
+
+#define sqr_add_c2(ai,aj,c1,c2,c3)	\
+	lg	%r7,ai*8(%r3);		\
+	mlg	%r6,aj*8(%r3);		\
+	algr	c1,%r7;			\
+	alcgr	c2,%r6;			\
+	alcgr	c3,zero;		\
+	algr	c1,%r7;			\
+	alcgr	c2,%r6;			\
+	alcgr	c3,zero
+
+// void bn_sqr_comba8(BN_ULONG *r2,BN_ULONG *r3);
+.globl	bn_sqr_comba8
+.type	bn_sqr_comba8,@function
+.align	4
+bn_sqr_comba8:
+	stmg	%r6,%r8,48(%r15)
+
+	lghi	c1,0
+	lghi	c2,0
+	lghi	c3,0
+	lghi	zero,0
+
+	sqr_add_c(0,c1,c2,c3);
+	stg	c1,0*8(%r2)
+	lghi	c1,0
+
+	sqr_add_c2(1,0,c2,c3,c1);
+	stg	c2,1*8(%r2)
+	lghi	c2,0
+
+	sqr_add_c(1,c3,c1,c2);
+	sqr_add_c2(2,0,c3,c1,c2);
+	stg	c3,2*8(%r2)
+	lghi	c3,0
+
+	sqr_add_c2(3,0,c1,c2,c3);
+	sqr_add_c2(2,1,c1,c2,c3);
+	stg	c1,3*8(%r2)
+	lghi	c1,0
+
+	sqr_add_c(2,c2,c3,c1);
+	sqr_add_c2(3,1,c2,c3,c1);
+	sqr_add_c2(4,0,c2,c3,c1);
+	stg	c2,4*8(%r2)
+	lghi	c2,0
+
+	sqr_add_c2(5,0,c3,c1,c2);
+	sqr_add_c2(4,1,c3,c1,c2);
+	sqr_add_c2(3,2,c3,c1,c2);
+	stg	c3,5*8(%r2)
+	lghi	c3,0
+
+	sqr_add_c(3,c1,c2,c3);
+	sqr_add_c2(4,2,c1,c2,c3);
+	sqr_add_c2(5,1,c1,c2,c3);
+	sqr_add_c2(6,0,c1,c2,c3);
+	stg	c1,6*8(%r2)
+	lghi	c1,0
+
+	sqr_add_c2(7,0,c2,c3,c1);
+	sqr_add_c2(6,1,c2,c3,c1);
+	sqr_add_c2(5,2,c2,c3,c1);
+	sqr_add_c2(4,3,c2,c3,c1);
+	stg	c2,7*8(%r2)
+	lghi	c2,0
+
+	sqr_add_c(4,c3,c1,c2);
+	sqr_add_c2(5,3,c3,c1,c2);
+	sqr_add_c2(6,2,c3,c1,c2);
+	sqr_add_c2(7,1,c3,c1,c2);
+	stg	c3,8*8(%r2)
+	lghi	c3,0
+
+	sqr_add_c2(7,2,c1,c2,c3);
+	sqr_add_c2(6,3,c1,c2,c3);
+	sqr_add_c2(5,4,c1,c2,c3);
+	stg	c1,9*8(%r2)
+	lghi	c1,0
+
+	sqr_add_c(5,c2,c3,c1);
+	sqr_add_c2(6,4,c2,c3,c1);
+	sqr_add_c2(7,3,c2,c3,c1);
+	stg	c2,10*8(%r2)
+	lghi	c2,0
+
+	sqr_add_c2(7,4,c3,c1,c2);
+	sqr_add_c2(6,5,c3,c1,c2);
+	stg	c3,11*8(%r2)
+	lghi	c3,0
+
+	sqr_add_c(6,c1,c2,c3);
+	sqr_add_c2(7,5,c1,c2,c3);
+	stg	c1,12*8(%r2)
+	lghi	c1,0
+
+	sqr_add_c2(7,6,c2,c3,c1);
+	stg	c2,13*8(%r2)
+	lghi	c2,0
+
+	sqr_add_c(7,c3,c1,c2);
+	stg	c3,14*8(%r2)
+	stg	c1,15*8(%r2)
+
+	lmg	%r6,%r8,48(%r15)
+	br	%r14
+.size	bn_sqr_comba8,.-bn_sqr_comba8
+
+// void bn_sqr_comba4(BN_ULONG *r2,BN_ULONG *r3);
+.globl bn_sqr_comba4
+.type	bn_sqr_comba4,@function
+.align	4
+bn_sqr_comba4:
+	stmg	%r6,%r8,48(%r15)
+
+	lghi	c1,0
+	lghi	c2,0
+	lghi	c3,0
+	lghi	zero,0
+
+	sqr_add_c(0,c1,c2,c3);
+	stg	c1,0*8(%r2)
+	lghi	c1,0
+
+	sqr_add_c2(1,0,c2,c3,c1);
+	stg	c2,1*8(%r2)
+	lghi	c2,0
+
+	sqr_add_c(1,c3,c1,c2);
+	sqr_add_c2(2,0,c3,c1,c2);
+	stg	c3,2*8(%r2)
+	lghi	c3,0
+
+	sqr_add_c2(3,0,c1,c2,c3);
+	sqr_add_c2(2,1,c1,c2,c3);
+	stg	c1,3*8(%r2)
+	lghi	c1,0
+
+	sqr_add_c(2,c2,c3,c1);
+	sqr_add_c2(3,1,c2,c3,c1);
+	stg	c2,4*8(%r2)
+	lghi	c2,0
+
+	sqr_add_c2(3,2,c3,c1,c2);
+	stg	c3,5*8(%r2)
+	lghi	c3,0
+
+	sqr_add_c(3,c1,c2,c3);
+	stg	c1,6*8(%r2)
+	stg	c2,7*8(%r2)
+
+	lmg	%r6,%r8,48(%r15)
+	br	%r14
+.size	bn_sqr_comba4,.-bn_sqr_comba4

crypto/bn/asm/x86-mont.pl

@@ -0,0 +1,591 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# October 2005
+#
+# This is a "teaser" code, as it can be improved in several ways...
+# First of all non-SSE2 path should be implemented (yes, for now it
+# performs Montgomery multiplication/convolution only on SSE2-capable
+# CPUs such as P4, others fall down to original code). Then inner loop
+# can be unrolled and modulo-scheduled to improve ILP and possibly
+# moved to 128-bit XMM register bank (though it would require input
+# rearrangement and/or increase bus bandwidth utilization). Dedicated
+# squaring procedure should give further performance improvement...
+# Yet, for being draft, the code improves rsa512 *sign* benchmark by
+# 110%(!), rsa1024 one - by 70% and rsa4096 - by 20%:-)
+
+# December 2006
+#
+# Modulo-scheduling SSE2 loops results in further 15-20% improvement.
+# Integer-only code [being equipped with dedicated squaring procedure]
+# gives ~40% on rsa512 sign benchmark...
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+push(@INC,"${dir}","${dir}../../perlasm");
+require "x86asm.pl";
+
+&asm_init($ARGV[0],$0);
+
+$sse2=0;
+for (@ARGV) { $sse2=1 if (/-DOPENSSL_IA32_SSE2/); }
+
+&external_label("OPENSSL_ia32cap_P") if ($sse2);
+
+&function_begin("bn_mul_mont");
+
+$i="edx";
+$j="ecx";
+$ap="esi";	$tp="esi";		# overlapping variables!!!
+$rp="edi";	$bp="edi";		# overlapping variables!!!
+$np="ebp";
+$num="ebx";
+
+$_num=&DWP(4*0,"esp");			# stack top layout
+$_rp=&DWP(4*1,"esp");
+$_ap=&DWP(4*2,"esp");
+$_bp=&DWP(4*3,"esp");
+$_np=&DWP(4*4,"esp");
+$_n0=&DWP(4*5,"esp");	$_n0q=&QWP(4*5,"esp");
+$_sp=&DWP(4*6,"esp");
+$_bpend=&DWP(4*7,"esp");
+$frame=32;				# size of above frame rounded up to 16n
+
+	&xor	("eax","eax");
+	&mov	("edi",&wparam(5));	# int num
+	&cmp	("edi",4);
+	&jl	(&label("just_leave"));
+
+	&lea	("esi",&wparam(0));	# put aside pointer to argument block
+	&lea	("edx",&wparam(1));	# load ap
+	&mov	("ebp","esp");		# saved stack pointer!
+	&add	("edi",2);		# extra two words on top of tp
+	&neg	("edi");
+	&lea	("esp",&DWP(-$frame,"esp","edi",4));	# alloca($frame+4*(num+2))
+	&neg	("edi");
+
+	# minimize cache contention by arraning 2K window between stack
+	# pointer and ap argument [np is also position sensitive vector,
+	# but it's assumed to be near ap, as it's allocated at ~same
+	# time].
+	&mov	("eax","esp");
+	&sub	("eax","edx");
+	&and	("eax",2047);
+	&sub	("esp","eax");		# this aligns sp and ap modulo 2048
+
+	&xor	("edx","esp");
+	&and	("edx",2048);
+	&xor	("edx",2048);
+	&sub	("esp","edx");		# this splits them apart modulo 4096
+
+	&and	("esp",-64);		# align to cache line
+
+	################################# load argument block...
+	&mov	("eax",&DWP(0*4,"esi"));# BN_ULONG *rp
+	&mov	("ebx",&DWP(1*4,"esi"));# const BN_ULONG *ap
+	&mov	("ecx",&DWP(2*4,"esi"));# const BN_ULONG *bp
+	&mov	("edx",&DWP(3*4,"esi"));# const BN_ULONG *np
+	&mov	("esi",&DWP(4*4,"esi"));# const BN_ULONG *n0
+	#&mov	("edi",&DWP(5*4,"esi"));# int num
+
+	&mov	("esi",&DWP(0,"esi"));	# pull n0[0]
+	&mov	($_rp,"eax");		# ... save a copy of argument block
+	&mov	($_ap,"ebx");
+	&mov	($_bp,"ecx");
+	&mov	($_np,"edx");
+	&mov	($_n0,"esi");
+	&lea	($num,&DWP(-3,"edi"));	# num=num-1 to assist modulo-scheduling
+	#&mov	($_num,$num);		# redundant as $num is not reused
+	&mov	($_sp,"ebp");		# saved stack pointer!
+
+if($sse2) {
+$acc0="mm0";	# mmx register bank layout
+$acc1="mm1";
+$car0="mm2";
+$car1="mm3";
+$mul0="mm4";
+$mul1="mm5";
+$temp="mm6";
+$mask="mm7";
+
+	&picmeup("eax","OPENSSL_ia32cap_P");
+	&bt	(&DWP(0,"eax"),26);
+	&jnc	(&label("non_sse2"));
+
+	&mov	("eax",-1);
+	&movd	($mask,"eax");		# mask 32 lower bits
+
+	&mov	($ap,$_ap);		# load input pointers
+	&mov	($bp,$_bp);
+	&mov	($np,$_np);
+
+	&xor	($i,$i);		# i=0
+	&xor	($j,$j);		# j=0
+
+	&movd	($mul0,&DWP(0,$bp));		# bp[0]
+	&movd	($mul1,&DWP(0,$ap));		# ap[0]
+	&movd	($car1,&DWP(0,$np));		# np[0]
+
+	&pmuludq($mul1,$mul0);			# ap[0]*bp[0]
+	&movq	($car0,$mul1);
+	&movq	($acc0,$mul1);			# I wish movd worked for
+	&pand	($acc0,$mask);			# inter-register transfers
+
+	&pmuludq($mul1,$_n0q);			# *=n0
+
+	&pmuludq($car1,$mul1);			# "t[0]"*np[0]*n0
+	&paddq	($car1,$acc0);
+
+	&movd	($acc1,&DWP(4,$np));		# np[1]
+	&movd	($acc0,&DWP(4,$ap));		# ap[1]
+
+	&psrlq	($car0,32);
+	&psrlq	($car1,32);
+
+	&inc	($j);				# j++
+&set_label("1st",16);
+	&pmuludq($acc0,$mul0);			# ap[j]*bp[0]
+	&pmuludq($acc1,$mul1);			# np[j]*m1
+	&paddq	($car0,$acc0);			# +=c0
+	&paddq	($car1,$acc1);			# +=c1
+
+	&movq	($acc0,$car0);
+	&pand	($acc0,$mask);
+	&movd	($acc1,&DWP(4,$np,$j,4));	# np[j+1]
+	&paddq	($car1,$acc0);			# +=ap[j]*bp[0];
+	&movd	($acc0,&DWP(4,$ap,$j,4));	# ap[j+1]
+	&psrlq	($car0,32);
+	&movd	(&DWP($frame-4,"esp",$j,4),$car1);	# tp[j-1]=
+	&psrlq	($car1,32);
+
+	&lea	($j,&DWP(1,$j));
+	&cmp	($j,$num);
+	&jl	(&label("1st"));
+
+	&pmuludq($acc0,$mul0);			# ap[num-1]*bp[0]
+	&pmuludq($acc1,$mul1);			# np[num-1]*m1
+	&paddq	($car0,$acc0);			# +=c0
+	&paddq	($car1,$acc1);			# +=c1
+
+	&movq	($acc0,$car0);
+	&pand	($acc0,$mask);
+	&paddq	($car1,$acc0);			# +=ap[num-1]*bp[0];
+	&movd	(&DWP($frame-4,"esp",$j,4),$car1);	# tp[num-2]=
+
+	&psrlq	($car0,32);
+	&psrlq	($car1,32);
+
+	&paddq	($car1,$car0);
+	&movq	(&QWP($frame,"esp",$num,4),$car1);	# tp[num].tp[num-1]
+
+	&inc	($i);				# i++
+&set_label("outer");
+	&xor	($j,$j);			# j=0
+
+	&movd	($mul0,&DWP(0,$bp,$i,4));	# bp[i]
+	&movd	($mul1,&DWP(0,$ap));		# ap[0]
+	&movd	($temp,&DWP($frame,"esp"));	# tp[0]
+	&movd	($car1,&DWP(0,$np));		# np[0]
+	&pmuludq($mul1,$mul0);			# ap[0]*bp[i]
+
+	&paddq	($mul1,$temp);			# +=tp[0]
+	&movq	($acc0,$mul1);
+	&movq	($car0,$mul1);
+	&pand	($acc0,$mask);
+
+	&pmuludq($mul1,$_n0q);			# *=n0
+
+	&pmuludq($car1,$mul1);
+	&paddq	($car1,$acc0);
+
+	&movd	($temp,&DWP($frame+4,"esp"));	# tp[1]
+	&movd	($acc1,&DWP(4,$np));		# np[1]
+	&movd	($acc0,&DWP(4,$ap));		# ap[1]
+
+	&psrlq	($car0,32);
+	&psrlq	($car1,32);
+	&paddq	($car0,$temp);			# +=tp[1]
+
+	&inc	($j);				# j++
+	&dec	($num);
+&set_label("inner");
+	&pmuludq($acc0,$mul0);			# ap[j]*bp[i]
+	&pmuludq($acc1,$mul1);			# np[j]*m1
+	&paddq	($car0,$acc0);			# +=c0
+	&paddq	($car1,$acc1);			# +=c1
+
+	&movq	($acc0,$car0);
+	&movd	($temp,&DWP($frame+4,"esp",$j,4));# tp[j+1]
+	&pand	($acc0,$mask);
+	&movd	($acc1,&DWP(4,$np,$j,4));	# np[j+1]
+	&paddq	($car1,$acc0);			# +=ap[j]*bp[i]+tp[j]
+	&movd	($acc0,&DWP(4,$ap,$j,4));	# ap[j+1]
+	&psrlq	($car0,32);
+	&movd	(&DWP($frame-4,"esp",$j,4),$car1);# tp[j-1]=
+	&psrlq	($car1,32);
+	&paddq	($car0,$temp);			# +=tp[j+1]
+
+	&dec	($num);
+	&lea	($j,&DWP(1,$j));		# j++
+	&jnz	(&label("inner"));
+
+	&mov	($num,$j);
+	&pmuludq($acc0,$mul0);			# ap[num-1]*bp[i]
+	&pmuludq($acc1,$mul1);			# np[num-1]*m1
+	&paddq	($car0,$acc0);			# +=c0
+	&paddq	($car1,$acc1);			# +=c1
+
+	&movq	($acc0,$car0);
+	&pand	($acc0,$mask);
+	&paddq	($car1,$acc0);			# +=ap[num-1]*bp[i]+tp[num-1]
+	&movd	(&DWP($frame-4,"esp",$j,4),$car1);	# tp[num-2]=
+	&psrlq	($car0,32);
+	&psrlq	($car1,32);
+
+	&movd	($temp,&DWP($frame+4,"esp",$num,4));	# += tp[num]
+	&paddq	($car1,$car0);
+	&paddq	($car1,$temp);
+	&movq	(&QWP($frame,"esp",$num,4),$car1);	# tp[num].tp[num-1]
+
+	&lea	($i,&DWP(1,$i));		# i++
+	&cmp	($i,$num);
+	&jle	(&label("outer"));
+
+	&emms	();				# done with mmx bank
+	&jmp	(&label("common_tail"));
+
+&set_label("non_sse2",16);
+}
+
+if (0) {
+	&mov	("esp",$_sp);
+	&xor	("eax","eax");	# signal "not fast enough [yet]"
+	&jmp	(&label("just_leave"));
+	# While the below code provides competitive performance for
+	# all key lengthes on modern Intel cores, it's still more
+	# than 10% slower for 4096-bit key elsewhere:-( "Competitive"
+	# means compared to the original integer-only assembler.
+	# 512-bit RSA sign is better by ~40%, but that's about all
+	# one can say about all CPUs...
+} else {
+$inp="esi";	# integer path uses these registers differently
+$word="edi";
+$carry="ebp";
+
+	&mov	($inp,$_ap);
+	&lea	($carry,&DWP(1,$num));
+	&mov	($word,$_bp);
+	&xor	($j,$j);				# j=0
+	&mov	("edx",$inp);
+	&and	($carry,1);				# see if num is even
+	&sub	("edx",$word);				# see if ap==bp
+	&lea	("eax",&DWP(4,$word,$num,4));		# &bp[num]
+	&or	($carry,"edx");
+	&mov	($word,&DWP(0,$word));			# bp[0]
+	&jz	(&label("bn_sqr_mont"));
+	&mov	($_bpend,"eax");
+	&mov	("eax",&DWP(0,$inp));
+	&xor	("edx","edx");
+
+&set_label("mull",16);
+	&mov	($carry,"edx");
+	&mul	($word);				# ap[j]*bp[0]
+	&add	($carry,"eax");
+	&lea	($j,&DWP(1,$j));
+	&adc	("edx",0);
+	&mov	("eax",&DWP(0,$inp,$j,4));		# ap[j+1]
+	&cmp	($j,$num);
+	&mov	(&DWP($frame-4,"esp",$j,4),$carry);	# tp[j]=
+	&jl	(&label("mull"));
+
+	&mov	($carry,"edx");
+	&mul	($word);				# ap[num-1]*bp[0]
+	 &mov	($word,$_n0);
+	&add	("eax",$carry);
+	 &mov	($inp,$_np);
+	&adc	("edx",0);
+	 &imul	($word,&DWP($frame,"esp"));		# n0*tp[0]
+
+	&mov	(&DWP($frame,"esp",$num,4),"eax");	# tp[num-1]=
+	&xor	($j,$j);
+	&mov	(&DWP($frame+4,"esp",$num,4),"edx");	# tp[num]=
+	&mov	(&DWP($frame+8,"esp",$num,4),$j);	# tp[num+1]=
+
+	&mov	("eax",&DWP(0,$inp));			# np[0]
+	&mul	($word);				# np[0]*m
+	&add	("eax",&DWP($frame,"esp"));		# +=tp[0]
+	&mov	("eax",&DWP(4,$inp));			# np[1]
+	&adc	("edx",0);
+	&inc	($j);
+
+	&jmp	(&label("2ndmadd"));
+
+&set_label("1stmadd",16);
+	&mov	($carry,"edx");
+	&mul	($word);				# ap[j]*bp[i]
+	&add	($carry,&DWP($frame,"esp",$j,4));	# +=tp[j]
+	&lea	($j,&DWP(1,$j));
+	&adc	("edx",0);
+	&add	($carry,"eax");
+	&mov	("eax",&DWP(0,$inp,$j,4));		# ap[j+1]
+	&adc	("edx",0);
+	&cmp	($j,$num);
+	&mov	(&DWP($frame-4,"esp",$j,4),$carry);	# tp[j]=
+	&jl	(&label("1stmadd"));
+
+	&mov	($carry,"edx");
+	&mul	($word);				# ap[num-1]*bp[i]
+	&add	("eax",&DWP($frame,"esp",$num,4));	# +=tp[num-1]
+	 &mov	($word,$_n0);
+	&adc	("edx",0);
+	 &mov	($inp,$_np);
+	&add	($carry,"eax");
+	&adc	("edx",0);
+	 &imul	($word,&DWP($frame,"esp"));		# n0*tp[0]
+
+	&xor	($j,$j);
+	&add	("edx",&DWP($frame+4,"esp",$num,4));	# carry+=tp[num]
+	&mov	(&DWP($frame,"esp",$num,4),$carry);	# tp[num-1]=
+	&adc	($j,0);
+	 &mov	("eax",&DWP(0,$inp));			# np[0]
+	&mov	(&DWP($frame+4,"esp",$num,4),"edx");	# tp[num]=
+	&mov	(&DWP($frame+8,"esp",$num,4),$j);	# tp[num+1]=
+
+	&mul	($word);				# np[0]*m
+	&add	("eax",&DWP($frame,"esp"));		# +=tp[0]
+	&mov	("eax",&DWP(4,$inp));			# np[1]
+	&adc	("edx",0);
+	&mov	($j,1);
+
+&set_label("2ndmadd",16);
+	&mov	($carry,"edx");
+	&mul	($word);				# np[j]*m
+	&add	($carry,&DWP($frame,"esp",$j,4));	# +=tp[j]
+	&lea	($j,&DWP(1,$j));
+	&adc	("edx",0);
+	&add	($carry,"eax");
+	&mov	("eax",&DWP(0,$inp,$j,4));		# np[j+1]
+	&adc	("edx",0);
+	&cmp	($j,$num);
+	&mov	(&DWP($frame-8,"esp",$j,4),$carry);	# tp[j-1]=
+	&jl	(&label("2ndmadd"));
+
+	&mov	($carry,"edx");
+	&mul	($word);				# np[j]*m
+	&add	($carry,&DWP($frame,"esp",$num,4));	# +=tp[num-1]
+	&adc	("edx",0);
+	&add	($carry,"eax");
+	&adc	("edx",0);
+	&mov	(&DWP($frame-4,"esp",$num,4),$carry);	# tp[num-2]=
+
+	&xor	("eax","eax");
+	 &mov	($j,$_bp);				# &bp[i]
+	&add	("edx",&DWP($frame+4,"esp",$num,4));	# carry+=tp[num]
+	&adc	("eax",&DWP($frame+8,"esp",$num,4));	# +=tp[num+1]
+	 &lea	($j,&DWP(4,$j));
+	&mov	(&DWP($frame,"esp",$num,4),"edx");	# tp[num-1]=
+	 &cmp	($j,$_bpend);
+	&mov	(&DWP($frame+4,"esp",$num,4),"eax");	# tp[num]=
+	&je	(&label("common_tail"));
+
+	&mov	($word,&DWP(0,$j));			# bp[i+1]
+	&mov	($inp,$_ap);
+	&mov	($_bp,$j);				# &bp[++i]
+	&xor	($j,$j);
+	&xor	("edx","edx");
+	&mov	("eax",&DWP(0,$inp));
+	&jmp	(&label("1stmadd"));
+
+&set_label("bn_sqr_mont",16);
+$sbit=$num;
+	&mov	($_num,$num);
+	&mov	($_bp,$j);				# i=0
+
+	&mov	("eax",$word);				# ap[0]
+	&mul	($word);				# ap[0]*ap[0]
+	&mov	(&DWP($frame,"esp"),"eax");		# tp[0]=
+	&mov	($sbit,"edx");
+	&shr	("edx",1);
+	&and	($sbit,1);
+	&inc	($j);
+&set_label("sqr",16);
+	&mov	("eax",&DWP(0,$inp,$j,4));		# ap[j]
+	&mov	($carry,"edx");
+	&mul	($word);				# ap[j]*ap[0]
+	&add	("eax",$carry);
+	&lea	($j,&DWP(1,$j));
+	&adc	("edx",0);
+	&lea	($carry,&DWP(0,$sbit,"eax",2));
+	&shr	("eax",31);
+	&cmp	($j,$_num);
+	&mov	($sbit,"eax");
+	&mov	(&DWP($frame-4,"esp",$j,4),$carry);	# tp[j]=
+	&jl	(&label("sqr"));
+
+	&mov	("eax",&DWP(0,$inp,$j,4));		# ap[num-1]
+	&mov	($carry,"edx");
+	&mul	($word);				# ap[num-1]*ap[0]
+	&add	("eax",$carry);
+	 &mov	($word,$_n0);
+	&adc	("edx",0);
+	 &mov	($inp,$_np);
+	&lea	($carry,&DWP(0,$sbit,"eax",2));
+	 &imul	($word,&DWP($frame,"esp"));		# n0*tp[0]
+	&shr	("eax",31);
+	&mov	(&DWP($frame,"esp",$j,4),$carry);	# tp[num-1]=
+
+	&lea	($carry,&DWP(0,"eax","edx",2));
+	 &mov	("eax",&DWP(0,$inp));			# np[0]
+	&shr	("edx",31);
+	&mov	(&DWP($frame+4,"esp",$j,4),$carry);	# tp[num]=
+	&mov	(&DWP($frame+8,"esp",$j,4),"edx");	# tp[num+1]=
+
+	&mul	($word);				# np[0]*m
+	&add	("eax",&DWP($frame,"esp"));		# +=tp[0]
+	&mov	($num,$j);
+	&adc	("edx",0);
+	&mov	("eax",&DWP(4,$inp));			# np[1]
+	&mov	($j,1);
+
+&set_label("3rdmadd",16);
+	&mov	($carry,"edx");
+	&mul	($word);				# np[j]*m
+	&add	($carry,&DWP($frame,"esp",$j,4));	# +=tp[j]
+	&adc	("edx",0);
+	&add	($carry,"eax");
+	&mov	("eax",&DWP(4,$inp,$j,4));		# np[j+1]
+	&adc	("edx",0);
+	&mov	(&DWP($frame-4,"esp",$j,4),$carry);	# tp[j-1]=
+
+	&mov	($carry,"edx");
+	&mul	($word);				# np[j+1]*m
+	&add	($carry,&DWP($frame+4,"esp",$j,4));	# +=tp[j+1]
+	&lea	($j,&DWP(2,$j));
+	&adc	("edx",0);
+	&add	($carry,"eax");
+	&mov	("eax",&DWP(0,$inp,$j,4));		# np[j+2]
+	&adc	("edx",0);
+	&cmp	($j,$num);
+	&mov	(&DWP($frame-8,"esp",$j,4),$carry);	# tp[j]=
+	&jl	(&label("3rdmadd"));
+
+	&mov	($carry,"edx");
+	&mul	($word);				# np[j]*m
+	&add	($carry,&DWP($frame,"esp",$num,4));	# +=tp[num-1]
+	&adc	("edx",0);
+	&add	($carry,"eax");
+	&adc	("edx",0);
+	&mov	(&DWP($frame-4,"esp",$num,4),$carry);	# tp[num-2]=
+
+	&mov	($j,$_bp);				# i
+	&xor	("eax","eax");
+	&mov	($inp,$_ap);
+	&add	("edx",&DWP($frame+4,"esp",$num,4));	# carry+=tp[num]
+	&adc	("eax",&DWP($frame+8,"esp",$num,4));	# +=tp[num+1]
+	&mov	(&DWP($frame,"esp",$num,4),"edx");	# tp[num-1]=
+	&cmp	($j,$num);
+	&mov	(&DWP($frame+4,"esp",$num,4),"eax");	# tp[num]=
+	&je	(&label("common_tail"));
+
+	&mov	($word,&DWP(4,$inp,$j,4));		# ap[i]
+	&lea	($j,&DWP(1,$j));
+	&mov	("eax",$word);
+	&mov	($_bp,$j);				# ++i
+	&mul	($word);				# ap[i]*ap[i]
+	&add	("eax",&DWP($frame,"esp",$j,4));	# +=tp[i]
+	&adc	("edx",0);
+	&mov	(&DWP($frame,"esp",$j,4),"eax");	# tp[i]=
+	&xor	($carry,$carry);
+	&cmp	($j,$num);
+	&lea	($j,&DWP(1,$j));
+	&je	(&label("sqrlast"));
+
+	&mov	($sbit,"edx");				# zaps $num
+	&shr	("edx",1);
+	&and	($sbit,1);
+&set_label("sqradd",16);
+	&mov	("eax",&DWP(0,$inp,$j,4));		# ap[j]
+	&mov	($carry,"edx");
+	&mul	($word);				# ap[j]*ap[i]
+	&add	("eax",$carry);
+	&lea	($carry,&DWP(0,"eax","eax"));
+	&adc	("edx",0);
+	&shr	("eax",31);
+	&add	($carry,&DWP($frame,"esp",$j,4));	# +=tp[j]
+	&lea	($j,&DWP(1,$j));
+	&adc	("eax",0);
+	&add	($carry,$sbit);
+	&adc	("eax",0);
+	&cmp	($j,$_num);
+	&mov	(&DWP($frame-4,"esp",$j,4),$carry);	# tp[j]=
+	&mov	($sbit,"eax");
+	&jle	(&label("sqradd"));
+
+	&mov	($carry,"edx");
+	&lea	("edx",&DWP(0,$sbit,"edx",2));
+	&shr	($carry,31);
+&set_label("sqrlast");
+	&mov	($word,$_n0);
+	&mov	($inp,$_np);
+	&imul	($word,&DWP($frame,"esp"));		# n0*tp[0]
+
+	&add	("edx",&DWP($frame,"esp",$j,4));	# +=tp[num]
+	&mov	("eax",&DWP(0,$inp));			# np[0]
+	&adc	($carry,0);
+	&mov	(&DWP($frame,"esp",$j,4),"edx");	# tp[num]=
+	&mov	(&DWP($frame+4,"esp",$j,4),$carry);	# tp[num+1]=
+
+	&mul	($word);				# np[0]*m
+	&add	("eax",&DWP($frame,"esp"));		# +=tp[0]
+	&lea	($num,&DWP(-1,$j));
+	&adc	("edx",0);
+	&mov	($j,1);
+	&mov	("eax",&DWP(4,$inp));			# np[1]
+
+	&jmp	(&label("3rdmadd"));
+}
+
+&set_label("common_tail",16);
+	&mov	($np,$_np);			# load modulus pointer
+	&mov	($rp,$_rp);			# load result pointer
+	&lea	($tp,&DWP($frame,"esp"));	# [$ap and $bp are zapped]
+
+	&mov	("eax",&DWP(0,$tp));		# tp[0]
+	&mov	($j,$num);			# j=num-1
+	&xor	($i,$i);			# i=0 and clear CF!
+
+&set_label("sub",16);
+	&sbb	("eax",&DWP(0,$np,$i,4));
+	&mov	(&DWP(0,$rp,$i,4),"eax");	# rp[i]=tp[i]-np[i]
+	&dec	($j);				# doesn't affect CF!
+	&mov	("eax",&DWP(4,$tp,$i,4));	# tp[i+1]
+	&lea	($i,&DWP(1,$i));		# i++
+	&jge	(&label("sub"));
+
+	&sbb	("eax",0);			# handle upmost overflow bit
+	&and	($tp,"eax");
+	&not	("eax");
+	&mov	($np,$rp);
+	&and	($np,"eax");
+	&or	($tp,$np);			# tp=carry?tp:rp
+
+&set_label("copy",16);				# copy or in-place refresh
+	&mov	("eax",&DWP(0,$tp,$num,4));
+	&mov	(&DWP(0,$rp,$num,4),"eax");	# rp[i]=tp[i]
+	&mov	(&DWP($frame,"esp",$num,4),$j);	# zap temporary vector
+	&dec	($num);
+	&jge	(&label("copy"));
+
+	&mov	("esp",$_sp);		# pull saved stack pointer
+	&mov	("eax",1);
+&set_label("just_leave");
+&function_end("bn_mul_mont");
+
+&asciz("Montgomery Multiplication for x86, CRYPTOGAMS by <appro\@openssl.org>");
+
+&asm_finish();

crypto/bn/asm/x86_64-mont.pl

@@ -0,0 +1,213 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# October 2005.
+#
+# Montgomery multiplication routine for x86_64. While it gives modest
+# 9% improvement of rsa4096 sign on Opteron, rsa512 sign runs more
+# than twice, >2x, as fast. Most common rsa1024 sign is improved by
+# respectful 50%. It remains to be seen if loop unrolling and
+# dedicated squaring routine can provide further improvement...
+
+$output=shift;
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open STDOUT,"| $^X $xlate $output";
+
+# int bn_mul_mont(
+$rp="%rdi";	# BN_ULONG *rp,
+$ap="%rsi";	# const BN_ULONG *ap,
+$bp="%rdx";	# const BN_ULONG *bp,
+$np="%rcx";	# const BN_ULONG *np,
+$n0="%r8";	# const BN_ULONG *n0,
+$num="%r9";	# int num);
+$lo0="%r10";
+$hi0="%r11";
+$bp="%r12";	# reassign $bp
+$hi1="%r13";
+$i="%r14";
+$j="%r15";
+$m0="%rbx";
+$m1="%rbp";
+
+$code=<<___;
+.text
+
+.globl	bn_mul_mont
+.type	bn_mul_mont,\@function,6
+.align	16
+bn_mul_mont:
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+
+	lea	2($num),%rax
+	mov	%rsp,%rbp
+	neg	%rax
+	lea	(%rsp,%rax,8),%rsp	# tp=alloca(8*(num+2))
+	and	\$-1024,%rsp		# minimize TLB usage
+
+	mov	%rbp,8(%rsp,$num,8)	# tp[num+1]=%rsp
+	mov	%rdx,$bp		# $bp reassigned, remember?
+
+	mov	($n0),$n0		# pull n0[0] value
+
+	xor	$i,$i			# i=0
+	xor	$j,$j			# j=0
+
+	mov	($bp),$m0		# m0=bp[0]
+	mov	($ap),%rax
+	mulq	$m0			# ap[0]*bp[0]
+	mov	%rax,$lo0
+	mov	%rdx,$hi0
+
+	imulq	$n0,%rax		# "tp[0]"*n0
+	mov	%rax,$m1
+
+	mulq	($np)			# np[0]*m1
+	add	$lo0,%rax		# discarded
+	adc	\$0,%rdx
+	mov	%rdx,$hi1
+
+	lea	1($j),$j		# j++
+.L1st:
+	mov	($ap,$j,8),%rax
+	mulq	$m0			# ap[j]*bp[0]
+	add	$hi0,%rax
+	adc	\$0,%rdx
+	mov	%rax,$lo0
+	mov	($np,$j,8),%rax
+	mov	%rdx,$hi0
+
+	mulq	$m1			# np[j]*m1
+	add	$hi1,%rax
+	lea	1($j),$j		# j++
+	adc	\$0,%rdx
+	add	$lo0,%rax		# np[j]*m1+ap[j]*bp[0]
+	adc	\$0,%rdx
+	mov	%rax,-16(%rsp,$j,8)	# tp[j-1]
+	cmp	$num,$j
+	mov	%rdx,$hi1
+	jl	.L1st
+
+	xor	%rdx,%rdx
+	add	$hi0,$hi1
+	adc	\$0,%rdx
+	mov	$hi1,-8(%rsp,$num,8)
+	mov	%rdx,(%rsp,$num,8)	# store upmost overflow bit
+
+	lea	1($i),$i		# i++
+.align	4
+.Louter:
+	xor	$j,$j			# j=0
+
+	mov	($bp,$i,8),$m0		# m0=bp[i]
+	mov	($ap),%rax		# ap[0]
+	mulq	$m0			# ap[0]*bp[i]
+	add	(%rsp),%rax		# ap[0]*bp[i]+tp[0]
+	adc	\$0,%rdx
+	mov	%rax,$lo0
+	mov	%rdx,$hi0
+
+	imulq	$n0,%rax		# tp[0]*n0
+	mov	%rax,$m1
+
+	mulq	($np,$j,8)		# np[0]*m1
+	add	$lo0,%rax		# discarded
+	mov	8(%rsp),$lo0		# tp[1]
+	adc	\$0,%rdx
+	mov	%rdx,$hi1
+
+	lea	1($j),$j		# j++
+.align	4
+.Linner:
+	mov	($ap,$j,8),%rax
+	mulq	$m0			# ap[j]*bp[i]
+	add	$hi0,%rax
+	adc	\$0,%rdx
+	add	%rax,$lo0		# ap[j]*bp[i]+tp[j]
+	mov	($np,$j,8),%rax
+	adc	\$0,%rdx
+	mov	%rdx,$hi0
+
+	mulq	$m1			# np[j]*m1
+	add	$hi1,%rax
+	lea	1($j),$j		# j++
+	adc	\$0,%rdx
+	add	$lo0,%rax		# np[j]*m1+ap[j]*bp[i]+tp[j]
+	adc	\$0,%rdx
+	mov	(%rsp,$j,8),$lo0
+	cmp	$num,$j
+	mov	%rax,-16(%rsp,$j,8)	# tp[j-1]
+	mov	%rdx,$hi1
+	jl	.Linner
+
+	xor	%rdx,%rdx
+	add	$hi0,$hi1
+	adc	\$0,%rdx
+	add	$lo0,$hi1		# pull upmost overflow bit
+	adc	\$0,%rdx
+	mov	$hi1,-8(%rsp,$num,8)
+	mov	%rdx,(%rsp,$num,8)	# store upmost overflow bit
+
+	lea	1($i),$i		# i++
+	cmp	$num,$i
+	jl	.Louter
+
+	lea	(%rsp),$ap		# borrow ap for tp
+	lea	-1($num),$j		# j=num-1
+
+	mov	($ap),%rax		# tp[0]
+	xor	$i,$i			# i=0 and clear CF!
+	jmp	.Lsub
+.align	16
+.Lsub:	sbb	($np,$i,8),%rax
+	mov	%rax,($rp,$i,8)		# rp[i]=tp[i]-np[i]
+	dec	$j			# doesn't affect CF!
+	mov	8($ap,$i,8),%rax	# tp[i+1]
+	lea	1($i),$i		# i++
+	jge	.Lsub
+
+	sbb	\$0,%rax		# handle upmost overflow bit
+	and	%rax,$ap
+	not	%rax
+	mov	$rp,$np
+	and	%rax,$np
+	lea	-1($num),$j
+	or	$np,$ap			# ap=borrow?tp:rp
+.align	16
+.Lcopy:					# copy or in-place refresh
+	mov	($ap,$j,8),%rax
+	mov	%rax,($rp,$j,8)		# rp[i]=tp[i]
+	mov	$i,(%rsp,$j,8)		# zap temporary vector
+	dec	$j
+	jge	.Lcopy
+
+	mov	8(%rsp,$num,8),%rsp	# restore %rsp
+	mov	\$1,%rax
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	ret
+.size	bn_mul_mont,.-bn_mul_mont
+.asciz	"Montgomery Multiplication for x86_64, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+print $code;
+close STDOUT;

crypto/bn/bn_lcl.h

@@ -481,6 +481,7 @@ BN_ULONG bn_add_part_words(BN_ULONG *r, const BN_ULONG *a, const BN_ULONG *b,
 	int cl, int dl);
 BN_ULONG bn_sub_part_words(BN_ULONG *r, const BN_ULONG *a, const BN_ULONG *b,
 	int cl, int dl);
+int bn_mul_mont(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp, const BN_ULONG *np,const BN_ULONG *n0, int num);
 
 #ifdef  __cplusplus
 }

crypto/bn/bn_mont.c

@@ -127,6 +127,21 @@ int BN_mod_mul_montgomery(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
 	{
 	BIGNUM *tmp;
 	int ret=0;
+#if defined(OPENSSL_BN_ASM_MONT) && defined(MONT_WORD)
+	int num = mont->N.top;
+
+	if (num>1 && a->top==num && b->top==num)
+		{
+		if (bn_wexpand(r,num) == NULL) return(0);
+		if (bn_mul_mont(r->d,a->d,b->d,mont->N.d,&mont->n0,num))
+			{
+			r->neg = a->neg^b->neg;
+			r->top = num;
+			bn_correct_top(r);
+			return(1);
+			}
+		}
+#endif
 
 	BN_CTX_start(ctx);
 	tmp = BN_CTX_get(ctx);
@@ -284,6 +299,8 @@ int BN_from_montgomery(BIGNUM *ret, const BIGNUM *a, BN_MONT_CTX *mont,
 		}
 	for (ri+=4; i<ri; i++)
 		rp[i]=nrp[i], ap[i]=0;
+	bn_correct_top(r);
+	bn_correct_top(ret);
 # else
 	if (bn_wexpand(ret,al) == NULL) goto err;
 	ret->top=al;

crypto/bn/bn_mul.c

@@ -655,16 +655,16 @@ void bn_mul_part_recursive(BN_ULONG *r, BN_ULONG *a, BN_ULONG *b, int n,
 				for (;;)
 					{
 					i/=2;
-					if (i < tna && i < tnb)
+					if (i <= tna && tna == tnb)
 						{
-						bn_mul_part_recursive(&(r[n2]),
+						bn_mul_recursive(&(r[n2]),
 							&(a[n]),&(b[n]),
 							i,tna-i,tnb-i,p);
 						break;
 						}
-					else if (i <= tna && i <= tnb)
+					else if (i < tna || i < tnb)
 						{
-						bn_mul_recursive(&(r[n2]),
+						bn_mul_part_recursive(&(r[n2]),
 							&(a[n]),&(b[n]),
 							i,tna-i,tnb-i,p);
 						break;

crypto/bn/bn_prime.c

@@ -377,14 +377,14 @@ static int witness(BIGNUM *w, const BIGNUM *a, const BIGNUM *a1,
 static int probable_prime(BIGNUM *rnd, int bits)
 	{
 	int i;
-	BN_ULONG mods[NUMPRIMES];
+	prime_t mods[NUMPRIMES];
 	BN_ULONG delta,maxdelta;
 
 again:
 	if (!BN_rand(rnd,bits,1,1)) return(0);
 	/* we now have a random number 'rand' to test. */
 	for (i=1; i<NUMPRIMES; i++)
-		mods[i]=BN_mod_word(rnd,(BN_ULONG)primes[i]);
+		mods[i]=(prime_t)BN_mod_word(rnd,(BN_ULONG)primes[i]);
 	maxdelta=BN_MASK2 - primes[NUMPRIMES-1];
 	delta=0;
 	loop: for (i=1; i<NUMPRIMES; i++)

crypto/bn/bn_prime.h

@@ -58,10 +58,12 @@
 
 #ifndef EIGHT_BIT
 #define NUMPRIMES 2048
+typedef unsigned short prime_t;
 #else
 #define NUMPRIMES 54
+typedef unsigned char prime_t;
 #endif
-static const unsigned int primes[NUMPRIMES]=
+static const prime_t primes[NUMPRIMES]=
 	{
 	   2,   3,   5,   7,  11,  13,  17,  19,
 	  23,  29,  31,  37,  41,  43,  47,  53,

crypto/bn/bn_prime.pl

@@ -101,10 +101,12 @@ for ($i=0; $i <= $#primes; $i++)
 
 printf "#ifndef EIGHT_BIT\n";
 printf "#define NUMPRIMES %d\n",$num;
+printf "typedef unsigned short prime_t;\n";
 printf "#else\n";
 printf "#define NUMPRIMES %d\n",$eight;
+printf "typedef unsigned char prime_t;\n";
 printf "#endif\n";
-print "static const unsigned int primes[NUMPRIMES]=\n\t{\n\t";
+print "static const prime_t primes[NUMPRIMES]=\n\t{\n\t";
 $init=0;
 for ($i=0; $i <= $#primes; $i++)
 	{

crypto/cryptlib.c

@@ -270,7 +270,7 @@ int CRYPTO_add_lock(int *pointer, int amount, int type, const char *file,
 
 #if	defined(__i386)   || defined(__i386__)   || defined(_M_IX86) || \
 	defined(__INTEL__) || \
-	defined(__x86_64) || defined(__x86_64__) || defined(_M_AMD64)
+	defined(__x86_64) || defined(__x86_64__) || defined(_M_AMD64) || defined(_M_X64)
 
 unsigned long  OPENSSL_ia32cap_P=0;
 unsigned long *OPENSSL_ia32cap_loc(void) { return &OPENSSL_ia32cap_P; }
@@ -306,6 +306,62 @@ void OPENSSL_cpuid_setup(void) {}
 #endif
 
 #if (defined(_WIN32) || defined(__CYGWIN__)) && defined(_WINDLL)
+
+#ifdef OPENSSL_FIPS
+
+#include <tlhelp32.h>
+#if defined(__GNUC__) && __GNUC__>=2
+static int DllInit(void) __attribute__((constructor));
+#elif defined(_MSC_VER)
+static int DllInit(void);
+# ifdef _WIN64
+# pragma section(".CRT$XCU",read)
+  __declspec(allocate(".CRT$XCU"))
+# else
+# pragma data_seg(".CRT$XCU")
+# endif
+  static int (*p)(void) = DllInit;
+# pragma data_seg()
+#endif
+
+static int DllInit(void)
+{
+#if defined(_WIN32_WINNT)
+	union	{ int(*f)(void); BYTE *p; } t = { DllInit };
+        HANDLE	hModuleSnap = INVALID_HANDLE_VALUE;
+	IMAGE_DOS_HEADER *dos_header;
+	IMAGE_NT_HEADERS *nt_headers;
+	MODULEENTRY32 me32 = {sizeof(me32)};
+
+	hModuleSnap = CreateToolhelp32Snapshot(TH32CS_SNAPMODULE,0);
+	if (hModuleSnap != INVALID_HANDLE_VALUE &&
+	    Module32First(hModuleSnap,&me32)) do
+		{
+		if (t.p >= me32.modBaseAddr &&
+		    t.p <  me32.modBaseAddr+me32.modBaseSize)
+			{
+			dos_header=(IMAGE_DOS_HEADER *)me32.modBaseAddr;
+			if (dos_header->e_magic==IMAGE_DOS_SIGNATURE)
+				{
+				nt_headers=(IMAGE_NT_HEADERS *)
+					((BYTE *)dos_header+dos_header->e_lfanew);
+				if (nt_headers->Signature==IMAGE_NT_SIGNATURE &&
+				    me32.modBaseAddr!=(BYTE*)nt_headers->OptionalHeader.ImageBase)
+					OPENSSL_NONPIC_relocated=1;
+				}
+			break;
+			}
+		} while (Module32Next(hModuleSnap,&me32));
+
+	if (hModuleSnap != INVALID_HANDLE_VALUE)
+		CloseHandle(hModuleSnap);
+#endif
+	OPENSSL_cpuid_setup();
+	return 0;
+}
+
+#else
+
 #ifdef __CYGWIN__
 /* pick DLL_[PROCESS|THREAD]_[ATTACH|DETACH] definitions */
 #include <windows.h>
@@ -349,6 +405,8 @@ BOOL WINAPI DllMain(HINSTANCE hinstDLL, DWORD fdwReason,
 	}
 #endif
 
+#endif
+
 #if defined(_WIN32) && !defined(__CYGWIN__)
 #include <tchar.h>
 

crypto/des/Makefile

@@ -156,7 +156,7 @@ des_enc.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
 des_enc.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
 des_enc.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 des_enc.o: ../../include/openssl/ui.h ../../include/openssl/ui_compat.h
-des_enc.o: des_enc.c des_locl.h
+des_enc.o: des_enc.c des_locl.h ncbc_enc.c
 des_lib.o: ../../include/openssl/bio.h ../../include/openssl/crypto.h
 des_lib.o: ../../include/openssl/des.h ../../include/openssl/des_old.h
 des_lib.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h

crypto/des/des_enc.c

@@ -58,8 +58,6 @@
 
 #include "des_locl.h"
 
-#ifndef OPENSSL_FIPS
-
 void DES_encrypt1(DES_LONG *data, DES_key_schedule *ks, int enc)
 	{
 	register DES_LONG l,r,t,u;
@@ -289,8 +287,6 @@ void DES_decrypt3(DES_LONG *data, DES_key_schedule *ks1,
 	data[1]=r;
 	}
 
-#endif
-
 #ifndef DES_DEFAULT_OPTIONS
 
 #if !defined(OPENSSL_FIPS_DES_ASM)

crypto/des/set_key.c

@@ -63,10 +63,9 @@
  * 1.1 added norm_expand_bits
  * 1.0 First working version
  */
+#include <openssl/fips.h>
 #include "des_locl.h"
 
-#ifndef OPENSSL_FIPS
-
 OPENSSL_IMPLEMENT_GLOBAL(int,DES_check_key);	/* defaults to false */
 
 static const unsigned char odd_parity[256]={
@@ -117,7 +116,7 @@ int DES_check_key_parity(const_DES_cblock *key)
  * (and actual cblock values).
  */
 #define NUM_WEAK_KEY	16
-static DES_cblock weak_keys[NUM_WEAK_KEY]={
+static const DES_cblock weak_keys[NUM_WEAK_KEY]={
 	/* weak keys */
 	{0x01,0x01,0x01,0x01,0x01,0x01,0x01,0x01},
 	{0xFE,0xFE,0xFE,0xFE,0xFE,0xFE,0xFE,0xFE},
@@ -351,6 +350,10 @@ void DES_set_key_unchecked(const_DES_cblock *key, DES_key_schedule *schedule)
 	k = &schedule->ks->deslong[0];
 	in = &(*key)[0];
 
+#ifdef OPENSSL_FIPS
+	FIPS_selftest_check();
+#endif
+
 	c2l(in,c);
 	c2l(in,d);
 
@@ -408,4 +411,3 @@ void des_fixup_key_parity(des_cblock *key)
 	}
 */
 
-#endif

crypto/dh/dh.h

@@ -77,6 +77,8 @@
 # define OPENSSL_DH_MAX_MODULUS_BITS	10000
 #endif
 
+#define OPENSSL_DH_FIPS_MIN_MODULUS_BITS 1024
+
 #define DH_FLAG_CACHE_MONT_P     0x01
 #define DH_FLAG_NO_EXP_CONSTTIME 0x02 /* new with 0.9.7h; the built-in DH
                                        * implementation now uses constant time
@@ -233,6 +235,7 @@ void ERR_load_DH_strings(void);
 /* Reason codes. */
 #define DH_R_BAD_GENERATOR				 101
 #define DH_R_INVALID_PUBKEY				 102
+#define DH_R_KEY_SIZE_TOO_SMALL				 104
 #define DH_R_MODULUS_TOO_LARGE				 103
 #define DH_R_NO_PRIVATE_VALUE				 100
 

crypto/dh/dh_err.c

@@ -1,6 +1,6 @@
 /* crypto/dh/dh_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -87,6 +87,7 @@ static ERR_STRING_DATA DH_str_reasons[]=
 	{
 {ERR_REASON(DH_R_BAD_GENERATOR)          ,"bad generator"},
 {ERR_REASON(DH_R_INVALID_PUBKEY)         ,"invalid public key"},
+{ERR_REASON(DH_R_KEY_SIZE_TOO_SMALL)     ,"key size too small"},
 {ERR_REASON(DH_R_MODULUS_TOO_LARGE)      ,"modulus too large"},
 {ERR_REASON(DH_R_NO_PRIVATE_VALUE)       ,"no private value"},
 {0,NULL}

crypto/dsa/Makefile

@@ -80,8 +80,8 @@ dsa_asn1.o: ../../e_os.h ../../include/openssl/asn1.h
 dsa_asn1.o: ../../include/openssl/asn1t.h ../../include/openssl/bio.h
 dsa_asn1.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 dsa_asn1.o: ../../include/openssl/dsa.h ../../include/openssl/e_os2.h
-dsa_asn1.o: ../../include/openssl/err.h ../../include/openssl/lhash.h
+dsa_asn1.o: ../../include/openssl/err.h ../../include/openssl/fips.h
-dsa_asn1.o: ../../include/openssl/opensslconf.h
+dsa_asn1.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
 dsa_asn1.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 dsa_asn1.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 dsa_asn1.o: ../../include/openssl/symhacks.h ../cryptlib.h dsa_asn1.c

crypto/dsa/dsa.h

@@ -88,6 +88,8 @@
 # define OPENSSL_DSA_MAX_MODULUS_BITS	10000
 #endif
 
+#define OPENSSL_DSA_FIPS_MIN_MODULUS_BITS 1024
+
 #define DSA_FLAG_CACHE_MONT_P	0x01
 #define DSA_FLAG_NO_EXP_CONSTTIME       0x02 /* new with 0.9.7h; the built-in DSA
                                               * implementation now uses constant time
@@ -290,8 +292,11 @@ void ERR_load_DSA_strings(void);
 #define DSA_F_D2I_DSA_SIG				 110
 #define DSA_F_DSAPARAMS_PRINT				 100
 #define DSA_F_DSAPARAMS_PRINT_FP			 101
+#define DSA_F_DSA_BUILTIN_KEYGEN			 119
+#define DSA_F_DSA_BUILTIN_PARAMGEN			 118
 #define DSA_F_DSA_DO_SIGN				 112
 #define DSA_F_DSA_DO_VERIFY				 113
+#define DSA_F_DSA_GENERATE_PARAMETERS			 117
 #define DSA_F_DSA_NEW_METHOD				 103
 #define DSA_F_DSA_PRINT					 104
 #define DSA_F_DSA_PRINT_FP				 105
@@ -307,6 +312,7 @@ void ERR_load_DSA_strings(void);
 /* Reason codes. */
 #define DSA_R_BAD_Q_VALUE				 102
 #define DSA_R_DATA_TOO_LARGE_FOR_KEY_SIZE		 100
+#define DSA_R_KEY_SIZE_TOO_SMALL			 106
 #define DSA_R_MISSING_PARAMETERS			 101
 #define DSA_R_MODULUS_TOO_LARGE				 103
 #define DSA_R_NON_FIPS_METHOD				 104

crypto/dsa/dsa_err.c

@@ -73,8 +73,11 @@ static ERR_STRING_DATA DSA_str_functs[]=
 {ERR_FUNC(DSA_F_D2I_DSA_SIG),	"d2i_DSA_SIG"},
 {ERR_FUNC(DSA_F_DSAPARAMS_PRINT),	"DSAparams_print"},
 {ERR_FUNC(DSA_F_DSAPARAMS_PRINT_FP),	"DSAparams_print_fp"},
+{ERR_FUNC(DSA_F_DSA_BUILTIN_KEYGEN),	"DSA_BUILTIN_KEYGEN"},
+{ERR_FUNC(DSA_F_DSA_BUILTIN_PARAMGEN),	"DSA_BUILTIN_PARAMGEN"},
 {ERR_FUNC(DSA_F_DSA_DO_SIGN),	"DSA_do_sign"},
 {ERR_FUNC(DSA_F_DSA_DO_VERIFY),	"DSA_do_verify"},
+{ERR_FUNC(DSA_F_DSA_GENERATE_PARAMETERS),	"DSA_generate_parameters"},
 {ERR_FUNC(DSA_F_DSA_NEW_METHOD),	"DSA_new_method"},
 {ERR_FUNC(DSA_F_DSA_PRINT),	"DSA_print"},
 {ERR_FUNC(DSA_F_DSA_PRINT_FP),	"DSA_print_fp"},
@@ -93,6 +96,7 @@ static ERR_STRING_DATA DSA_str_reasons[]=
 	{
 {ERR_REASON(DSA_R_BAD_Q_VALUE)           ,"bad q value"},
 {ERR_REASON(DSA_R_DATA_TOO_LARGE_FOR_KEY_SIZE),"data too large for key size"},
+{ERR_REASON(DSA_R_KEY_SIZE_TOO_SMALL)    ,"key size too small"},
 {ERR_REASON(DSA_R_MISSING_PARAMETERS)    ,"missing parameters"},
 {ERR_REASON(DSA_R_MODULUS_TOO_LARGE)     ,"modulus too large"},
 {ERR_REASON(DSA_R_NON_FIPS_METHOD)       ,"non fips method"},

crypto/dsa/dsa_gen.c

@@ -119,13 +119,20 @@ static int dsa_builtin_paramgen(DSA *ret, int bits,
 	if (bits < 512) bits=512;
 	bits=(bits+63)/64*64;
 
-	if (seed_len < 20)
+	/* NB: seed_len == 0 is special case: copy generated seed to
+ 	 * seed_in if it is not NULL.
+ 	 */
+	if (seed_len && (seed_len < 20))
 		seed_in = NULL; /* seed buffer too small -- ignore */
 	if (seed_len > 20) 
 		seed_len = 20; /* App. 2.2 of FIPS PUB 186 allows larger SEED,
 		                * but our internal buffers are restricted to 160 bits*/
 	if ((seed_in != NULL) && (seed_len == 20))
+		{
 		memcpy(seed,seed_in,seed_len);
+		/* set seed_in to NULL to avoid it being copied back */
+		seed_in = NULL;
+		}
 
 	if ((ctx=BN_CTX_new()) == NULL) goto err;
 
@@ -302,7 +309,7 @@ err:
 			ok=0;
 			goto err;
 			}
-		if ((m > 1) && (seed_in != NULL)) memcpy(seed_in,seed,20);
+		if (seed_in != NULL) memcpy(seed_in,seed,20);
 		if (counter_ret != NULL) *counter_ret=counter;
 		if (h_ret != NULL) *h_ret=h;
 		}

crypto/err/err_str.c

@@ -223,7 +223,7 @@ static ERR_STRING_DATA SYS_str_reasons[NUM_SYS_STR_REASONS + 1];
  * others will be displayed numerically by ERR_error_string.
  * It is crucial that we have something for each reason code
  * that occurs in ERR_str_reasons, or bogus reason strings
- * will be returned for SYSerr(), which always gets an errno
+ * will be returned for SYSerr, which always gets an errno
  * value and never one of those 'standard' reason codes. */
 
 static void build_SYS_str_reasons(void)

crypto/err/openssl.ec

@@ -31,7 +31,7 @@ L COMP		crypto/comp/comp.h		crypto/comp/comp_err.c
 L ECDSA		crypto/ecdsa/ecdsa.h		crypto/ecdsa/ecs_err.c
 L ECDH		crypto/ecdh/ecdh.h		crypto/ecdh/ech_err.c
 L STORE		crypto/store/store.h		crypto/store/str_err.c
-L FIPS		fips-1.0/fips.h			crypto/fips_err.h
+L FIPS		fips/fips.h			crypto/fips_err.h
 
 # additional header files to be scanned for function names
 L NONE		crypto/x509/x509_vfy.h		NONE

crypto/evp/Makefile

@@ -293,7 +293,15 @@ e_rc5.o: ../../include/openssl/lhash.h ../../include/openssl/opensslconf.h
 e_rc5.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 e_rc5.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 e_rc5.o: ../../include/openssl/symhacks.h ../cryptlib.h e_rc5.c
-e_seed.o: ../../include/openssl/opensslconf.h e_seed.c
+e_seed.o: ../../include/openssl/asn1.h ../../include/openssl/bio.h
+e_seed.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+e_seed.o: ../../include/openssl/err.h ../../include/openssl/evp.h
+e_seed.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
+e_seed.o: ../../include/openssl/obj_mac.h ../../include/openssl/objects.h
+e_seed.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+e_seed.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
+e_seed.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+e_seed.o: e_seed.c
 e_xcbc_d.o: ../../e_os.h ../../include/openssl/asn1.h
 e_xcbc_d.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 e_xcbc_d.o: ../../include/openssl/crypto.h ../../include/openssl/des.h
@@ -306,6 +314,17 @@ e_xcbc_d.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 e_xcbc_d.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 e_xcbc_d.o: ../../include/openssl/symhacks.h ../../include/openssl/ui.h
 e_xcbc_d.o: ../../include/openssl/ui_compat.h ../cryptlib.h e_xcbc_d.c
+enc_min.o: ../../e_os.h ../../include/openssl/asn1.h
+enc_min.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+enc_min.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
+enc_min.o: ../../include/openssl/engine.h ../../include/openssl/err.h
+enc_min.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
+enc_min.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+enc_min.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+enc_min.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+enc_min.o: ../../include/openssl/rand.h ../../include/openssl/safestack.h
+enc_min.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
+enc_min.o: ../cryptlib.h enc_min.c evp_locl.h
 encode.o: ../../e_os.h ../../include/openssl/asn1.h ../../include/openssl/bio.h
 encode.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
 encode.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
@@ -325,6 +344,21 @@ evp_acnf.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
 evp_acnf.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 evp_acnf.o: ../../include/openssl/safestack.h ../../include/openssl/stack.h
 evp_acnf.o: ../../include/openssl/symhacks.h ../cryptlib.h evp_acnf.c
+evp_cnf.o: ../../e_os.h ../../include/openssl/asn1.h
+evp_cnf.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
+evp_cnf.o: ../../include/openssl/conf.h ../../include/openssl/crypto.h
+evp_cnf.o: ../../include/openssl/dso.h ../../include/openssl/e_os2.h
+evp_cnf.o: ../../include/openssl/ec.h ../../include/openssl/ecdh.h
+evp_cnf.o: ../../include/openssl/ecdsa.h ../../include/openssl/err.h
+evp_cnf.o: ../../include/openssl/evp.h ../../include/openssl/fips.h
+evp_cnf.o: ../../include/openssl/lhash.h ../../include/openssl/obj_mac.h
+evp_cnf.o: ../../include/openssl/objects.h ../../include/openssl/opensslconf.h
+evp_cnf.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
+evp_cnf.o: ../../include/openssl/pkcs7.h ../../include/openssl/safestack.h
+evp_cnf.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+evp_cnf.o: ../../include/openssl/symhacks.h ../../include/openssl/x509.h
+evp_cnf.o: ../../include/openssl/x509_vfy.h ../../include/openssl/x509v3.h
+evp_cnf.o: ../cryptlib.h evp_cnf.c
 evp_enc.o: ../../e_os.h ../../include/openssl/asn1.h
 evp_enc.o: ../../include/openssl/bio.h ../../include/openssl/buffer.h
 evp_enc.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h

crypto/evp/dig_eng.c

@@ -135,7 +135,7 @@ static int do_evp_md_engine_full(EVP_MD_CTX *ctx, const EVP_MD **ptype, ENGINE *
 			{
 			if (!ENGINE_init(impl))
 				{
-				EVPerr(EVP_F_EVP_DIGESTINIT_EX,EVP_R_INITIALIZATION_ERROR);
+				EVPerr(EVP_F_DO_EVP_MD_ENGINE_FULL,EVP_R_INITIALIZATION_ERROR);
 				return 0;
 				}
 			}
@@ -149,7 +149,7 @@ static int do_evp_md_engine_full(EVP_MD_CTX *ctx, const EVP_MD **ptype, ENGINE *
 			if(!d)
 				{
 				/* Same comment from evp_enc.c */
-				EVPerr(EVP_F_EVP_DIGESTINIT_EX,EVP_R_INITIALIZATION_ERROR);
+				EVPerr(EVP_F_DO_EVP_MD_ENGINE_FULL,EVP_R_INITIALIZATION_ERROR);
 				return 0;
 				}
 			/* We'll use the ENGINE's private digest definition */
@@ -165,7 +165,7 @@ static int do_evp_md_engine_full(EVP_MD_CTX *ctx, const EVP_MD **ptype, ENGINE *
 	else
 	if(!ctx->digest)
 		{
-		EVPerr(EVP_F_EVP_DIGESTINIT_EX,EVP_R_NO_DIGEST_SET);
+		EVPerr(EVP_F_DO_EVP_MD_ENGINE_FULL,EVP_R_NO_DIGEST_SET);
 		return 0;
 		}
 	return 1;

crypto/evp/digest.c

@@ -120,9 +120,6 @@
 
 void EVP_MD_CTX_init(EVP_MD_CTX *ctx)
 	{
-#ifdef OPENSSL_FIPS
-	FIPS_selftest_check();
-#endif
 	memset(ctx,'\0',sizeof *ctx);
 	}
 
@@ -222,7 +219,7 @@ static int do_evp_md_engine(EVP_MD_CTX *ctx, const EVP_MD **ptype, ENGINE *impl)
 			{
 			if (!ENGINE_init(impl))
 				{
-				EVPerr(EVP_F_EVP_DIGESTINIT_EX,EVP_R_INITIALIZATION_ERROR);
+				EVPerr(EVP_F_DO_EVP_MD_ENGINE,EVP_R_INITIALIZATION_ERROR);
 				return 0;
 				}
 			}
@@ -236,7 +233,7 @@ static int do_evp_md_engine(EVP_MD_CTX *ctx, const EVP_MD **ptype, ENGINE *impl)
 			if(!d)
 				{
 				/* Same comment from evp_enc.c */
-				EVPerr(EVP_F_EVP_DIGESTINIT_EX,EVP_R_INITIALIZATION_ERROR);
+				EVPerr(EVP_F_DO_EVP_MD_ENGINE,EVP_R_INITIALIZATION_ERROR);
 				return 0;
 				}
 			/* We'll use the ENGINE's private digest definition */
@@ -252,7 +249,7 @@ static int do_evp_md_engine(EVP_MD_CTX *ctx, const EVP_MD **ptype, ENGINE *impl)
 	else
 	if(!ctx->digest)
 		{
-		EVPerr(EVP_F_EVP_DIGESTINIT_EX,EVP_R_NO_DIGEST_SET);
+		EVPerr(EVP_F_DO_EVP_MD_ENGINE,EVP_R_NO_DIGEST_SET);
 		return 0;
 		}
 	return 1;
@@ -265,6 +262,14 @@ static int do_evp_md_engine(EVP_MD_CTX *ctx, const EVP_MD **ptype, ENGINE *impl)
 int EVP_DigestInit_ex(EVP_MD_CTX *ctx, const EVP_MD *type, ENGINE *impl)
 	{
 	M_EVP_MD_CTX_clear_flags(ctx,EVP_MD_CTX_FLAG_CLEANED);
+#ifdef OPENSSL_FIPS
+	if(FIPS_selftest_failed())
+		{
+		FIPSerr(FIPS_F_EVP_DIGESTINIT_EX,FIPS_R_FIPS_SELFTEST_FAILED);
+		ctx->digest = &bad_md;
+		return 0;
+		}
+#endif
 #ifndef OPENSSL_NO_ENGINE
 	/* Whether it's nice or not, "Inits" can be used on "Final"'d contexts
 	 * so this context may already have an ENGINE! Try to avoid releasing
@@ -284,7 +289,7 @@ int EVP_DigestInit_ex(EVP_MD_CTX *ctx, const EVP_MD *type, ENGINE *impl)
 			if (!(type->flags & EVP_MD_FLAG_FIPS) 
 			 && !(ctx->flags & EVP_MD_CTX_FLAG_NON_FIPS_ALLOW))
 				{
-				EVPerr(EVP_F_EVP_DIGESTINIT, EVP_R_DISABLED_FOR_FIPS);
+				EVPerr(EVP_F_EVP_DIGESTINIT_EX, EVP_R_DISABLED_FOR_FIPS);
 				ctx->digest = &bad_md;
 				return 0;
 				}
@@ -305,6 +310,9 @@ int EVP_DigestInit_ex(EVP_MD_CTX *ctx, const EVP_MD *type, ENGINE *impl)
 int EVP_DigestUpdate(EVP_MD_CTX *ctx, const void *data,
 	     size_t count)
 	{
+#ifdef OPENSSL_FIPS
+	FIPS_selftest_check();
+#endif
 	return ctx->digest->update(ctx,data,count);
 	}
 
@@ -321,6 +329,9 @@ int EVP_DigestFinal(EVP_MD_CTX *ctx, unsigned char *md, unsigned int *size)
 int EVP_DigestFinal_ex(EVP_MD_CTX *ctx, unsigned char *md, unsigned int *size)
 	{
 	int ret;
+#ifdef OPENSSL_FIPS
+	FIPS_selftest_check();
+#endif
 
 	OPENSSL_assert(ctx->digest->md_size <= EVP_MAX_MD_SIZE);
 	ret=ctx->digest->final(ctx,md);

crypto/evp/enc_min.c

@@ -152,7 +152,7 @@ static int do_evp_enc_engine(EVP_CIPHER_CTX *ctx, const EVP_CIPHER **pcipher, EN
 		{
 		if (!ENGINE_init(impl))
 			{
-			EVPerr(EVP_F_EVP_CIPHERINIT_EX, EVP_R_INITIALIZATION_ERROR);
+			EVPerr(EVP_F_DO_EVP_ENC_ENGINE, EVP_R_INITIALIZATION_ERROR);
 			return 0;
 			}
 		}
@@ -169,7 +169,7 @@ static int do_evp_enc_engine(EVP_CIPHER_CTX *ctx, const EVP_CIPHER **pcipher, EN
 			 * control history, is that we should at least
 			 * be able to avoid using US mispellings of
 			 * "initialisation"? */
-			EVPerr(EVP_F_EVP_CIPHERINIT_EX, EVP_R_INITIALIZATION_ERROR);
+			EVPerr(EVP_F_DO_EVP_ENC_ENGINE, EVP_R_INITIALIZATION_ERROR);
 			return 0;
 			}
 		/* We'll use the ENGINE's private cipher definition */
@@ -199,6 +199,14 @@ int EVP_CipherInit_ex(EVP_CIPHER_CTX *ctx, const EVP_CIPHER *cipher, ENGINE *imp
 			enc = 1;
 		ctx->encrypt = enc;
 		}
+#ifdef OPENSSL_NO_FIPS
+	if(FIPS_selftest_failed())
+		{
+		FIPSerr(FIPS_F_EVP_CIPHERINIT_EX,FIPS_R_FIPS_SELFTEST_FAILED);
+		ctx->cipher = &bad_cipher;
+		return 0;
+		}
+#endif
 #ifndef OPENSSL_NO_ENGINE
 	/* Whether it's nice or not, "Inits" can be used on "Final"'d contexts
 	 * so this context may already have an ENGINE! Try to avoid releasing
@@ -295,7 +303,7 @@ skip_to_init:
 		if (!(ctx->cipher->flags & EVP_CIPH_FLAG_FIPS)
 			& !(ctx->flags & EVP_CIPH_FLAG_NON_FIPS_ALLOW))
 			{
-			EVPerr(EVP_F_EVP_CIPHERINIT, EVP_R_DISABLED_FOR_FIPS);
+			EVPerr(EVP_F_EVP_CIPHERINIT_EX, EVP_R_DISABLED_FOR_FIPS);
 #if 0
 			ERR_add_error_data(2, "cipher=",
 						EVP_CIPHER_name(ctx->cipher));
@@ -339,6 +347,9 @@ int EVP_CIPHER_CTX_cleanup(EVP_CIPHER_CTX *c)
 
 int EVP_Cipher(EVP_CIPHER_CTX *ctx, unsigned char *out, const unsigned char *in, unsigned int inl)
 	{
+#ifdef OPENSSL_FIPS
+	FIPS_selftest_check();
+#endif
 	return ctx->cipher->do_cipher(ctx,out,in,inl);
 	}
 

crypto/evp/evp.h

@@ -956,6 +956,10 @@ void ERR_load_EVP_strings(void);
 #define EVP_F_ALG_MODULE_INIT				 138
 #define EVP_F_CAMELLIA_INIT_KEY				 159
 #define EVP_F_D2I_PKEY					 100
+#define EVP_F_DO_EVP_ENC_ENGINE				 140
+#define EVP_F_DO_EVP_ENC_ENGINE_FULL			 141
+#define EVP_F_DO_EVP_MD_ENGINE				 139
+#define EVP_F_DO_EVP_MD_ENGINE_FULL			 142
 #define EVP_F_DSAPKEY2PKCS8				 134
 #define EVP_F_DSA_PKEY2PKCS8				 135
 #define EVP_F_ECDSA_PKEY2PKCS8				 129
@@ -1032,6 +1036,7 @@ void ERR_load_EVP_strings(void);
 #define EVP_R_NO_VERIFY_FUNCTION_CONFIGURED		 105
 #define EVP_R_PKCS8_UNKNOWN_BROKEN_TYPE			 117
 #define EVP_R_PUBLIC_KEY_NOT_RSA			 106
+#define EVP_R_SEED_KEY_SETUP_FAILED			 162
 #define EVP_R_UNKNOWN_OPTION				 149
 #define EVP_R_UNKNOWN_PBE_ALGORITHM			 121
 #define EVP_R_UNSUPORTED_NUMBER_OF_ROUNDS		 135
@@ -1044,7 +1049,6 @@ void ERR_load_EVP_strings(void);
 #define EVP_R_UNSUPPORTED_SALT_TYPE			 126
 #define EVP_R_WRONG_FINAL_BLOCK_LENGTH			 109
 #define EVP_R_WRONG_PUBLIC_KEY_TYPE			 110
-#define EVP_R_SEED_KEY_SETUP_FAILED			 162
 
 #ifdef  __cplusplus
 }

crypto/evp/evp_enc.c

@@ -66,6 +66,14 @@
 #endif
 #include "evp_locl.h"
 
+#ifdef OPENSSL_FIPS
+	#define M_do_cipher(ctx, out, in, inl) \
+		EVP_Cipher(ctx,out,in,inl)
+#else
+	#define M_do_cipher(ctx, out, in, inl) \
+		ctx->cipher->do_cipher(ctx,out,in,inl)
+#endif
+
 const char EVP_version[]="EVP" OPENSSL_VERSION_PTEXT;
 
 EVP_CIPHER_CTX *EVP_CIPHER_CTX_new(void)
@@ -138,7 +146,7 @@ int EVP_EncryptUpdate(EVP_CIPHER_CTX *ctx, unsigned char *out, int *outl,
 	OPENSSL_assert(inl > 0);
 	if(ctx->buf_len == 0 && (inl&(ctx->block_mask)) == 0)
 		{
-		if(ctx->cipher->do_cipher(ctx,out,in,inl))
+		if(M_do_cipher(ctx,out,in,inl))
 			{
 			*outl=inl;
 			return 1;
@@ -165,7 +173,7 @@ int EVP_EncryptUpdate(EVP_CIPHER_CTX *ctx, unsigned char *out, int *outl,
 			{
 			j=bl-i;
 			memcpy(&(ctx->buf[i]),in,j);
-			if(!ctx->cipher->do_cipher(ctx,out,ctx->buf,bl)) return 0;
+			if(!M_do_cipher(ctx,out,ctx->buf,bl)) return 0;
 			inl-=j;
 			in+=j;
 			out+=bl;
@@ -178,7 +186,7 @@ int EVP_EncryptUpdate(EVP_CIPHER_CTX *ctx, unsigned char *out, int *outl,
 	inl-=i;
 	if (inl > 0)
 		{
-		if(!ctx->cipher->do_cipher(ctx,out,in,inl)) return 0;
+		if(!M_do_cipher(ctx,out,in,inl)) return 0;
 		*outl+=inl;
 		}
 
@@ -222,7 +230,7 @@ int EVP_EncryptFinal_ex(EVP_CIPHER_CTX *ctx, unsigned char *out, int *outl)
 	n=b-bl;
 	for (i=bl; i<b; i++)
 		ctx->buf[i]=n;
-	ret=ctx->cipher->do_cipher(ctx,out,ctx->buf,b);
+	ret=M_do_cipher(ctx,out,ctx->buf,b);
 
 
 	if(ret)
@@ -384,7 +392,7 @@ static int do_evp_enc_engine_full(EVP_CIPHER_CTX *ctx, const EVP_CIPHER **pciphe
 		{
 		if (!ENGINE_init(impl))
 			{
-			EVPerr(EVP_F_EVP_CIPHERINIT_EX, EVP_R_INITIALIZATION_ERROR);
+			EVPerr(EVP_F_DO_EVP_ENC_ENGINE_FULL, EVP_R_INITIALIZATION_ERROR);
 			return 0;
 			}
 		}
@@ -401,7 +409,7 @@ static int do_evp_enc_engine_full(EVP_CIPHER_CTX *ctx, const EVP_CIPHER **pciphe
 			 * control history, is that we should at least
 			 * be able to avoid using US mispellings of
 			 * "initialisation"? */
-			EVPerr(EVP_F_EVP_CIPHERINIT_EX, EVP_R_INITIALIZATION_ERROR);
+			EVPerr(EVP_F_DO_EVP_ENC_ENGINE_FULL, EVP_R_INITIALIZATION_ERROR);
 			return 0;
 			}
 		/* We'll use the ENGINE's private cipher definition */

crypto/evp/evp_err.c

@@ -74,6 +74,10 @@ static ERR_STRING_DATA EVP_str_functs[]=
 {ERR_FUNC(EVP_F_ALG_MODULE_INIT),	"ALG_MODULE_INIT"},
 {ERR_FUNC(EVP_F_CAMELLIA_INIT_KEY),	"CAMELLIA_INIT_KEY"},
 {ERR_FUNC(EVP_F_D2I_PKEY),	"D2I_PKEY"},
+{ERR_FUNC(EVP_F_DO_EVP_ENC_ENGINE),	"DO_EVP_ENC_ENGINE"},
+{ERR_FUNC(EVP_F_DO_EVP_ENC_ENGINE_FULL),	"DO_EVP_ENC_ENGINE_FULL"},
+{ERR_FUNC(EVP_F_DO_EVP_MD_ENGINE),	"DO_EVP_MD_ENGINE"},
+{ERR_FUNC(EVP_F_DO_EVP_MD_ENGINE_FULL),	"DO_EVP_MD_ENGINE_FULL"},
 {ERR_FUNC(EVP_F_DSAPKEY2PKCS8),	"DSAPKEY2PKCS8"},
 {ERR_FUNC(EVP_F_DSA_PKEY2PKCS8),	"DSA_PKEY2PKCS8"},
 {ERR_FUNC(EVP_F_ECDSA_PKEY2PKCS8),	"ECDSA_PKEY2PKCS8"},
@@ -153,6 +157,7 @@ static ERR_STRING_DATA EVP_str_reasons[]=
 {ERR_REASON(EVP_R_NO_VERIFY_FUNCTION_CONFIGURED),"no verify function configured"},
 {ERR_REASON(EVP_R_PKCS8_UNKNOWN_BROKEN_TYPE),"pkcs8 unknown broken type"},
 {ERR_REASON(EVP_R_PUBLIC_KEY_NOT_RSA)    ,"public key not rsa"},
+{ERR_REASON(EVP_R_SEED_KEY_SETUP_FAILED) ,"seed key setup failed"},
 {ERR_REASON(EVP_R_UNKNOWN_OPTION)        ,"unknown option"},
 {ERR_REASON(EVP_R_UNKNOWN_PBE_ALGORITHM) ,"unknown pbe algorithm"},
 {ERR_REASON(EVP_R_UNSUPORTED_NUMBER_OF_ROUNDS),"unsuported number of rounds"},

crypto/evp/evp_locl.h

@@ -92,7 +92,7 @@ static int cname##_cbc_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out, const uns
 #define BLOCK_CIPHER_func_cfb(cname, cprefix, cbits, kstruct, ksched) \
 static int cname##_cfb##cbits##_cipher(EVP_CIPHER_CTX *ctx, unsigned char *out, const unsigned char *in, unsigned int inl) \
 {\
-	cprefix##_cfb##cbits##_encrypt(in, out, (long)inl, &((kstruct *)ctx->cipher_data)->ksched, ctx->iv, &ctx->num, ctx->encrypt);\
+	cprefix##_cfb##cbits##_encrypt(in, out, (long)(cbits==1?inl*8:inl), &((kstruct *)ctx->cipher_data)->ksched, ctx->iv, &ctx->num, ctx->encrypt);\
 	return 1;\
 }
 

crypto/fips_err.h

@@ -70,29 +70,31 @@
 
 static ERR_STRING_DATA FIPS_str_functs[]=
 	{
-{ERR_FUNC(FIPS_F_DH_GENERATE_PARAMETERS),	"DH_generate_parameters"},
+{ERR_FUNC(FIPS_F_DH_BUILTIN_GENPARAMS),	"DH_BUILTIN_GENPARAMS"},
+{ERR_FUNC(FIPS_F_DSA_BUILTIN_PARAMGEN),	"DSA_BUILTIN_PARAMGEN"},
 {ERR_FUNC(FIPS_F_DSA_DO_SIGN),	"DSA_do_sign"},
 {ERR_FUNC(FIPS_F_DSA_DO_VERIFY),	"DSA_do_verify"},
-{ERR_FUNC(FIPS_F_DSA_GENERATE_PARAMETERS),	"DSA_generate_parameters"},
+{ERR_FUNC(FIPS_F_EVP_CIPHERINIT_EX),	"EVP_CipherInit_ex"},
+{ERR_FUNC(FIPS_F_EVP_DIGESTINIT_EX),	"EVP_DigestInit_ex"},
 {ERR_FUNC(FIPS_F_FIPS_CHECK_DSA),	"FIPS_CHECK_DSA"},
-{ERR_FUNC(FIPS_F_FIPS_CHECK_DSO),	"FIPS_CHECK_DSO"},
+{ERR_FUNC(FIPS_F_FIPS_CHECK_INCORE_FINGERPRINT),	"FIPS_CHECK_INCORE_FINGERPRINT"},
-{ERR_FUNC(FIPS_F_FIPS_CHECK_EXE),	"FIPS_CHECK_EXE"},
-{ERR_FUNC(FIPS_F_FIPS_CHECK_FINGERPRINT),	"FIPS_CHECK_FINGERPRINT"},
-{ERR_FUNC(FIPS_F_FIPS_CHECK_PK_SIG),	"FIPS_CHECK_PK_SIG"},
 {ERR_FUNC(FIPS_F_FIPS_CHECK_RSA),	"FIPS_CHECK_RSA"},
-{ERR_FUNC(FIPS_F_FIPS_DSA_CHECK),	"FIPS_dsa_check"},
+{ERR_FUNC(FIPS_F_FIPS_DSA_CHECK),	"FIPS_DSA_CHECK"},
 {ERR_FUNC(FIPS_F_FIPS_MODE_SET),	"FIPS_mode_set"},
 {ERR_FUNC(FIPS_F_FIPS_PKEY_SIGNATURE_TEST),	"fips_pkey_signature_test"},
 {ERR_FUNC(FIPS_F_FIPS_SELFTEST_AES),	"FIPS_selftest_aes"},
 {ERR_FUNC(FIPS_F_FIPS_SELFTEST_DES),	"FIPS_selftest_des"},
 {ERR_FUNC(FIPS_F_FIPS_SELFTEST_DSA),	"FIPS_selftest_dsa"},
+{ERR_FUNC(FIPS_F_FIPS_SELFTEST_HMAC),	"FIPS_selftest_hmac"},
 {ERR_FUNC(FIPS_F_FIPS_SELFTEST_RNG),	"FIPS_selftest_rng"},
-{ERR_FUNC(FIPS_F_FIPS_SELFTEST_RSA),	"FIPS_selftest_rsa"},
+{ERR_FUNC(FIPS_F_FIPS_SELFTEST_SHA1),	"FIPS_selftest_sha1"},
-{ERR_FUNC(FIPS_F_FIPS_SELFTEST_SHA),	"FIPS_SELFTEST_SHA"},
 {ERR_FUNC(FIPS_F_HASH_FINAL),	"HASH_FINAL"},
+{ERR_FUNC(FIPS_F_RSA_BUILTIN_KEYGEN),	"RSA_BUILTIN_KEYGEN"},
+{ERR_FUNC(FIPS_F_RSA_EAY_PRIVATE_DECRYPT),	"RSA_EAY_PRIVATE_DECRYPT"},
+{ERR_FUNC(FIPS_F_RSA_EAY_PRIVATE_ENCRYPT),	"RSA_EAY_PRIVATE_ENCRYPT"},
+{ERR_FUNC(FIPS_F_RSA_EAY_PUBLIC_DECRYPT),	"RSA_EAY_PUBLIC_DECRYPT"},
 {ERR_FUNC(FIPS_F_RSA_EAY_PUBLIC_ENCRYPT),	"RSA_EAY_PUBLIC_ENCRYPT"},
-{ERR_FUNC(FIPS_F_RSA_GENERATE_KEY),	"RSA_generate_key"},
+{ERR_FUNC(FIPS_F_RSA_X931_GENERATE_KEY_EX),	"RSA_X931_generate_key_ex"},
-{ERR_FUNC(FIPS_F_RSA_X931_GENERATE_KEY),	"RSA_X931_generate_key"},
 {ERR_FUNC(FIPS_F_SSLEAY_RAND_BYTES),	"SSLEAY_RAND_BYTES"},
 {0,NULL}
 	};

crypto/md32_common.h

@@ -1,6 +1,6 @@
 /* crypto/md32_common.h */
 /* ====================================================================
- * Copyright (c) 1999-2002 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -47,10 +47,6 @@
  * OF THE POSSIBILITY OF SUCH DAMAGE.
  * ====================================================================
  *
- * This product includes cryptographic software written by Eric Young
- * (eay@cryptsoft.com).  This product includes software written by Tim
- * Hudson (tjh@cryptsoft.com).
- *
  */
 
 /*
@@ -76,40 +72,27 @@
  *		typedef struct {
  *			...
  *			HASH_LONG	Nl,Nh;
+ *			either {
  *			HASH_LONG	data[HASH_LBLOCK];
+ *			unsigned char	data[HASH_CBLOCK];
+ *			};
  *			unsigned int	num;
  *			...
  *			} HASH_CTX;
+ *	data[] vector is expected to be zeroed upon first call to
+ *	HASH_UPDATE.
  * HASH_UPDATE
  *	name of "Update" function, implemented here.
  * HASH_TRANSFORM
  *	name of "Transform" function, implemented here.
  * HASH_FINAL
  *	name of "Final" function, implemented here.
- * HASH_BLOCK_HOST_ORDER
- *	name of "block" function treating *aligned* input message
- *	in host byte order, implemented externally.
  * HASH_BLOCK_DATA_ORDER
- *	name of "block" function treating *unaligned* input message
+ *	name of "block" function capable of treating *unaligned* input
- *	in original (data) byte order, implemented externally (it
+ *	message in original (data) byte order, implemented externally.
- *	actually is optional if data and host are of the same
- *	"endianess").
  * HASH_MAKE_STRING
  *	macro convering context variables to an ASCII hash string.
  *
- * Optional macros:
- *
- * B_ENDIAN or L_ENDIAN
- *	defines host byte-order.
- * HASH_LONG_LOG2
- *	defaults to 2 if not states otherwise.
- * HASH_LBLOCK
- *	assumed to be HASH_CBLOCK/4 if not stated otherwise.
- * HASH_BLOCK_DATA_ORDER_ALIGNED
- *	alternative "block" function capable of treating
- *	aligned input message in original (data) order,
- *	implemented externally.
- *
  * MD5 example:
  *
  *	#define DATA_ORDER_IS_LITTLE_ENDIAN
@@ -118,11 +101,9 @@
  *	#define HASH_LONG_LOG2		MD5_LONG_LOG2
  *	#define HASH_CTX		MD5_CTX
  *	#define HASH_CBLOCK		MD5_CBLOCK
- *	#define HASH_LBLOCK		MD5_LBLOCK
  *	#define HASH_UPDATE		MD5_Update
  *	#define HASH_TRANSFORM		MD5_Transform
  *	#define HASH_FINAL		MD5_Final
- *	#define HASH_BLOCK_HOST_ORDER	md5_block_host_order
  *	#define HASH_BLOCK_DATA_ORDER	md5_block_data_order
  *
  *					<appro@fy.chalmers.se>
@@ -152,27 +133,9 @@
 #error "HASH_FINAL must be defined!"
 #endif
 
-#ifndef HASH_BLOCK_HOST_ORDER
-#error "HASH_BLOCK_HOST_ORDER must be defined!"
-#endif
-
-#if 0
-/*
- * Moved below as it's required only if HASH_BLOCK_DATA_ORDER_ALIGNED
- * isn't defined.
- */
 #ifndef HASH_BLOCK_DATA_ORDER
 #error "HASH_BLOCK_DATA_ORDER must be defined!"
 #endif
-#endif
-
-#ifndef HASH_LBLOCK
-#define HASH_LBLOCK	(HASH_CBLOCK/4)
-#endif
-
-#ifndef HASH_LONG_LOG2
-#define HASH_LONG_LOG2	2
-#endif
 
 /*
  * Engage compiler specific rotate intrinsic function if available.
@@ -206,7 +169,8 @@
 				: "cc");		\
 			   ret;				\
 			})
-#  elif defined(__powerpc) || defined(__ppc__) || defined(__powerpc64__)
+#  elif defined(_ARCH_PPC) || defined(_ARCH_PPC64) || \
+	defined(__powerpc) || defined(__ppc__) || defined(__powerpc64__)
 #   define ROTATE(a,n)	({ register unsigned int ret;	\
 				asm (			\
 				"rlwinm %0,%1,%2,0,31"	\
@@ -214,80 +178,28 @@
 				: "r"(a), "I"(n));	\
 			   ret;				\
 			})
+#  elif defined(__s390x__)
+#   define ROTATE(a,n) ({ register unsigned int ret;	\
+				asm ("rll %0,%1,%2"	\
+				: "=r"(ret)		\
+				: "r"(a), "I"(n));	\
+			  ret;				\
+			})
 #  endif
 # endif
 #endif /* PEDANTIC */
 
-#if HASH_LONG_LOG2==2	/* Engage only if sizeof(HASH_LONG)== 4 */
-/* A nice byte order reversal from Wei Dai <weidai@eskimo.com> */
-#ifdef ROTATE
-/* 5 instructions with rotate instruction, else 9 */
-#define REVERSE_FETCH32(a,l)	(					\
-		l=*(const HASH_LONG *)(a),				\
-		((ROTATE(l,8)&0x00FF00FF)|(ROTATE((l&0x00FF00FF),24)))	\
-				)
-#else
-/* 6 instructions with rotate instruction, else 8 */
-#define REVERSE_FETCH32(a,l)	(				\
-		l=*(const HASH_LONG *)(a),			\
-		l=(((l>>8)&0x00FF00FF)|((l&0x00FF00FF)<<8)),	\
-		ROTATE(l,16)					\
-				)
-/*
- * Originally the middle line started with l=(((l&0xFF00FF00)>>8)|...
- * It's rewritten as above for two reasons:
- *	- RISCs aren't good at long constants and have to explicitely
- *	  compose 'em with several (well, usually 2) instructions in a
- *	  register before performing the actual operation and (as you
- *	  already realized:-) having same constant should inspire the
- *	  compiler to permanently allocate the only register for it;
- *	- most modern CPUs have two ALUs, but usually only one has
- *	  circuitry for shifts:-( this minor tweak inspires compiler
- *	  to schedule shift instructions in a better way...
- *
- *				<appro@fy.chalmers.se>
- */
-#endif
-#endif
-
 #ifndef ROTATE
 #define ROTATE(a,n)     (((a)<<(n))|(((a)&0xffffffff)>>(32-(n))))
 #endif
 
-/*
- * Make some obvious choices. E.g., HASH_BLOCK_DATA_ORDER_ALIGNED
- * and HASH_BLOCK_HOST_ORDER ought to be the same if input data
- * and host are of the same "endianess". It's possible to mask
- * this with blank #define HASH_BLOCK_DATA_ORDER though...
- *
- *				<appro@fy.chalmers.se>
- */
-#if defined(B_ENDIAN)
-#  if defined(DATA_ORDER_IS_BIG_ENDIAN)
-#    if !defined(HASH_BLOCK_DATA_ORDER_ALIGNED) && HASH_LONG_LOG2==2
-#      define HASH_BLOCK_DATA_ORDER_ALIGNED	HASH_BLOCK_HOST_ORDER
-#    endif
-#  endif
-#elif defined(L_ENDIAN)
-#  if defined(DATA_ORDER_IS_LITTLE_ENDIAN)
-#    if !defined(HASH_BLOCK_DATA_ORDER_ALIGNED) && HASH_LONG_LOG2==2
-#      define HASH_BLOCK_DATA_ORDER_ALIGNED	HASH_BLOCK_HOST_ORDER
-#    endif
-#  endif
-#endif
-
-#if !defined(HASH_BLOCK_DATA_ORDER_ALIGNED)
-#ifndef HASH_BLOCK_DATA_ORDER
-#error "HASH_BLOCK_DATA_ORDER must be defined!"
-#endif
-#endif
-
 #if defined(DATA_ORDER_IS_BIG_ENDIAN)
 
 #ifndef PEDANTIC
 # if defined(__GNUC__) && __GNUC__>=2 && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM)
 #  if ((defined(__i386) || defined(__i386__)) && !defined(I386_ONLY)) || \
       (defined(__x86_64) || defined(__x86_64__))
+#   if !defined(B_ENDIAN)
     /*
      * This gives ~30-40% performance improvement in SHA-256 compiled
      * with gcc [on P4]. Well, first macro to be frank. We can pull
@@ -302,6 +214,11 @@
 				   *((unsigned int *)(c))=r; (c)+=4; r;	})
 #   endif
 #  endif
+# endif
+#endif
+#if defined(__s390__) || defined(__s390x__)
+# define HOST_c2l(c,l) ((l)=*((const unsigned int *)(c)), (c)+=4, (l))
+# define HOST_l2c(l,c) (*((unsigned int *)(c))=(l), (c)+=4, (l))
 #endif
 
 #ifndef HOST_c2l
@@ -311,29 +228,6 @@
 			 l|=(((unsigned long)(*((c)++)))    ),		\
 			 l)
 #endif
-#define HOST_p_c2l(c,l,n)	{					\
-			switch (n) {					\
-			case 0: l =((unsigned long)(*((c)++)))<<24;	\
-			case 1: l|=((unsigned long)(*((c)++)))<<16;	\
-			case 2: l|=((unsigned long)(*((c)++)))<< 8;	\
-			case 3: l|=((unsigned long)(*((c)++)));		\
-				} }
-#define HOST_p_c2l_p(c,l,sc,len) {					\
-			switch (sc) {					\
-			case 0: l =((unsigned long)(*((c)++)))<<24;	\
-				if (--len == 0) break;			\
-			case 1: l|=((unsigned long)(*((c)++)))<<16;	\
-				if (--len == 0) break;			\
-			case 2: l|=((unsigned long)(*((c)++)))<< 8;	\
-				} }
-/* NOTE the pointer is not incremented at the end of this */
-#define HOST_c2l_p(c,l,n)	{					\
-			l=0; (c)+=n;					\
-			switch (n) {					\
-			case 3: l =((unsigned long)(*(--(c))))<< 8;	\
-			case 2: l|=((unsigned long)(*(--(c))))<<16;	\
-			case 1: l|=((unsigned long)(*(--(c))))<<24;	\
-				} }
 #ifndef HOST_l2c
 #define HOST_l2c(l,c)	(*((c)++)=(unsigned char)(((l)>>24)&0xff),	\
 			 *((c)++)=(unsigned char)(((l)>>16)&0xff),	\
@@ -344,6 +238,18 @@
 
 #elif defined(DATA_ORDER_IS_LITTLE_ENDIAN)
 
+#ifndef PEDANTIC
+# if defined(__GNUC__) && __GNUC__>=2 && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM)
+#  if defined(__s390x__)
+#   define HOST_c2l(c,l)	({ asm ("lrv	%0,0(%1)"		\
+					:"=r"(l) : "r"(c));		\
+				   (c)+=4; (l);				})
+#   define HOST_l2c(l,c)	({ asm ("strv	%0,0(%1)"		\
+					: : "r"(l),"r"(c) : "memory");	\
+				   (c)+=4; (l);				})
+#  endif
+# endif
+#endif
 #if defined(__i386) || defined(__i386__) || defined(__x86_64) || defined(__x86_64__)
 # ifndef B_ENDIAN
    /* See comment in DATA_ORDER_IS_BIG_ENDIAN section. */
@@ -359,29 +265,6 @@
 			 l|=(((unsigned long)(*((c)++)))<<24),		\
 			 l)
 #endif
-#define HOST_p_c2l(c,l,n)	{					\
-			switch (n) {					\
-			case 0: l =((unsigned long)(*((c)++)));		\
-			case 1: l|=((unsigned long)(*((c)++)))<< 8;	\
-			case 2: l|=((unsigned long)(*((c)++)))<<16;	\
-			case 3: l|=((unsigned long)(*((c)++)))<<24;	\
-				} }
-#define HOST_p_c2l_p(c,l,sc,len) {					\
-			switch (sc) {					\
-			case 0: l =((unsigned long)(*((c)++)));		\
-				if (--len == 0) break;			\
-			case 1: l|=((unsigned long)(*((c)++)))<< 8;	\
-				if (--len == 0) break;			\
-			case 2: l|=((unsigned long)(*((c)++)))<<16;	\
-				} }
-/* NOTE the pointer is not incremented at the end of this */
-#define HOST_c2l_p(c,l,n)	{					\
-			l=0; (c)+=n;					\
-			switch (n) {					\
-			case 3: l =((unsigned long)(*(--(c))))<<16;	\
-			case 2: l|=((unsigned long)(*(--(c))))<< 8;	\
-			case 1: l|=((unsigned long)(*(--(c))));		\
-				} }
 #ifndef HOST_l2c
 #define HOST_l2c(l,c)	(*((c)++)=(unsigned char)(((l)    )&0xff),	\
 			 *((c)++)=(unsigned char)(((l)>> 8)&0xff),	\
@@ -399,9 +282,9 @@
 int HASH_UPDATE (HASH_CTX *c, const void *data_, size_t len)
 	{
 	const unsigned char *data=data_;
-	register HASH_LONG * p;
+	unsigned char *p;
-	register HASH_LONG l;
+	HASH_LONG l;
-	size_t sw,sc,ew,ec;
+	size_t n;
 
 	if (len==0) return 1;
 
@@ -413,101 +296,43 @@ int HASH_UPDATE (HASH_CTX *c, const void *data_, size_t len)
 	c->Nh+=(len>>29);	/* might cause compiler warning on 16-bit */
 	c->Nl=l;
 
-	if (c->num != 0)
+	n = c->num;
+	if (n != 0)
 		{
-		p=c->data;
+		p=(unsigned char *)c->data;
-		sw=c->num>>2;
-		sc=c->num&0x03;
 
-		if ((c->num+len) >= HASH_CBLOCK)
+		if ((n+len) >= HASH_CBLOCK)
 			{
-			l=p[sw]; HOST_p_c2l(data,l,sc); p[sw++]=l;
+			memcpy (p+n,data,HASH_CBLOCK-n);
-			for (; sw<HASH_LBLOCK; sw++)
+			HASH_BLOCK_DATA_ORDER (c,p,1);
-				{
+			n      = HASH_CBLOCK-n;
-				HOST_c2l(data,l); p[sw]=l;
+			data  += n;
-				}
+			len   -= n;
-			HASH_BLOCK_HOST_ORDER (c,p,1);
+			c->num = 0;
-			len-=(HASH_CBLOCK-c->num);
+			memset (p,0,HASH_CBLOCK);	/* keep it zeroed */
-			c->num=0;
-			/* drop through and do the rest */
 			}
 		else
 			{
-			c->num+=(unsigned int)len;
+			memcpy (p+n,data,len);
-			if ((sc+len) < 4) /* ugly, add char's to a word */
+			c->num += (unsigned int)len;
-				{
-				l=p[sw]; HOST_p_c2l_p(data,l,sc,len); p[sw]=l;
-				}
-			else
-				{
-				ew=(c->num>>2);
-				ec=(c->num&0x03);
-				if (sc)
-					l=p[sw];
-				HOST_p_c2l(data,l,sc);
-				p[sw++]=l;
-				for (; sw < ew; sw++)
-					{
-					HOST_c2l(data,l); p[sw]=l;
-					}
-				if (ec)
-					{
-					HOST_c2l_p(data,l,ec); p[sw]=l;
-					}
-				}
 			return 1;
 			}
 		}
 
-	sw=len/HASH_CBLOCK;
+	n = len/HASH_CBLOCK;
-	if (sw > 0)
+	if (n > 0)
 		{
-#if defined(HASH_BLOCK_DATA_ORDER_ALIGNED)
+		HASH_BLOCK_DATA_ORDER (c,data,n);
-		/*
+		n    *= HASH_CBLOCK;
-		 * Note that HASH_BLOCK_DATA_ORDER_ALIGNED gets defined
+		data += n;
-		 * only if sizeof(HASH_LONG)==4.
+		len  -= n;
-		 */
-		if ((((size_t)data)%4) == 0)
-			{
-			/* data is properly aligned so that we can cast it: */
-			HASH_BLOCK_DATA_ORDER_ALIGNED (c,(const HASH_LONG *)data,sw);
-			sw*=HASH_CBLOCK;
-			data+=sw;
-			len-=sw;
-			}
-		else
-#if !defined(HASH_BLOCK_DATA_ORDER)
-			while (sw--)
-				{
-				memcpy (p=c->data,data,HASH_CBLOCK);
-				HASH_BLOCK_DATA_ORDER_ALIGNED(c,p,1);
-				data+=HASH_CBLOCK;
-				len-=HASH_CBLOCK;
-				}
-#endif
-#endif
-#if defined(HASH_BLOCK_DATA_ORDER)
-			{
-			HASH_BLOCK_DATA_ORDER(c,data,sw);
-			sw*=HASH_CBLOCK;
-			data+=sw;
-			len-=sw;
-			}
-#endif
 		}
 
-	if (len!=0)
+	if (len != 0)
 		{
-		p = c->data;
+		p = (unsigned char *)c->data;
 		c->num = len;
-		ew=len>>2;	/* words to copy */
+		memcpy (p,data,len);
-		ec=len&0x03;
-		for (; ew; ew--,p++)
-			{
-			HOST_c2l(data,l); *p=l;
-			}
-		HOST_c2l_p(data,l,ec);
-		*p=l;
 		}
 	return 1;
 	}
@@ -515,73 +340,38 @@ int HASH_UPDATE (HASH_CTX *c, const void *data_, size_t len)
 
 void HASH_TRANSFORM (HASH_CTX *c, const unsigned char *data)
 	{
-#if defined(HASH_BLOCK_DATA_ORDER_ALIGNED)
-	if ((((size_t)data)%4) == 0)
-		/* data is properly aligned so that we can cast it: */
-		HASH_BLOCK_DATA_ORDER_ALIGNED (c,(const HASH_LONG *)data,1);
-	else
-#if !defined(HASH_BLOCK_DATA_ORDER)
-		{
-		memcpy (c->data,data,HASH_CBLOCK);
-		HASH_BLOCK_DATA_ORDER_ALIGNED (c,c->data,1);
-		}
-#endif
-#endif
-#if defined(HASH_BLOCK_DATA_ORDER)
 	HASH_BLOCK_DATA_ORDER (c,data,1);
-#endif
 	}
 
 
 int HASH_FINAL (unsigned char *md, HASH_CTX *c)
 	{
-	register HASH_LONG *p;
+	unsigned char *p = (unsigned char *)c->data;
-	register unsigned long l;
+	size_t n = c->num;
-	register int i,j;
-	static const unsigned char end[4]={0x80,0x00,0x00,0x00};
-	const unsigned char *cp=end;
 
-	/* c->num should definitly have room for at least one more byte. */
+	p[n] = 0x80; /* there is always room for one */
-	p=c->data;
+	n++;
-	i=c->num>>2;
-	j=c->num&0x03;
 
-#if 0
+	if (n > (HASH_CBLOCK-8))
-	/* purify often complains about the following line as an
-	 * Uninitialized Memory Read.  While this can be true, the
-	 * following p_c2l macro will reset l when that case is true.
-	 * This is because j&0x03 contains the number of 'valid' bytes
-	 * already in p[i].  If and only if j&0x03 == 0, the UMR will
-	 * occur but this is also the only time p_c2l will do
-	 * l= *(cp++) instead of l|= *(cp++)
-	 * Many thanks to Alex Tang <altitude@cic.net> for pickup this
-	 * 'potential bug' */
-#ifdef PURIFY
-	if (j==0) p[i]=0; /* Yeah, but that's not the way to fix it:-) */
-#endif
-	l=p[i];
-#else
-	l = (j==0) ? 0 : p[i];
-#endif
-	HOST_p_c2l(cp,l,j); p[i++]=l; /* i is the next 'undefined word' */
-
-	if (i>(HASH_LBLOCK-2)) /* save room for Nl and Nh */
 		{
-		if (i<HASH_LBLOCK) p[i]=0;
+		memset (p+n,0,HASH_CBLOCK-n);
-		HASH_BLOCK_HOST_ORDER (c,p,1);
+		n=0;
-		i=0;
+		HASH_BLOCK_DATA_ORDER (c,p,1);
 		}
-	for (; i<(HASH_LBLOCK-2); i++)
+	memset (p+n,0,HASH_CBLOCK-8-n);
-		p[i]=0;
 
+	p += HASH_CBLOCK-8;
 #if   defined(DATA_ORDER_IS_BIG_ENDIAN)
-	p[HASH_LBLOCK-2]=c->Nh;
+	(void)HOST_l2c(c->Nh,p);
-	p[HASH_LBLOCK-1]=c->Nl;
+	(void)HOST_l2c(c->Nl,p);
 #elif defined(DATA_ORDER_IS_LITTLE_ENDIAN)
-	p[HASH_LBLOCK-2]=c->Nl;
+	(void)HOST_l2c(c->Nl,p);
-	p[HASH_LBLOCK-1]=c->Nh;
+	(void)HOST_l2c(c->Nh,p);
 #endif
-	HASH_BLOCK_HOST_ORDER (c,p,1);
+	p -= HASH_CBLOCK;
+	HASH_BLOCK_DATA_ORDER (c,p,1);
+	c->num=0;
+	memset (p,0,HASH_CBLOCK);
 
 #ifndef HASH_MAKE_STRING
 #error "HASH_MAKE_STRING must be defined!"
@@ -589,11 +379,6 @@ int HASH_FINAL (unsigned char *md, HASH_CTX *c)
 	HASH_MAKE_STRING(c,md);
 #endif
 
-	c->num=0;
-	/* clear stuff, HASH_BLOCK may be leaving some stuff on the stack
-	 * but I'm not worried :-)
-	OPENSSL_cleanse((void *)c,sizeof(HASH_CTX));
-	 */
 	return 1;
 	}
 

crypto/md4/md4_dgst.c

@@ -84,79 +84,6 @@ FIPS_NON_FIPS_MD_Init(MD4)
 	return 1;
 	}
 
-#ifndef md4_block_host_order
-void md4_block_host_order (MD4_CTX *c, const void *data, size_t num)
-	{
-	const MD4_LONG *X=data;
-	register unsigned MD32_REG_T A,B,C,D;
-
-	A=c->A;
-	B=c->B;
-	C=c->C;
-	D=c->D;
-
-	for (;num--;X+=HASH_LBLOCK)
-		{
-	/* Round 0 */
-	R0(A,B,C,D,X[ 0], 3,0);
-	R0(D,A,B,C,X[ 1], 7,0);
-	R0(C,D,A,B,X[ 2],11,0);
-	R0(B,C,D,A,X[ 3],19,0);
-	R0(A,B,C,D,X[ 4], 3,0);
-	R0(D,A,B,C,X[ 5], 7,0);
-	R0(C,D,A,B,X[ 6],11,0);
-	R0(B,C,D,A,X[ 7],19,0);
-	R0(A,B,C,D,X[ 8], 3,0);
-	R0(D,A,B,C,X[ 9], 7,0);
-	R0(C,D,A,B,X[10],11,0);
-	R0(B,C,D,A,X[11],19,0);
-	R0(A,B,C,D,X[12], 3,0);
-	R0(D,A,B,C,X[13], 7,0);
-	R0(C,D,A,B,X[14],11,0);
-	R0(B,C,D,A,X[15],19,0);
-	/* Round 1 */
-	R1(A,B,C,D,X[ 0], 3,0x5A827999L);
-	R1(D,A,B,C,X[ 4], 5,0x5A827999L);
-	R1(C,D,A,B,X[ 8], 9,0x5A827999L);
-	R1(B,C,D,A,X[12],13,0x5A827999L);
-	R1(A,B,C,D,X[ 1], 3,0x5A827999L);
-	R1(D,A,B,C,X[ 5], 5,0x5A827999L);
-	R1(C,D,A,B,X[ 9], 9,0x5A827999L);
-	R1(B,C,D,A,X[13],13,0x5A827999L);
-	R1(A,B,C,D,X[ 2], 3,0x5A827999L);
-	R1(D,A,B,C,X[ 6], 5,0x5A827999L);
-	R1(C,D,A,B,X[10], 9,0x5A827999L);
-	R1(B,C,D,A,X[14],13,0x5A827999L);
-	R1(A,B,C,D,X[ 3], 3,0x5A827999L);
-	R1(D,A,B,C,X[ 7], 5,0x5A827999L);
-	R1(C,D,A,B,X[11], 9,0x5A827999L);
-	R1(B,C,D,A,X[15],13,0x5A827999L);
-	/* Round 2 */
-	R2(A,B,C,D,X[ 0], 3,0x6ED9EBA1);
-	R2(D,A,B,C,X[ 8], 9,0x6ED9EBA1);
-	R2(C,D,A,B,X[ 4],11,0x6ED9EBA1);
-	R2(B,C,D,A,X[12],15,0x6ED9EBA1);
-	R2(A,B,C,D,X[ 2], 3,0x6ED9EBA1);
-	R2(D,A,B,C,X[10], 9,0x6ED9EBA1);
-	R2(C,D,A,B,X[ 6],11,0x6ED9EBA1);
-	R2(B,C,D,A,X[14],15,0x6ED9EBA1);
-	R2(A,B,C,D,X[ 1], 3,0x6ED9EBA1);
-	R2(D,A,B,C,X[ 9], 9,0x6ED9EBA1);
-	R2(C,D,A,B,X[ 5],11,0x6ED9EBA1);
-	R2(B,C,D,A,X[13],15,0x6ED9EBA1);
-	R2(A,B,C,D,X[ 3], 3,0x6ED9EBA1);
-	R2(D,A,B,C,X[11], 9,0x6ED9EBA1);
-	R2(C,D,A,B,X[ 7],11,0x6ED9EBA1);
-	R2(B,C,D,A,X[15],15,0x6ED9EBA1);
-
-	A = c->A += A;
-	B = c->B += B;
-	C = c->C += C;
-	D = c->D += D;
-		}
-	}
-#endif
-
 #ifndef md4_block_data_order
 #ifdef X
 #undef X
@@ -242,19 +169,3 @@ void md4_block_data_order (MD4_CTX *c, const void *data_, size_t num)
 		}
 	}
 #endif
-
-#ifdef undef
-int printit(unsigned long *l)
-	{
-	int i,ii;
-
-	for (i=0; i<2; i++)
-		{
-		for (ii=0; ii<8; ii++)
-			{
-			fprintf(stderr,"%08lx ",l[i*8+ii]);
-			}
-		fprintf(stderr,"\n");
-		}
-	}
-#endif

crypto/md4/md4_locl.h

@@ -65,43 +65,13 @@
 #define MD4_LONG_LOG2 2 /* default to 32 bits */
 #endif
 
-void md4_block_host_order (MD4_CTX *c, const void *p,size_t num);
 void md4_block_data_order (MD4_CTX *c, const void *p,size_t num);
 
-#if defined(__i386) || defined(__i386__) || defined(_M_IX86) || defined(__INTEL__)
-# if !defined(B_ENDIAN)
-/*
- * *_block_host_order is expected to handle aligned data while
- * *_block_data_order - unaligned. As algorithm and host (x86)
- * are in this case of the same "endianness" these two are
- * otherwise indistinguishable. But normally you don't want to
- * call the same function because unaligned access in places
- * where alignment is expected is usually a "Bad Thing". Indeed,
- * on RISCs you get punished with BUS ERROR signal or *severe*
- * performance degradation. Intel CPUs are in turn perfectly
- * capable of loading unaligned data without such drastic side
- * effect. Yes, they say it's slower than aligned load, but no
- * exception is generated and therefore performance degradation
- * is *incomparable* with RISCs. What we should weight here is
- * costs of unaligned access against costs of aligning data.
- * According to my measurements allowing unaligned access results
- * in ~9% performance improvement on Pentium II operating at
- * 266MHz. I won't be surprised if the difference will be higher
- * on faster systems:-)
- *
- *				<appro@fy.chalmers.se>
- */
-# define md4_block_data_order md4_block_host_order
-# endif
-#endif
-
 #define DATA_ORDER_IS_LITTLE_ENDIAN
 
 #define HASH_LONG		MD4_LONG
-#define HASH_LONG_LOG2		MD4_LONG_LOG2
 #define HASH_CTX		MD4_CTX
 #define HASH_CBLOCK		MD4_CBLOCK
-#define HASH_LBLOCK		MD4_LBLOCK
 #define HASH_UPDATE		MD4_Update
 #define HASH_TRANSFORM		MD4_Transform
 #define HASH_FINAL		MD4_Final
@@ -112,21 +82,7 @@ void md4_block_data_order (MD4_CTX *c, const void *p,size_t num);
 	ll=(c)->C; HOST_l2c(ll,(s));	\
 	ll=(c)->D; HOST_l2c(ll,(s));	\
 	} while (0)
-#define HASH_BLOCK_HOST_ORDER	md4_block_host_order
-#if !defined(L_ENDIAN) || defined(md4_block_data_order)
 #define	HASH_BLOCK_DATA_ORDER	md4_block_data_order
-/*
- * Little-endians (Intel and Alpha) feel better without this.
- * It looks like memcpy does better job than generic
- * md4_block_data_order on copying-n-aligning input data.
- * But frankly speaking I didn't expect such result on Alpha.
- * On the other hand I've got this with egcs-1.0.2 and if
- * program is compiled with another (better?) compiler it
- * might turn out other way around.
- *
- *				<appro@fy.chalmers.se>
- */
-#endif
 
 #include "md32_common.h"
 

crypto/md5/Makefile

@@ -52,24 +52,6 @@ mx86-cof.s: asm/md5-586.pl ../perlasm/x86asm.pl
 mx86-out.s: asm/md5-586.pl ../perlasm/x86asm.pl
 	(cd asm; $(PERL) md5-586.pl a.out $(CFLAGS) > ../$@)
 
-md5-sparcv8plus.o: asm/md5-sparcv9.S
-	$(CC) $(ASFLAGS) -DMD5_BLOCK_DATA_ORDER -c \
-		-o md5-sparcv8plus.o asm/md5-sparcv9.S
-
-# Old GNU assembler doesn't understand V9 instructions, so we
-# hire /usr/ccs/bin/as to do the job. Note that option is called
-# *-gcc27, but even gcc 2>=8 users may experience similar problem
-# if they didn't bother to upgrade GNU assembler. Such users should
-# not choose this option, but be adviced to *remove* GNU assembler
-# or upgrade it.
-md5-sparcv8plus-gcc27.o: asm/md5-sparcv9.S
-	$(CC) $(ASFLAGS) -DMD5_BLOCK_DATA_ORDER -E asm/md5-sparcv9.S | \
-		/usr/ccs/bin/as -xarch=v8plus - -o md5-sparcv8plus-gcc27.o
-
-md5-sparcv9.o: asm/md5-sparcv9.S
-	$(CC) $(ASFLAGS) -DMD5_BLOCK_DATA_ORDER -c \
-		-o md5-sparcv9.o asm/md5-sparcv9.S
-
 md5-x86_64.s:	asm/md5-x86_64.pl;	$(PERL) asm/md5-x86_64.pl $@
 
 files:

crypto/md5/asm/md5-586.pl

@@ -29,7 +29,7 @@ $X="esi";
  0, 7, 14, 5, 12, 3, 10, 1, 8, 15, 6, 13, 4, 11, 2, 9,	# R3
  );
 
-&md5_block("md5_block_asm_host_order");
+&md5_block("md5_block_asm_data_order");
 &asm_finish();
 
 sub Np

crypto/md5/asm/md5-x86_64.pl

@@ -111,9 +111,9 @@ $code .= <<EOF;
 .text
 .align 16
 
-.globl md5_block_asm_host_order
+.globl md5_block_asm_data_order
-.type md5_block_asm_host_order,\@function,3
+.type md5_block_asm_data_order,\@function,3
-md5_block_asm_host_order:
+md5_block_asm_data_order:
 	push	%rbp
 	push	%rbx
 	push	%r14
@@ -237,7 +237,7 @@ $code .= <<EOF;
 	pop	%rbx
 	pop	%rbp
 	ret
-.size md5_block_asm_host_order,.-md5_block_asm_host_order
+.size md5_block_asm_data_order,.-md5_block_asm_data_order
 EOF
 
 print $code;

crypto/md5/md5_dgst.c

@@ -84,96 +84,6 @@ FIPS_NON_FIPS_MD_Init(MD5)
 	return 1;
 	}
 
-#ifndef md5_block_host_order
-void md5_block_host_order (MD5_CTX *c, const void *data, size_t num)
-	{
-	const MD5_LONG *X=data;
-	register unsigned MD32_REG_T A,B,C,D;
-
-	A=c->A;
-	B=c->B;
-	C=c->C;
-	D=c->D;
-
-	for (;num--;X+=HASH_LBLOCK)
-		{
-	/* Round 0 */
-	R0(A,B,C,D,X[ 0], 7,0xd76aa478L);
-	R0(D,A,B,C,X[ 1],12,0xe8c7b756L);
-	R0(C,D,A,B,X[ 2],17,0x242070dbL);
-	R0(B,C,D,A,X[ 3],22,0xc1bdceeeL);
-	R0(A,B,C,D,X[ 4], 7,0xf57c0fafL);
-	R0(D,A,B,C,X[ 5],12,0x4787c62aL);
-	R0(C,D,A,B,X[ 6],17,0xa8304613L);
-	R0(B,C,D,A,X[ 7],22,0xfd469501L);
-	R0(A,B,C,D,X[ 8], 7,0x698098d8L);
-	R0(D,A,B,C,X[ 9],12,0x8b44f7afL);
-	R0(C,D,A,B,X[10],17,0xffff5bb1L);
-	R0(B,C,D,A,X[11],22,0x895cd7beL);
-	R0(A,B,C,D,X[12], 7,0x6b901122L);
-	R0(D,A,B,C,X[13],12,0xfd987193L);
-	R0(C,D,A,B,X[14],17,0xa679438eL);
-	R0(B,C,D,A,X[15],22,0x49b40821L);
-	/* Round 1 */
-	R1(A,B,C,D,X[ 1], 5,0xf61e2562L);
-	R1(D,A,B,C,X[ 6], 9,0xc040b340L);
-	R1(C,D,A,B,X[11],14,0x265e5a51L);
-	R1(B,C,D,A,X[ 0],20,0xe9b6c7aaL);
-	R1(A,B,C,D,X[ 5], 5,0xd62f105dL);
-	R1(D,A,B,C,X[10], 9,0x02441453L);
-	R1(C,D,A,B,X[15],14,0xd8a1e681L);
-	R1(B,C,D,A,X[ 4],20,0xe7d3fbc8L);
-	R1(A,B,C,D,X[ 9], 5,0x21e1cde6L);
-	R1(D,A,B,C,X[14], 9,0xc33707d6L);
-	R1(C,D,A,B,X[ 3],14,0xf4d50d87L);
-	R1(B,C,D,A,X[ 8],20,0x455a14edL);
-	R1(A,B,C,D,X[13], 5,0xa9e3e905L);
-	R1(D,A,B,C,X[ 2], 9,0xfcefa3f8L);
-	R1(C,D,A,B,X[ 7],14,0x676f02d9L);
-	R1(B,C,D,A,X[12],20,0x8d2a4c8aL);
-	/* Round 2 */
-	R2(A,B,C,D,X[ 5], 4,0xfffa3942L);
-	R2(D,A,B,C,X[ 8],11,0x8771f681L);
-	R2(C,D,A,B,X[11],16,0x6d9d6122L);
-	R2(B,C,D,A,X[14],23,0xfde5380cL);
-	R2(A,B,C,D,X[ 1], 4,0xa4beea44L);
-	R2(D,A,B,C,X[ 4],11,0x4bdecfa9L);
-	R2(C,D,A,B,X[ 7],16,0xf6bb4b60L);
-	R2(B,C,D,A,X[10],23,0xbebfbc70L);
-	R2(A,B,C,D,X[13], 4,0x289b7ec6L);
-	R2(D,A,B,C,X[ 0],11,0xeaa127faL);
-	R2(C,D,A,B,X[ 3],16,0xd4ef3085L);
-	R2(B,C,D,A,X[ 6],23,0x04881d05L);
-	R2(A,B,C,D,X[ 9], 4,0xd9d4d039L);
-	R2(D,A,B,C,X[12],11,0xe6db99e5L);
-	R2(C,D,A,B,X[15],16,0x1fa27cf8L);
-	R2(B,C,D,A,X[ 2],23,0xc4ac5665L);
-	/* Round 3 */
-	R3(A,B,C,D,X[ 0], 6,0xf4292244L);
-	R3(D,A,B,C,X[ 7],10,0x432aff97L);
-	R3(C,D,A,B,X[14],15,0xab9423a7L);
-	R3(B,C,D,A,X[ 5],21,0xfc93a039L);
-	R3(A,B,C,D,X[12], 6,0x655b59c3L);
-	R3(D,A,B,C,X[ 3],10,0x8f0ccc92L);
-	R3(C,D,A,B,X[10],15,0xffeff47dL);
-	R3(B,C,D,A,X[ 1],21,0x85845dd1L);
-	R3(A,B,C,D,X[ 8], 6,0x6fa87e4fL);
-	R3(D,A,B,C,X[15],10,0xfe2ce6e0L);
-	R3(C,D,A,B,X[ 6],15,0xa3014314L);
-	R3(B,C,D,A,X[13],21,0x4e0811a1L);
-	R3(A,B,C,D,X[ 4], 6,0xf7537e82L);
-	R3(D,A,B,C,X[11],10,0xbd3af235L);
-	R3(C,D,A,B,X[ 2],15,0x2ad7d2bbL);
-	R3(B,C,D,A,X[ 9],21,0xeb86d391L);
-
-	A = c->A += A;
-	B = c->B += B;
-	C = c->C += C;
-	D = c->D += D;
-		}
-	}
-#endif
-
 #ifndef md5_block_data_order
 #ifdef X
 #undef X
@@ -276,19 +186,3 @@ void md5_block_data_order (MD5_CTX *c, const void *data_, size_t num)
 		}
 	}
 #endif
-
-#ifdef undef
-int printit(unsigned long *l)
-	{
-	int i,ii;
-
-	for (i=0; i<2; i++)
-		{
-		for (ii=0; ii<8; ii++)
-			{
-			fprintf(stderr,"%08lx ",l[i*8+ii]);
-			}
-		fprintf(stderr,"\n");
-		}
-	}
-#endif

crypto/md5/md5_locl.h

@@ -66,53 +66,19 @@
 #endif
 
 #ifdef MD5_ASM
-# if defined(__i386) || defined(__i386__) || defined(_M_IX86) || defined(__INTEL__) || defined(__x86_64) || defined(__x86_64__)
+# if defined(__i386) || defined(__i386__) || defined(_M_IX86) || defined(__INTEL__) || \
-#  if !defined(B_ENDIAN)
+     defined(__x86_64) || defined(__x86_64__) || defined(_M_AMD64) || defined(_M_X64)
-#   define md5_block_host_order md5_block_asm_host_order
+#  define md5_block_data_order md5_block_asm_data_order
-#  endif
-# elif defined(__sparc) && defined(OPENSSL_SYS_ULTRASPARC)
-   void md5_block_asm_data_order_aligned (MD5_CTX *c, const MD5_LONG *p,size_t num);
-#  define HASH_BLOCK_DATA_ORDER_ALIGNED md5_block_asm_data_order_aligned
 # endif
 #endif
 
-void md5_block_host_order (MD5_CTX *c, const void *p,size_t num);
 void md5_block_data_order (MD5_CTX *c, const void *p,size_t num);
 
-#if defined(__i386) || defined(__i386__) || defined(_M_IX86) || defined(__INTEL__) || defined(__x86_64) || defined(__x86_64__)
-# if !defined(B_ENDIAN)
-/*
- * *_block_host_order is expected to handle aligned data while
- * *_block_data_order - unaligned. As algorithm and host (x86)
- * are in this case of the same "endianness" these two are
- * otherwise indistinguishable. But normally you don't want to
- * call the same function because unaligned access in places
- * where alignment is expected is usually a "Bad Thing". Indeed,
- * on RISCs you get punished with BUS ERROR signal or *severe*
- * performance degradation. Intel CPUs are in turn perfectly
- * capable of loading unaligned data without such drastic side
- * effect. Yes, they say it's slower than aligned load, but no
- * exception is generated and therefore performance degradation
- * is *incomparable* with RISCs. What we should weight here is
- * costs of unaligned access against costs of aligning data.
- * According to my measurements allowing unaligned access results
- * in ~9% performance improvement on Pentium II operating at
- * 266MHz. I won't be surprised if the difference will be higher
- * on faster systems:-)
- *
- *				<appro@fy.chalmers.se>
- */
-# define md5_block_data_order md5_block_host_order
-# endif
-#endif
-
 #define DATA_ORDER_IS_LITTLE_ENDIAN
 
 #define HASH_LONG		MD5_LONG
-#define HASH_LONG_LOG2		MD5_LONG_LOG2
 #define HASH_CTX		MD5_CTX
 #define HASH_CBLOCK		MD5_CBLOCK
-#define HASH_LBLOCK		MD5_LBLOCK
 #define HASH_UPDATE		MD5_Update
 #define HASH_TRANSFORM		MD5_Transform
 #define HASH_FINAL		MD5_Final
@@ -123,21 +89,7 @@ void md5_block_data_order (MD5_CTX *c, const void *p,size_t num);
 	ll=(c)->C; HOST_l2c(ll,(s));	\
 	ll=(c)->D; HOST_l2c(ll,(s));	\
 	} while (0)
-#define HASH_BLOCK_HOST_ORDER	md5_block_host_order
-#if !defined(L_ENDIAN) || defined(md5_block_data_order)
 #define	HASH_BLOCK_DATA_ORDER	md5_block_data_order
-/*
- * Little-endians (Intel and Alpha) feel better without this.
- * It looks like memcpy does better job than generic
- * md5_block_data_order on copying-n-aligning input data.
- * But frankly speaking I didn't expect such result on Alpha.
- * On the other hand I've got this with egcs-1.0.2 and if
- * program is compiled with another (better?) compiler it
- * might turn out other way around.
- *
- *				<appro@fy.chalmers.se>
- */
-#endif
 
 #include "md32_common.h"
 

crypto/mem_clr.c

@@ -64,12 +64,14 @@ unsigned char cleanse_ctr = 0;
 void OPENSSL_cleanse(void *ptr, size_t len)
 	{
 	unsigned char *p = ptr;
-	size_t loop = len;
+	size_t loop = len, ctr = cleanse_ctr;
 	while(loop--)
 		{
-		*(p++) = cleanse_ctr;
+		*(p++) = (unsigned char)ctr;
-		cleanse_ctr += (17 + (unsigned char)((unsigned long)p & 0xF));
+		ctr += (17 + ((size_t)p & 0xF));
 		}
-	if(memchr(ptr, cleanse_ctr, len))
+	p=memchr(ptr, (unsigned char)ctr, len);
-		cleanse_ctr += 63;
+	if(p)
+		ctr += (63 + (size_t)p);
+	cleanse_ctr = (unsigned char)ctr;
 	}

crypto/perlasm/x86_64-xlate.pl

@@ -57,7 +57,13 @@
 #		lea		.Label-.Lpic_point(%rcx),%rbp
 
 my $output = shift;
-open STDOUT,">$output" || die "can't open $output: $!";
+
+{ my ($stddev,$stdino,@junk)=stat(STDOUT);
+  my ($outdev,$outino,@junk)=stat($output);
+
+    open STDOUT,">$output" || die "can't open $output: $!"
+	if ($stddev!=$outdev || $stdino!=$outino);
+}
 
 my $masm=1 if ($output =~ /\.asm/);
 
@@ -70,7 +76,7 @@ my $current_function;
 	local	*line = shift;
 	undef	$ret;
 
-	if ($line =~ /^([a-z]+)/i) {
+	if ($line =~ /^([a-z][a-z0-9]*)/i) {
 	    $self->{op} = $1;
 	    $ret = $self;
 	    $line = substr($line,@+[0]); $line =~ s/^\s+//;
@@ -79,6 +85,8 @@ my $current_function;
 	    if ($self->{op} =~ /(movz)b.*/) {	# movz is pain...
 		$self->{op} = $1;
 		$self->{sz} = "b";
+	    } elsif ($self->{op} =~ /call/) {
+		$self->{sz} = ""
 	    } elsif ($self->{op} =~ /([a-z]{3,})([qlwb])/) {
 		$self->{op} = $1;
 		$self->{sz} = $2;
@@ -95,8 +103,10 @@ my $current_function;
     sub out {
 	my $self = shift;
 	if (!$masm) {
-	    if ($self->{op} eq "movz") {	# movz in pain...
+	    if ($self->{op} eq "movz") {	# movz is pain...
 		sprintf "%s%s%s",$self->{op},$self->{sz},shift;
+	    } elsif ($self->{op} =~ /^set/) { 
+		"$self->{op}";
 	    } elsif ($self->{op} eq "ret") {
 	    	".byte	0xf3,0xc3";
 	    } else {
@@ -133,6 +143,10 @@ my $current_function;
     	my $self = shift;
 
 	if (!$masm) {
+	    # Solaris /usr/ccs/bin/as can't handle multiplications
+	    # in $self->{value}
+	    $self->{value} =~ s/(?<![0-9a-f])(0[x0-9a-f]+)/oct($1)/egi;
+	    $self->{value} =~ s/([0-9]+\s*[\*\/\%]\s*[0-9]+)/eval($1)/eg;
 	    sprintf "\$%s",$self->{value};
 	} else {
 	    $self->{value} =~ s/0x([0-9a-f]+)/0$1h/ig;
@@ -163,14 +177,16 @@ my $current_function;
     	my $self = shift;
 	my $sz = shift;
 
+	# Silently convert all EAs to 64-bit. This is required for
+	# elder GNU assembler and results in more compact code,
+	# *but* most importantly AES module depends on this feature!
+	$self->{index} =~ s/^[er](.?[0-9xpi])[d]?$/r\1/;
+	$self->{base}  =~ s/^[er](.?[0-9xpi])[d]?$/r\1/;
+
 	if (!$masm) {
-	    # elder GNU assembler insists on 64-bit EAs:-(
-	    # on pros side, this results in more compact code:-)
-	    $self->{index} =~ s/^[er](.?[0-9xp])[d]?$/r\1/;
-	    $self->{base}  =~ s/^[er](.?[0-9xp])[d]?$/r\1/;
 	    # Solaris /usr/ccs/bin/as can't handle multiplications
 	    # in $self->{label}
-	    $self->{label} =~ s/(?<![0-9a-f])(0[x0-9a-f]+)/oct($1)/eg;
+	    $self->{label} =~ s/(?<![0-9a-f])(0[x0-9a-f]+)/oct($1)/egi;
 	    $self->{label} =~ s/([0-9]+\s*[\*\/\%]\s*[0-9]+)/eval($1)/eg;
 
 	    if (defined($self->{index})) {
@@ -192,6 +208,8 @@ my $current_function;
 					$self->{label},
 					$self->{index},$self->{scale},
 					$self->{base};
+	    } elsif ($self->{base} eq "rip") {
+		sprintf "%s PTR %s",$szmap{$sz},$self->{label};
 	    } else {
 		sprintf "%s PTR %s[%s]",$szmap{$sz},
 					$self->{label},$self->{base};
@@ -317,6 +335,10 @@ my $current_function;
 		$line =~ s/\@function.*/\@function/;
 		if ($line =~ /\.picmeup\s+(%r[\w]+)/i) {
 		    $self->{value} = sprintf "\t.long\t0x%x,0x90000000",$opcode{$1};
+		} elsif ($line =~ /\.asciz\s+"(.*)"$/) {
+		    $self->{value} = ".byte\t".join(",",unpack("C*",$1),0);
+		} elsif ($line =~ /\.extern/) {
+		    $self->{value} = ""; # swallow extern
 		} else {
 		    $self->{value} = $line;
 		}
@@ -338,6 +360,7 @@ my $current_function;
 				    $self->{value} = $v;
 				    last;
 				  };
+		/\.extern/  && do { $self->{value} = "EXTRN\t".$line.":BYTE"; last;  };
 		/\.globl/   && do { $self->{value} = "PUBLIC\t".$line; last; };
 		/\.type/    && do { ($sym,$type,$narg) = split(',',$line);
 				    if ($type eq "\@function") {
@@ -372,6 +395,19 @@ my $current_function;
 		/\.picmeup/ && do { $self->{value} = sprintf"\tDD\t 0%Xh,090000000h",$opcode{$line};
 				    last;
 				  };
+		/\.asciz/   && do { if ($line =~ /^"(.*)"$/) {
+					my @str=unpack("C*",$1);
+					push @str,0;
+					while ($#str>15) {
+					    $self->{value}.="DB\t"
+						.join(",",@str[0..15])."\n";
+					    foreach (0..15) { shift @str; }
+					}
+					$self->{value}.="DB\t"
+						.join(",",@str) if (@str);
+				    }
+				    last;
+				  };
 	    }
 	    $line = "";
 	}
@@ -480,7 +516,10 @@ close STDOUT;
 # arguments passed to callee, *but* not less than 4! This means that
 # upon function entry point 5th argument resides at 40(%rsp), as well
 # as that 32 bytes from 8(%rsp) can always be used as temporal
-# storage [without allocating a frame].
+# storage [without allocating a frame]. One can actually argue that
+# one can assume a "red zone" above stack pointer under Win64 as well.
+# Point is that at apparently no occasion Windows kernel would alter
+# the area above user stack pointer in true asynchronous manner...
 #
 # All the above means that if assembler programmer adheres to Unix
 # register and stack layout, but disregards the "red zone" existense,

crypto/perlasm/x86asm.pl

@@ -1,48 +1,131 @@
-#!/usr/local/bin/perl
+#!/usr/bin/env perl
 
 # require 'x86asm.pl';
-# &asm_init("cpp","des-586.pl");
+# &asm_init(<flavor>,"des-586.pl"[,$i386only]);
-# XXX
+# &function_begin("foo");
-# XXX
+# ...
-# main'asm_finish
+# &function_end("foo");
+# &asm_finish
 
-sub main'asm_finish
+# AUTOLOAD is this context has quite unpleasant side effect, namely
-	{
+# that typos in function calls effectively go to assembler output,
-	&file_end();
+# but on the pros side we don't have to implement one subroutine per
-	&asm_finish_cpp() if $cpp;
+# each opcode...
-	print &asm_get_output();
+sub ::AUTOLOAD
+{ my $opcode = $AUTOLOAD;
+
+    die "more than 2 arguments passed to $opcode" if ($#_>1);
+
+    $opcode =~ s/.*:://;
+    if    ($opcode =~ /^push/) { $stack+=4; }
+    elsif ($opcode =~ /^pop/)  { $stack-=4; }
+
+    &generic($opcode,@_) or die "undefined subroutine \&$AUTOLOAD";
+}
+
+$out=();
+$i386=0;
+
+sub ::emit
+{ my $opcode=shift;
+
+    if ($#_==-1)    { push(@out,"\t$opcode\n");				}
+    else            { push(@out,"\t$opcode\t".join(',',@_)."\n");	}
+}
+
+sub ::LB
+{   $_[0] =~ m/^e?([a-d])x$/o or die "$_[0] does not have a 'low byte'";
+  $1."l";
+}
+sub ::HB
+{   $_[0] =~ m/^e?([a-d])x$/o or die "$_[0] does not have a 'high byte'";
+  $1."h";
+}
+sub ::stack_push{ my $num=$_[0]*4; $stack+=$num; &sub("esp",$num);	}
+sub ::stack_pop	{ my $num=$_[0]*4; $stack-=$num; &add("esp",$num);	}
+sub ::blindpop	{ &pop($_[0]); $stack+=4;				}
+sub ::wparam	{ &DWP($stack+4*$_[0],"esp");				}
+sub ::swtmp	{ &DWP(4*$_[0],"esp");					}
+
+sub ::bswap
+{   if ($i386)	# emulate bswap for i386
+    {	&comment("bswap @_");
+	&xchg(&HB(@_),&LB(@_));
+	&ror (@_,16);
+	&xchg(&HB(@_),&LB(@_));
     }
+    else
+    {	&generic("bswap",@_);	}
+}
+# These are made-up opcodes introduced over the years essentially
+# by ignorance, just alias them to real ones...
+sub ::movb	{ &mov(@_);	}
+sub ::xorb	{ &xor(@_);	}
+sub ::rotl	{ &rol(@_);	}
+sub ::rotr	{ &ror(@_);	}
+sub ::exch	{ &xchg(@_);	}
+sub ::halt	{ &hlt;		}
+
+sub ::function_begin
+{   &function_begin_B(@_);
+    $stack=4;
+    &push("ebp");
+    &push("ebx");
+    &push("esi");
+    &push("edi");
+}
+
+sub ::function_end
+{   &pop("edi");
+    &pop("esi");
+    &pop("ebx");
+    &pop("ebp");
+    &ret();
+    $stack=0;
+    &function_end_B(@_);
+}
+
+sub ::function_end_A
+{   &pop("edi");
+    &pop("esi");
+    &pop("ebx");
+    &pop("ebp");
+    &ret();
+    $stack+=16;	# readjust esp as if we didn't pop anything
+}
+
+sub ::asciz {   foreach (@_) { &data_byte(unpack("C*",$_),0); }   }
+
+sub ::asm_finish
+{   &file_end();
+    print @out;
+}
+
+sub ::asm_init
+{ my ($type,$fn,$cpu)=@_;
 
-sub main'asm_init
-	{
-	($type,$fn,$i386)=@_;
     $filename=$fn;
+    $i386=$cpu;
 
     $elf=$cpp=$coff=$aout=$win32=$netware=$mwerks=0;
-	if (	($type eq "elf"))
+    if    (($type eq "elf"))
     {	$elf=1;			require "x86unix.pl";	}
-	elsif (	($type eq "a.out"))
+    elsif (($type eq "a\.out"))
     {	$aout=1;		require "x86unix.pl";	}
-	elsif (	($type eq "coff" or $type eq "gaswin"))
+    elsif (($type eq "coff" or $type eq "gaswin"))
     {	$coff=1;		require "x86unix.pl";	}
-	elsif (	($type eq "cpp"))
+    elsif (($type eq "win32n"))
-		{ $cpp=1; require "x86unix.pl"; }
-	elsif (	($type eq "win32"))
-		{ $win32=1; require "x86ms.pl"; }
-	elsif (	($type eq "win32n"))
     {	$win32=1;		require "x86nasm.pl";	}
-	elsif (	($type eq "nw-nasm"))
+    elsif (($type eq "nw-nasm"))
     {	$netware=1;		require "x86nasm.pl";	}
-	elsif (	($type eq "nw-mwasm"))
+    elsif (($type eq "nw-mwasm"))
     {	$netware=1; $mwerks=1;	require "x86nasm.pl";	}
     else
-		{
+    {	print STDERR <<"EOF";
-		print STDERR <<"EOF";
 Pick one target type from
 	elf	- Linux, FreeBSD, Solaris x86, etc.
-	a.out	- OpenBSD, DJGPP, etc.
+	a.out	- DJGPP, elder OpenBSD, etc.
 	coff	- GAS/COFF such as Win32 targets
-	win32	- Windows 95/Windows NT
 	win32n	- Windows 95/Windows NT NASM format
 	nw-nasm - NetWare NASM format
 	nw-mwasm- NetWare Metrowerks Assembler
@@ -53,78 +136,8 @@ EOF
     $pic=0;
     for (@ARGV) { $pic=1 if (/\-[fK]PIC/i); }
 
-	&asm_init_output();
-
-&comment("Don't even think of reading this code");
-&comment("It was automatically generated by $filename");
-&comment("Which is a perl program used to generate the x86 assember for");
-&comment("any of ELF, a.out, COFF, Win32, ...");
-&comment("eric <eay\@cryptsoft.com>");
-&comment("");
-
     $filename =~ s/\.pl$//;
     &file($filename);
-	}
+}
-
-sub asm_finish_cpp
-	{
-	return unless $cpp;
-
-	local($tmp,$i);
-	foreach $i (&get_labels())
-		{
-		$tmp.="#define $i _$i\n";
-		}
-	print <<"EOF";
-/* Run the C pre-processor over this file with one of the following defined
- * ELF - elf object files,
- * OUT - a.out object files,
- * BSDI - BSDI style a.out object files
- * SOL - Solaris style elf
- */
-
-#define TYPE(a,b)       .type   a,b
-#define SIZE(a,b)       .size   a,b
-
-#if defined(OUT) || (defined(BSDI) && !defined(ELF))
-$tmp
-#endif
-
-#ifdef OUT
-#define OK	1
-#define ALIGN	4
-#if defined(__CYGWIN__) || defined(__DJGPP__) || (__MINGW32__)
-#undef SIZE
-#undef TYPE
-#define SIZE(a,b)
-#define TYPE(a,b)	.def a; .scl 2; .type 32; .endef
-#endif /* __CYGWIN || __DJGPP */
-#endif
-
-#if defined(BSDI) && !defined(ELF)
-#define OK              1
-#define ALIGN           4
-#undef SIZE
-#undef TYPE
-#define SIZE(a,b)
-#define TYPE(a,b)
-#endif
-
-#if defined(ELF) || defined(SOL)
-#define OK              1
-#define ALIGN           16
-#endif
-
-#ifndef OK
-You need to define one of
-ELF - elf systems - linux-elf, NetBSD and DG-UX
-OUT - a.out systems - linux-a.out and FreeBSD
-SOL - solaris systems, which are elf with strange comment lines
-BSDI - a.out with a very primative version of as.
-#endif
-
-/* Let the Assembler begin :-) */
-EOF
-	}
 
 1;

crypto/perlasm/x86ms.pl

@@ -1,464 +0,0 @@
-#!/usr/local/bin/perl
-
-package x86ms;
-
-$label="L000";
-
-%lb=(	'eax',	'al',
-	'ebx',	'bl',
-	'ecx',	'cl',
-	'edx',	'dl',
-	'ax',	'al',
-	'bx',	'bl',
-	'cx',	'cl',
-	'dx',	'dl',
-	);
-
-%hb=(	'eax',	'ah',
-	'ebx',	'bh',
-	'ecx',	'ch',
-	'edx',	'dh',
-	'ax',	'ah',
-	'bx',	'bh',
-	'cx',	'ch',
-	'dx',	'dh',
-	);
-
-sub main'asm_init_output { @out=(); }
-sub main'asm_get_output { return(@out); }
-sub main'get_labels { return(@labels); }
-sub main'external_label
-{
-	push(@labels,@_);
-	foreach (@_) {
-		push(@out, "EXTRN\t_$_:DWORD\n");
-	}
-}
-
-sub main'LB
-	{
-	(defined($lb{$_[0]})) || die "$_[0] does not have a 'low byte'\n";
-	return($lb{$_[0]});
-	}
-
-sub main'HB
-	{
-	(defined($hb{$_[0]})) || die "$_[0] does not have a 'high byte'\n";
-	return($hb{$_[0]});
-	}
-
-sub main'BP
-	{
-	&get_mem("BYTE",@_);
-	}
-
-sub main'DWP
-	{
-	&get_mem("DWORD",@_);
-	}
-
-sub main'QWP
-	{
-	&get_mem("QWORD",@_);
-	}
-
-sub main'BC
-	{
-	return @_;
-	}
-
-sub main'DWC
-	{
-	return @_;
-	}
-
-sub main'stack_push
-	{
-	local($num)=@_;
-	$stack+=$num*4;
-	&main'sub("esp",$num*4);
-	}
-
-sub main'stack_pop
-	{
-	local($num)=@_;
-	$stack-=$num*4;
-	&main'add("esp",$num*4);
-	}
-
-sub get_mem
-	{
-	local($size,$addr,$reg1,$reg2,$idx)=@_;
-	local($t,$post);
-	local($ret)="$size PTR ";
-
-	$addr =~ s/^\s+//;
-	if ($addr =~ /^(.+)\+(.+)$/)
-		{
-		$reg2=&conv($1);
-		$addr="_$2";
-		}
-	elsif ($addr =~ /^[_a-z][_a-z0-9]*$/i)
-		{
-		$addr="_$addr";
-		}
-
-	if ($addr =~ /^.+\-.+$/) { $addr="($addr)"; }
-
-	$reg1="$regs{$reg1}" if defined($regs{$reg1});
-	$reg2="$regs{$reg2}" if defined($regs{$reg2});
-	if (($addr ne "") && ($addr ne 0))
-		{
-		if ($addr !~ /^-/)
-			{ $ret.=$addr; }
-		else	{ $post=$addr; }
-		}
-	if ($reg2 ne "")
-		{
-		$t="";
-		$t="*$idx" if ($idx != 0);
-		$reg1="+".$reg1 if ("$reg1$post" ne "");
-		$ret.="[$reg2$t$reg1$post]";
-		}
-	else
-		{
-		$ret.="[$reg1$post]"
-		}
-	$ret =~ s/\[\]//;	# in case $addr was the only argument
-	return($ret);
-	}
-
-sub main'mov	{ &out2("mov",@_); }
-sub main'movb	{ &out2("mov",@_); }
-sub main'and	{ &out2("and",@_); }
-sub main'or	{ &out2("or",@_); }
-sub main'shl	{ &out2("shl",@_); }
-sub main'shr	{ &out2("shr",@_); }
-sub main'xor	{ &out2("xor",@_); }
-sub main'xorb	{ &out2("xor",@_); }
-sub main'add	{ &out2("add",@_); }
-sub main'adc	{ &out2("adc",@_); }
-sub main'sub	{ &out2("sub",@_); }
-sub main'sbb	{ &out2("sbb",@_); }
-sub main'rotl	{ &out2("rol",@_); }
-sub main'rotr	{ &out2("ror",@_); }
-sub main'exch	{ &out2("xchg",@_); }
-sub main'cmp	{ &out2("cmp",@_); }
-sub main'lea	{ &out2("lea",@_); }
-sub main'mul	{ &out1("mul",@_); }
-sub main'div	{ &out1("div",@_); }
-sub main'dec	{ &out1("dec",@_); }
-sub main'inc	{ &out1("inc",@_); }
-sub main'jmp	{ &out1("jmp",@_); }
-sub main'jmp_ptr { &out1p("jmp",@_); }
-sub main'je	{ &out1("je",@_); }
-sub main'jle	{ &out1("jle",@_); }
-sub main'jz	{ &out1("jz",@_); }
-sub main'jge	{ &out1("jge",@_); }
-sub main'jl	{ &out1("jl",@_); }
-sub main'ja	{ &out1("ja",@_); }
-sub main'jae	{ &out1("jae",@_); }
-sub main'jb	{ &out1("jb",@_); }
-sub main'jbe	{ &out1("jbe",@_); }
-sub main'jc	{ &out1("jc",@_); }
-sub main'jnc	{ &out1("jnc",@_); }
-sub main'jnz	{ &out1("jnz",@_); }
-sub main'jne	{ &out1("jne",@_); }
-sub main'jno	{ &out1("jno",@_); }
-sub main'push	{ &out1("push",@_); $stack+=4; }
-sub main'pop	{ &out1("pop",@_); $stack-=4; }
-sub main'pushf	{ &out0("pushfd"); $stack+=4; }
-sub main'popf	{ &out0("popfd"); $stack-=4; }
-sub main'bswap	{ &out1("bswap",@_); &using486(); }
-sub main'not	{ &out1("not",@_); }
-sub main'call	{ &out1("call",($_[0]=~/^\$L/?'':'_').$_[0]); }
-sub main'call_ptr { &out1p("call",@_); }
-sub main'ret	{ &out0("ret"); }
-sub main'nop	{ &out0("nop"); }
-sub main'test	{ &out2("test",@_); }
-sub main'bt	{ &out2("bt",@_); }
-sub main'leave	{ &out0("leave"); }
-sub main'cpuid  { &out0("DW\t0A20Fh"); }
-sub main'rdtsc  { &out0("DW\t0310Fh"); }
-sub main'halt	{ &out0("hlt"); }
-sub main'movz	{ &out2("movzx",@_); }
-sub main'neg	{ &out1("neg",@_); }
-sub main'cld	{ &out0("cld"); }
-
-# SSE2
-sub main'emms	{ &out0("emms"); }
-sub main'movd	{ &out2("movd",@_); }
-sub main'movq	{ &out2("movq",@_); }
-sub main'movdqu	{ &out2("movdqu",@_); }
-sub main'movdqa	{ &out2("movdqa",@_); }
-sub main'movdq2q{ &out2("movdq2q",@_); }
-sub main'movq2dq{ &out2("movq2dq",@_); }
-sub main'paddq	{ &out2("paddq",@_); }
-sub main'pmuludq{ &out2("pmuludq",@_); }
-sub main'psrlq	{ &out2("psrlq",@_); }
-sub main'psllq	{ &out2("psllq",@_); }
-sub main'pxor	{ &out2("pxor",@_); }
-sub main'por	{ &out2("por",@_); }
-sub main'pand	{ &out2("pand",@_); }
-
-sub out2
-	{
-	local($name,$p1,$p2)=@_;
-	local($l,$t);
-
-	push(@out,"\t$name\t");
-	$t=&conv($p1).",";
-	$l=length($t);
-	push(@out,$t);
-	$l=4-($l+9)/8;
-	push(@out,"\t" x $l);
-	push(@out,&conv($p2));
-	push(@out,"\n");
-	}
-
-sub out0
-	{
-	local($name)=@_;
-
-	push(@out,"\t$name\n");
-	}
-
-sub out1
-	{
-	local($name,$p1)=@_;
-	local($l,$t);
-
-	push(@out,"\t$name\t".&conv($p1)."\n");
-	}
-
-sub conv
-	{
-	local($p)=@_;
-
-	$p =~ s/0x([0-9A-Fa-f]+)/0$1h/;
-	return $p;
-	}
-
-sub using486
-	{
-	return if $using486;
-	$using486++;
-	grep(s/\.386/\.486/,@out);
-	}
-
-sub main'file
-	{
-	local($file)=@_;
-
-	local($tmp)=<<"EOF";
-	TITLE	$file.asm
-        .386
-.model	FLAT
-_TEXT\$	SEGMENT PAGE 'CODE'
-
-EOF
-	push(@out,$tmp);
-	}
-
-sub main'function_begin
-	{
-	local($func,$extra)=@_;
-
-	push(@labels,$func);
-
-	local($tmp)=<<"EOF";
-PUBLIC	_$func
-$extra
-_$func PROC NEAR
-	push	ebp
-	push	ebx
-	push	esi
-	push	edi
-EOF
-	push(@out,$tmp);
-	$stack=20;
-	}
-
-sub main'function_begin_B
-	{
-	local($func,$extra)=@_;
-
-	local($tmp)=<<"EOF";
-PUBLIC	_$func
-$extra
-_$func PROC NEAR
-EOF
-	push(@out,$tmp);
-	$stack=4;
-	}
-
-sub main'function_end
-	{
-	local($func)=@_;
-
-	local($tmp)=<<"EOF";
-	pop	edi
-	pop	esi
-	pop	ebx
-	pop	ebp
-	ret
-_$func ENDP
-EOF
-	push(@out,$tmp);
-	$stack=0;
-	%label=();
-	}
-
-sub main'function_end_B
-	{
-	local($func)=@_;
-
-	local($tmp)=<<"EOF";
-_$func ENDP
-EOF
-	push(@out,$tmp);
-	$stack=0;
-	%label=();
-	}
-
-sub main'function_end_A
-	{
-	local($func)=@_;
-
-	local($tmp)=<<"EOF";
-	pop	edi
-	pop	esi
-	pop	ebx
-	pop	ebp
-	ret
-EOF
-	push(@out,$tmp);
-	}
-
-sub main'file_end
-	{
-	# try to detect if SSE2 or MMX extensions were used...
-	if (grep {/xmm[0-7]\s*,/i} @out) {
-		grep {s/\.[3-7]86/\.686\n\t\.XMM/} @out;
-		}
-	elsif (grep {/mm[0-7]\s*,/i} @out) {
-		grep {s/\.[3-7]86/\.686\n\t\.MMX/} @out;
-		}
-	push(@out,"_TEXT\$	ENDS\n");
-	push(@out,"END\n");
-	}
-
-sub main'wparam
-	{
-	local($num)=@_;
-
-	return(&main'DWP($stack+$num*4,"esp","",0));
-	}
-
-sub main'swtmp
-	{
-	return(&main'DWP($_[0]*4,"esp","",0));
-	}
-
-# Should use swtmp, which is above esp.  Linix can trash the stack above esp
-#sub main'wtmp
-#	{
-#	local($num)=@_;
-#
-#	return(&main'DWP(-(($num+1)*4),"esp","",0));
-#	}
-
-sub main'comment
-	{
-	foreach (@_)
-		{
-		push(@out,"\t; $_\n");
-		}
-	}
-
-sub main'public_label
-	{
-	$label{$_[0]}="_$_[0]"	if (!defined($label{$_[0]}));
-	push(@out,"PUBLIC\t$label{$_[0]}\n");
-	}
-
-sub main'label
-	{
-	if (!defined($label{$_[0]}))
-		{
-		$label{$_[0]}="\$${label}${_[0]}";
-		$label++;
-		}
-	return($label{$_[0]});
-	}
-
-sub main'set_label
-	{
-	if (!defined($label{$_[0]}))
-		{
-		$label{$_[0]}="\$${label}${_[0]}";
-		$label++;
-		}
-	if ($_[1]!=0 && $_[1]>1)
-		{
-		main'align($_[1]);
-		}
-	if((defined $_[2]) && ($_[2] == 1))
-		{
-		push(@out,"$label{$_[0]}::\n");
-		}
-	elsif ($label{$_[0]} !~ /^\$/)
-		{
-		push(@out,"$label{$_[0]}\tLABEL PTR\n");
-		}
-	else
-		{
-		push(@out,"$label{$_[0]}:\n");
-		}
-	}
-
-sub main'data_byte
-	{
-	push(@out,"\tDB\t".join(',',@_)."\n");
-	}
-
-sub main'data_word
-	{
-	push(@out,"\tDD\t".join(',',@_)."\n");
-	}
-
-sub main'align
-	{
-	push(@out,"\tALIGN\t$_[0]\n");
-	}
-
-sub out1p
-	{
-	local($name,$p1)=@_;
-	local($l,$t);
-
-	push(@out,"\t$name\t".&conv($p1)."\n");
-	}
-
-sub main'picmeup
-	{
-	local($dst,$sym)=@_;
-	&main'lea($dst,&main'DWP($sym));
-	}
-
-sub main'blindpop { &out1("pop",@_); }
-
-sub main'initseg 
-	{
-	local($f)=@_;
-	local($tmp)=<<___;
-OPTION	DOTNAME
-.CRT\$XCU	SEGMENT DWORD PUBLIC 'DATA'
-EXTRN	_$f:NEAR
-DD	_$f
-.CRT\$XCU	ENDS
-___
-	push(@out,$tmp);
-	}
-
-1;

crypto/perlasm/x86nasm.pl

@@ -1,451 +1,256 @@
-#!/usr/local/bin/perl
+#!/usr/bin/env perl
 
 package x86nasm;
 
-$label="L000";
+*out=\@::out;
-$under=($main'netware)?'':'_';
 
-%lb=(	'eax',	'al',
+$lprfx="\@L";
-	'ebx',	'bl',
+$label="000";
-	'ecx',	'cl',
+$under=($::netware)?'':'_';
-	'edx',	'dl',
+$initseg="";
-	'ax',	'al',
-	'bx',	'bl',
-	'cx',	'cl',
-	'dx',	'dl',
-	);
 
-%hb=(	'eax',	'ah',
+sub ::generic
-	'ebx',	'bh',
+{ my $opcode=shift;
-	'ecx',	'ch',
+  my $tmp;
-	'edx',	'dh',
-	'ax',	'ah',
-	'bx',	'bh',
-	'cx',	'ch',
-	'dx',	'dh',
-	);
 
-sub main'asm_init_output { @out=(); }
+    if (!$::mwerks)
-sub main'asm_get_output { return(@out); }
+    {   if    ($opcode =~ m/^j/o && $#_==0) # optimize jumps
-sub main'get_labels { return(@labels); }
+	{   $_[0] = "NEAR $_[0]";   	}
-
+	elsif ($opcode eq "lea" && $#_==1)# wipe storage qualifier from lea
-sub main'external_label
+	{   $_[1] =~ s/^[^\[]*\[/\[/o;	}
-{
-	push(@labels,@_);
-	foreach (@_) {
-		push(@out,".") if ($main'mwerks);
-		push(@out, "extern\t${under}$_\n");
     }
+    &::emit($opcode,@_);
+  1;
 }
+#
+# opcodes not covered by ::generic above, mostly inconsistent namings...
+#
+sub ::movz	{ &::movzx(@_);		}
+sub ::pushf	{ &::pushfd;		}
+sub ::popf	{ &::popfd;		}
 
-sub main'LB
+sub ::call	{ &::emit("call",(&islabel($_[0]) or "$under$_[0]")); }
-	{
+sub ::call_ptr	{ &::emit("call",@_);	}
-	(defined($lb{$_[0]})) || die "$_[0] does not have a 'low byte'\n";
+sub ::jmp_ptr	{ &::emit("jmp",@_);	}
-	return($lb{$_[0]});
-	}
 
-sub main'HB
+# chosen SSE instructions
-	{
+sub ::movq
-	(defined($hb{$_[0]})) || die "$_[0] does not have a 'high byte'\n";
+{ my($p1,$p2,$optimize)=@_;
-	return($hb{$_[0]});
-	}
 
-sub main'BP
+    if ($optimize && $p1=~/^mm[0-7]$/ && $p2=~/^mm[0-7]$/)
-	{
+    # movq between mmx registers can sink Intel CPUs
-	&get_mem("BYTE",@_);
+    {	&::pshufw($p1,$p2,0xe4);		}
-	}
+    else
-
+    {	&::emit("movq",@_);			}
-sub main'DWP
+}
-	{
+sub ::pshufw { &::emit("pshufw",@_); }
-	&get_mem("DWORD",@_);
-	}
-
-sub main'QWP
-	{
-	&get_mem("",@_);
-	}
-
-sub main'BC
-	{
-	return (($main'mwerks)?"":"BYTE ")."@_";
-	}
-
-sub main'DWC
-	{
-	return (($main'mwerks)?"":"DWORD ")."@_";
-	}
-
-sub main'stack_push
-	{
-	my($num)=@_;
-	$stack+=$num*4;
-	&main'sub("esp",$num*4);
-	}
-
-sub main'stack_pop
-	{
-	my($num)=@_;
-	$stack-=$num*4;
-	&main'add("esp",$num*4);
-	}
 
 sub get_mem
-	{
+{ my($size,$addr,$reg1,$reg2,$idx)=@_;
-	my($size,$addr,$reg1,$reg2,$idx)=@_;
+  my($post,$ret);
-	my($t,$post);
+
-	my($ret)=$size;
+    if ($size ne "")
-	if ($ret ne "")
+    {	$ret .= "$size";
-		{
+	$ret .= " PTR" if ($::mwerks);
-		$ret .= " PTR" if ($main'mwerks);
 	$ret .= " ";
     }
     $ret .= "[";
+
     $addr =~ s/^\s+//;
-	if ($addr =~ /^(.+)\+(.+)$/)
+    # prepend global references with optional underscore
-		{
+    $addr =~ s/^([^\+\-0-9][^\+\-]*)/islabel($1) or "$under$1"/ige;
-		$reg2=&conv($1);
+    # put address arithmetic expression in parenthesis
-		$addr="$under$2";
+    $addr="($addr)" if ($addr =~ /^.+[\-\+].+$/);
-		}
-	elsif ($addr =~ /^[_a-z][_a-z0-9]*$/i)
-		{
-		$addr="$under$addr";
-		}
 
-	if ($addr =~ /^.+\-.+$/) { $addr="($addr)"; }
-
-	$reg1="$regs{$reg1}" if defined($regs{$reg1});
-	$reg2="$regs{$reg2}" if defined($regs{$reg2});
     if (($addr ne "") && ($addr ne 0))
-		{
+    {	if ($addr !~ /^-/)	{ $ret .= "$addr+"; }
-		if ($addr !~ /^-/)
-			{ $ret.="${addr}+"; }
 	else			{ $post=$addr;      }
     }
+
     if ($reg2 ne "")
-		{
+    {	$idx!=0 or $idx=1;
-		$t="";
+	$ret .= "$reg2*$idx";
-		$t="*$idx" if ($idx != 0);
+	$ret .= "+$reg1" if ($reg1 ne "");
-		$reg1="+".$reg1 if ("$reg1$post" ne "");
-		$ret.="$reg2$t$reg1$post]";
     }
     else
-		{
+    {	$ret .= "$reg1";   }
-		$ret.="$reg1$post]"
+
-		}
+    $ret .= "$post]";
     $ret =~ s/\+\]/]/; # in case $addr was the only argument
-	return($ret);
-	}
 
-sub main'mov	{ &out2("mov",@_); }
+  $ret;
-sub main'movb	{ &out2("mov",@_); }
+}
-sub main'and	{ &out2("and",@_); }
+sub ::BP	{ &get_mem("BYTE",@_);  }
-sub main'or	{ &out2("or",@_); }
+sub ::DWP	{ &get_mem("DWORD",@_); }
-sub main'shl	{ &out2("shl",@_); }
+sub ::QWP	{ &get_mem("",@_);      }
-sub main'shr	{ &out2("shr",@_); }
+sub ::BC	{ (($::mwerks)?"":"BYTE ")."@_";  }
-sub main'xor	{ &out2("xor",@_); }
+sub ::DWC	{ (($::mwerks)?"":"DWORD ")."@_"; }
-sub main'xorb	{ &out2("xor",@_); }
-sub main'add	{ &out2("add",@_); }
-sub main'adc	{ &out2("adc",@_); }
-sub main'sub	{ &out2("sub",@_); }
-sub main'sbb	{ &out2("sbb",@_); }
-sub main'rotl	{ &out2("rol",@_); }
-sub main'rotr	{ &out2("ror",@_); }
-sub main'exch	{ &out2("xchg",@_); }
-sub main'cmp	{ &out2("cmp",@_); }
-sub main'lea	{ &out2("lea",@_); }
-sub main'mul	{ &out1("mul",@_); }
-sub main'div	{ &out1("div",@_); }
-sub main'dec	{ &out1("dec",@_); }
-sub main'inc	{ &out1("inc",@_); }
-sub main'jmp	{ &out1("jmp",@_); }
-sub main'jmp_ptr { &out1p("jmp",@_); }
 
-# This is a bit of a kludge: declare all branches as NEAR.
+sub ::file
-$near=($main'mwerks)?'':'NEAR';
+{   if ($::mwerks)	{ push(@out,".section\t.text\n"); }
-sub main'je	{ &out1("je $near",@_); }
+    else
-sub main'jle	{ &out1("jle $near",@_); }
+    { my $tmp=<<___;
-sub main'jz	{ &out1("jz $near",@_); }
-sub main'jge	{ &out1("jge $near",@_); }
-sub main'jl	{ &out1("jl $near",@_); }
-sub main'ja	{ &out1("ja $near",@_); }
-sub main'jae	{ &out1("jae $near",@_); }
-sub main'jb	{ &out1("jb $near",@_); }
-sub main'jbe	{ &out1("jbe $near",@_); }
-sub main'jc	{ &out1("jc $near",@_); }
-sub main'jnc	{ &out1("jnc $near",@_); }
-sub main'jnz	{ &out1("jnz $near",@_); }
-sub main'jne	{ &out1("jne $near",@_); }
-sub main'jno	{ &out1("jno $near",@_); }
-
-sub main'push	{ &out1("push",@_); $stack+=4; }
-sub main'pop	{ &out1("pop",@_); $stack-=4; }
-sub main'pushf	{ &out0("pushfd"); $stack+=4; }
-sub main'popf	{ &out0("popfd"); $stack-=4; }
-sub main'bswap	{ &out1("bswap",@_); &using486(); }
-sub main'not	{ &out1("not",@_); }
-sub main'call	{ &out1("call",($_[0]=~/^\@L/?'':$under).$_[0]); }
-sub main'call_ptr { &out1p("call",@_); }
-sub main'ret	{ &out0("ret"); }
-sub main'nop	{ &out0("nop"); }
-sub main'test	{ &out2("test",@_); }
-sub main'bt	{ &out2("bt",@_); }
-sub main'leave	{ &out0("leave"); }
-sub main'cpuid	{ &out0("cpuid"); }
-sub main'rdtsc	{ &out0("rdtsc"); }
-sub main'halt	{ &out0("hlt"); }
-sub main'movz	{ &out2("movzx",@_); }
-sub main'neg	{ &out1("neg",@_); }
-sub main'cld	{ &out0("cld"); }
-
-# SSE2
-sub main'emms	{ &out0("emms"); }
-sub main'movd	{ &out2("movd",@_); }
-sub main'movq	{ &out2("movq",@_); }
-sub main'movdqu	{ &out2("movdqu",@_); }
-sub main'movdqa	{ &out2("movdqa",@_); }
-sub main'movdq2q{ &out2("movdq2q",@_); }
-sub main'movq2dq{ &out2("movq2dq",@_); }
-sub main'paddq	{ &out2("paddq",@_); }
-sub main'pmuludq{ &out2("pmuludq",@_); }
-sub main'psrlq	{ &out2("psrlq",@_); }
-sub main'psllq	{ &out2("psllq",@_); }
-sub main'pxor	{ &out2("pxor",@_); }
-sub main'por	{ &out2("por",@_); }
-sub main'pand	{ &out2("pand",@_); }
-
-sub out2
-	{
-	my($name,$p1,$p2)=@_;
-	my($l,$t);
-
-	push(@out,"\t$name\t");
-	if (!$main'mwerks and $name eq "lea")
-		{
-		$p1 =~ s/^[^\[]*\[/\[/;
-		$p2 =~ s/^[^\[]*\[/\[/;
-		}
-	$t=&conv($p1).",";
-	$l=length($t);
-	push(@out,$t);
-	$l=4-($l+9)/8;
-	push(@out,"\t" x $l);
-	push(@out,&conv($p2));
-	push(@out,"\n");
-	}
-
-sub out0
-	{
-	my($name)=@_;
-
-	push(@out,"\t$name\n");
-	}
-
-sub out1
-	{
-	my($name,$p1)=@_;
-	my($l,$t);
-	push(@out,"\t$name\t".&conv($p1)."\n");
-	}
-
-sub conv
-	{
-	my($p)=@_;
-	$p =~ s/0x([0-9A-Fa-f]+)/0$1h/;
-	return $p;
-	}
-
-sub using486
-	{
-	return if $using486;
-	$using486++;
-	grep(s/\.386/\.486/,@out);
-	}
-
-sub main'file
-	{
-	if ($main'mwerks)	{ push(@out,".section\t.text\n"); }
-	else	{
-		local $tmp=<<___;
 %ifdef __omf__
-section	code	use32 class=code
+section	code	use32 class=code align=64
 %else
-section	.text
+section	.text	code align=64
 %endif
 ___
 	push(@out,$tmp);
     }
-	}
+}
 
-sub main'function_begin
+sub ::function_begin_B
-	{
+{ my $func=$under.shift;
-	my($func,$extra)=@_;
+  my $tmp=<<___;
-
+global	$func
-	push(@labels,$func);
+align	16
-	my($tmp)=<<"EOF";
+$func:
-global	$under$func
-$under$func:
-	push	ebp
-	push	ebx
-	push	esi
-	push	edi
-EOF
-	push(@out,$tmp);
-	$stack=20;
-	}
-
-sub main'function_begin_B
-	{
-	my($func,$extra)=@_;
-	my($tmp)=<<"EOF";
-global	$under$func
-$under$func:
-EOF
-	push(@out,$tmp);
-	$stack=4;
-	}
-
-sub main'function_end
-	{
-	my($func)=@_;
-
-	my($tmp)=<<"EOF";
-	pop	edi
-	pop	esi
-	pop	ebx
-	pop	ebp
-	ret
-EOF
-	push(@out,$tmp);
-	$stack=0;
-	%label=();
-	}
-
-sub main'function_end_B
-	{
-	$stack=0;
-	%label=();
-	}
-
-sub main'function_end_A
-	{
-	my($func)=@_;
-
-	my($tmp)=<<"EOF";
-	pop	edi
-	pop	esi
-	pop	ebx
-	pop	ebp
-	ret
-EOF
-	push(@out,$tmp);
-	}
-
-sub main'file_end
-	{
-	}
-
-sub main'wparam
-	{
-	my($num)=@_;
-
-	return(&main'DWP($stack+$num*4,"esp","",0));
-	}
-
-sub main'swtmp
-	{
-	return(&main'DWP($_[0]*4,"esp","",0));
-	}
-
-# Should use swtmp, which is above esp.  Linix can trash the stack above esp
-#sub main'wtmp
-#	{
-#	my($num)=@_;
-#
-#	return(&main'DWP(-(($num+1)*4),"esp","",0));
-#	}
-
-sub main'comment
-	{
-	foreach (@_)
-		{
-		push(@out,"\t; $_\n");
-		}
-	}
-
-sub main'public_label
-	{
-	$label{$_[0]}="${under}${_[0]}"	if (!defined($label{$_[0]}));
-	push(@out,"global\t$label{$_[0]}\n");
-	}
-
-sub main'label
-	{
-	if (!defined($label{$_[0]}))
-		{
-		$label{$_[0]}="\@${label}${_[0]}";
-		$label++;
-		}
-	return($label{$_[0]});
-	}
-
-sub main'set_label
-	{
-	if (!defined($label{$_[0]}))
-		{
-		$label{$_[0]}="\@${label}${_[0]}";
-		$label++;
-		}
-	if ($_[1]!=0 && $_[1]>1)
-		{
-		main'align($_[1]);
-		}
-	push(@out,"$label{$_[0]}:\n");
-	}
-
-sub main'data_byte
-	{
-	push(@out,(($main'mwerks)?".byte\t":"DB\t").join(',',@_)."\n");
-	}
-
-sub main'data_word
-	{
-	push(@out,(($main'mwerks)?".long\t":"DD\t").join(',',@_)."\n");
-	}
-
-sub main'align
-	{
-	push(@out,".") if ($main'mwerks);
-	push(@out,"align\t$_[0]\n");
-	}
-
-sub out1p
-	{
-	my($name,$p1)=@_;
-	my($l,$t);
-
-	push(@out,"\t$name\t".&conv($p1)."\n");
-	}
-
-sub main'picmeup
-	{
-	local($dst,$sym)=@_;
-	&main'lea($dst,&main'DWP($sym));
-	}
-
-sub main'blindpop { &out1("pop",@_); }
-
-sub main'initseg
-	{
-	local($f)=@_;
-	if ($main'win32)
-		{
-		local($tmp)=<<___;
-segment	.CRT\$XCU data
-extern	$under$f
-DD	$under$f
 ___
     push(@out,$tmp);
+    $::stack=4;
+}
+sub ::function_end_B
+{ my $i;
+    foreach $i (%label) { undef $label{$i} if ($label{$i} =~ /^$prfx/);  }
+    $::stack=0;
+}
+
+sub ::file_end
+{   # try to detect if SSE2 or MMX extensions were used on Win32...
+    if ($::win32 && grep {/\b[x]?mm[0-7]\b|OPENSSL_ia32cap_P\b/i} @out)
+    {	# $1<<10 sets a reserved bit to signal that variable
+	# was initialized already...
+	my $code=<<___;
+align	16
+${lprfx}OPENSSL_ia32cap_init:
+	lea	edx,[${under}OPENSSL_ia32cap_P]
+	cmp	DWORD [edx],0
+	jne	NEAR ${lprfx}nocpuid
+	mov	DWORD [edx],1<<10
+	pushfd
+	pop	eax
+	mov	ecx,eax
+	xor	eax,1<<21
+	push	eax
+	popfd
+	pushfd
+	pop	eax
+	xor	eax,ecx
+	bt	eax,21
+	jnc	NEAR ${lprfx}nocpuid
+	push	ebp
+	push	edi
+	push	ebx
+	mov	edi,edx
+	xor	eax,eax
+	cpuid
+	xor	eax,eax
+	cmp	ebx,'Genu'
+	setne	al
+	mov	ebp,eax
+	cmp	edx,'ineI'
+	setne	al
+	or	ebp,eax
+	cmp	eax,'ntel'
+	setne	al
+	or	ebp,eax
+	mov	eax,1
+	cpuid
+	cmp	ebp,0
+	jne	${lprfx}notP4
+	and	ah,15
+	cmp	ah,15
+	jne	${lprfx}notP4
+	or	edx,1<<20
+${lprfx}notP4:
+	bt	edx,28
+	jnc	${lprfx}done
+	shr	ebx,16
+	cmp	bl,1
+	ja	${lprfx}done
+	and	edx,0xefffffff
+${lprfx}done:
+	or	edx,1<<10
+	mov	DWORD [edi],edx
+	pop	ebx
+	pop	edi
+	pop	ebp
+${lprfx}nocpuid:
+	ret
+segment	.CRT\$XCU data align=4
+dd	${lprfx}OPENSSL_ia32cap_init
+___
+	my $data=<<___;
+segment	.bss
+common	${under}OPENSSL_ia32cap_P 4
+___
+
+	#<not needed in OpenSSL context>#push (@out,$code);
+
+	# comment out OPENSSL_ia32cap_P declarations
+	grep {s/(^extern\s+${under}OPENSSL_ia32cap_P)/\;$1/} @out;
+	push (@out,$data)
     }
+    push (@out,$initseg) if ($initseg);		
+}
+
+sub ::comment {   foreach (@_) { push(@out,"\t; $_\n"); }   }
+
+sub islabel	# see is argument is known label
+{ my $i;
+    foreach $i (%label) { return $label{$i} if ($label{$i} eq $_[0]); }
+  undef;
+}
+
+sub ::external_label
+{   push(@labels,@_);
+    foreach (@_)
+    {	push(@out,".") if ($::mwerks);
+	push(@out, "extern\t${under}$_\n");
     }
+}
+
+sub ::public_label
+{   $label{$_[0]}="${under}${_[0]}" if (!defined($label{$_[0]}));
+    push(@out,"global\t$label{$_[0]}\n");
+}
+
+sub ::label
+{   if (!defined($label{$_[0]}))
+    {	$label{$_[0]}="${lprfx}${label}${_[0]}"; $label++;   }
+  $label{$_[0]};
+}
+
+sub ::set_label
+{ my $label=&::label($_[0]);
+    &::align($_[1]) if ($_[1]>1);
+    push(@out,"$label{$_[0]}:\n");
+}
+
+sub ::data_byte
+{   push(@out,(($::mwerks)?".byte\t":"db\t").join(',',@_)."\n");	}
+
+sub ::data_word
+{   push(@out,(($::mwerks)?".long\t":"dd\t").join(',',@_)."\n");	}
+
+sub ::align
+{   push(@out,".") if ($::mwerks); push(@out,"align\t$_[0]\n");	}
+
+sub ::picmeup
+{ my($dst,$sym)=@_;
+    &::lea($dst,&::DWP($sym));
+}
+
+sub ::initseg
+{ my($f)=$under.shift;
+    if ($::win32)
+    {	$initseg=<<___;
+segment	.CRT\$XCU data align=4
+extern	$f
+dd	$f
+___
+    }
+}
 
 1;

crypto/perlasm/x86unix.pl

@@ -1,499 +1,152 @@
-#!/usr/local/bin/perl
+#!/usr/bin/env perl
 
 package x86unix;	# GAS actually...
 
+*out=\@::out;
+
 $label="L000";
-$const="";
-$constl=0;
 
-$align=($main'aout)?"4":"16";
+$align=($::aout)?"4":"16";
-$under=($main'aout or $main'coff)?"_":"";
+$under=($::aout or $::coff)?"_":"";
-$dot=($main'aout)?"":".";
+$dot=($::aout)?"":".";
-$com_start="#" if ($main'aout or $main'coff);
+$com_start="#" if ($::aout or $::coff);
 
-sub main'asm_init_output { @out=(); }
+sub opsize()
-sub main'asm_get_output { return(@out); }
+{ my $reg=shift;
-sub main'get_labels { return(@labels); }
+    if    ($reg =~ m/^%e/o)		{ "l"; }
-sub main'external_label { push(@labels,@_); }
+    elsif ($reg =~ m/^%[a-d][hl]$/o)	{ "b"; }
+    elsif ($reg =~ m/^%[xm]/o)		{ undef; }
+    else				{ "w"; }
+}
 
-if ($main'cpp)
+# swap arguments;
-	{
+# expand opcode with size suffix;
-	$align="ALIGN";
+# prefix numeric constants with $;
-	$under="";
+sub ::generic
-	$com_start='/*';
+{ my($opcode,$dst,$src)=@_;
-	$com_end='*/';
+  my($tmp,$suffix,@arg);
+
+    if (defined($src))
+    {	$src =~ s/^(e?[a-dsixphl]{2})$/%$1/o;
+	$src =~ s/^(x?mm[0-7])$/%$1/o;
+	$src =~ s/^(\-?[0-9]+)$/\$$1/o;
+	$src =~ s/^(\-?0x[0-9a-f]+)$/\$$1/o;
+	push(@arg,$src);
+    }
+    if (defined($dst))
+    {	$dst =~ s/^(\*?)(e?[a-dsixphl]{2})$/$1%$2/o;
+	$dst =~ s/^(x?mm[0-7])$/%$1/o;
+	$dst =~ s/^(\-?[0-9]+)$/\$$1/o		if(!defined($src));
+	$dst =~ s/^(\-?0x[0-9a-f]+)$/\$$1/o	if(!defined($src));
+	push(@arg,$dst);
     }
 
-%lb=(	'eax',	'%al',
+    if    ($dst =~ m/^%/o)	{ $suffix=&opsize($dst); }
-	'ebx',	'%bl',
+    elsif ($src =~ m/^%/o)	{ $suffix=&opsize($src); }
-	'ecx',	'%cl',
+    else			{ $suffix="l";           }
-	'edx',	'%dl',
+    undef $suffix if ($dst =~ m/^%[xm]/o || $src =~ m/^%[xm]/o);
-	'ax',	'%al',
-	'bx',	'%bl',
-	'cx',	'%cl',
-	'dx',	'%dl',
-	);
 
-%hb=(	'eax',	'%ah',
+    if ($#_==0)				{ &::emit($opcode);		}
-	'ebx',	'%bh',
+    elsif ($opcode =~ m/^j/o && $#_==1)	{ &::emit($opcode,@arg);	}
-	'ecx',	'%ch',
+    elsif ($opcode eq "call" && $#_==1)	{ &::emit($opcode,@arg);	}
-	'edx',	'%dh',
+    elsif ($opcode =~ m/^set/&& $#_==1)	{ &::emit($opcode,@arg);	}
-	'ax',	'%ah',
+    else				{ &::emit($opcode.$suffix,@arg);}
-	'bx',	'%bh',
-	'cx',	'%ch',
-	'dx',	'%dh',
-	);
 
-%regs=(	'eax',	'%eax',
+  1;
-	'ebx',	'%ebx',
+}
-	'ecx',	'%ecx',
-	'edx',	'%edx',
-	'esi',	'%esi',
-	'edi',	'%edi',
-	'ebp',	'%ebp',
-	'esp',	'%esp',
-
-	'mm0',	'%mm0',
-	'mm1',	'%mm1',
-	'mm2',	'%mm2',
-	'mm3',	'%mm3',
-	'mm4',	'%mm4',
-	'mm5',	'%mm5',
-	'mm6',	'%mm6',
-	'mm7',	'%mm7',
-
-	'xmm0',	'%xmm0',
-	'xmm1',	'%xmm1',
-	'xmm2',	'%xmm2',
-	'xmm3',	'%xmm3',
-	'xmm4',	'%xmm4',
-	'xmm5',	'%xmm5',
-	'xmm6',	'%xmm6',
-	'xmm7',	'%xmm7',
-	);
-
-%reg_val=(
-	'eax',	0x00,
-	'ebx',	0x03,
-	'ecx',	0x01,
-	'edx',	0x02,
-	'esi',	0x06,
-	'edi',	0x07,
-	'ebp',	0x05,
-	'esp',	0x04,
-	);
-
-sub main'LB
-	{
-	(defined($lb{$_[0]})) || die "$_[0] does not have a 'low byte'\n";
-	return($lb{$_[0]});
-	}
-
-sub main'HB
-	{
-	(defined($hb{$_[0]})) || die "$_[0] does not have a 'high byte'\n";
-	return($hb{$_[0]});
-	}
-
-sub main'DWP
-	{
-	local($addr,$reg1,$reg2,$idx)=@_;
-
-	$ret="";
-	$addr =~ s/(^|[+ \t])([A-Za-z_]+[A-Za-z0-9_]+)($|[+ \t])/$1$under$2$3/;
-	$reg1="$regs{$reg1}" if defined($regs{$reg1});
-	$reg2="$regs{$reg2}" if defined($regs{$reg2});
-	$ret.=$addr if ($addr ne "") && ($addr ne 0);
-	if ($reg2 ne "")
-		{
-		if($idx ne "" && $idx != 0)
-		    { $ret.="($reg1,$reg2,$idx)"; }
-		else
-		    { $ret.="($reg1,$reg2)"; }
-	        }
-	elsif ($reg1 ne "")
-		{ $ret.="($reg1)" }
-	return($ret);
-	}
-
-sub main'QWP
-	{
-	return(&main'DWP(@_));
-	}
-
-sub main'BP
-	{
-	return(&main'DWP(@_));
-	}
-
-sub main'BC
-	{
-	return @_;
-	}
-
-sub main'DWC
-	{
-	return @_;
-	}
-
-#sub main'BP
-#	{
-#	local($addr,$reg1,$reg2,$idx)=@_;
 #
-#	$ret="";
+# opcodes not covered by ::generic above, mostly inconsistent namings...
 #
-#	$addr =~ s/(^|[+ \t])([A-Za-z_]+)($|[+ \t])/$1$under$2$3/;
+sub ::movz	{ &::movzb(@_);			}
-#	$reg1="$regs{$reg1}" if defined($regs{$reg1});
+sub ::pushf	{ &::pushfl;			}
-#	$reg2="$regs{$reg2}" if defined($regs{$reg2});
+sub ::popf	{ &::popfl;			}
-#	$ret.=$addr if ($addr ne "") && ($addr ne 0);
+sub ::cpuid	{ &::emit(".byte\t0x0f,0xa2");	}
-#	if ($reg2 ne "")
+sub ::rdtsc	{ &::emit(".byte\t0x0f,0x31");	}
-#		{ $ret.="($reg1,$reg2,$idx)"; }
-#	else
-#		{ $ret.="($reg1)" }
-#	return($ret);
-#	}
 
-sub main'mov	{ &out2("movl",@_); }
+sub ::call	{ &::emit("call",(&islabel($_[0]) or "$under$_[0]")); }
-sub main'movb	{ &out2("movb",@_); }
+sub ::call_ptr	{ &::generic("call","*$_[0]");	}
-sub main'and	{ &out2("andl",@_); }
+sub ::jmp_ptr	{ &::generic("jmp","*$_[0]");	}
-sub main'or	{ &out2("orl",@_); }
-sub main'shl	{ &out2("sall",@_); }
-sub main'shr	{ &out2("shrl",@_); }
-sub main'xor	{ &out2("xorl",@_); }
-sub main'xorb	{ &out2("xorb",@_); }
-sub main'add	{ &out2($_[0]=~/%[a-d][lh]/?"addb":"addl",@_); }
-sub main'adc	{ &out2("adcl",@_); }
-sub main'sub	{ &out2("subl",@_); }
-sub main'sbb	{ &out2("sbbl",@_); }
-sub main'rotl	{ &out2("roll",@_); }
-sub main'rotr	{ &out2("rorl",@_); }
-sub main'exch	{ &out2($_[0]=~/%[a-d][lh]/?"xchgb":"xchgl",@_); }
-sub main'cmp	{ &out2("cmpl",@_); }
-sub main'lea	{ &out2("leal",@_); }
-sub main'mul	{ &out1("mull",@_); }
-sub main'div	{ &out1("divl",@_); }
-sub main'jmp	{ &out1("jmp",@_); }
-sub main'jmp_ptr { &out1p("jmp",@_); }
-sub main'je	{ &out1("je",@_); }
-sub main'jle	{ &out1("jle",@_); }
-sub main'jne	{ &out1("jne",@_); }
-sub main'jnz	{ &out1("jnz",@_); }
-sub main'jz	{ &out1("jz",@_); }
-sub main'jge	{ &out1("jge",@_); }
-sub main'jl	{ &out1("jl",@_); }
-sub main'ja	{ &out1("ja",@_); }
-sub main'jae	{ &out1("jae",@_); }
-sub main'jb	{ &out1("jb",@_); }
-sub main'jbe	{ &out1("jbe",@_); }
-sub main'jc	{ &out1("jc",@_); }
-sub main'jnc	{ &out1("jnc",@_); }
-sub main'jno	{ &out1("jno",@_); }
-sub main'dec	{ &out1("decl",@_); }
-sub main'inc	{ &out1($_[0]=~/%[a-d][hl]/?"incb":"incl",@_); }
-sub main'push	{ &out1("pushl",@_); $stack+=4; }
-sub main'pop	{ &out1("popl",@_); $stack-=4; }
-sub main'pushf	{ &out0("pushfl"); $stack+=4; }
-sub main'popf	{ &out0("popfl"); $stack-=4; }
-sub main'not	{ &out1("notl",@_); }
-sub main'call	{	my $pre=$under;
-			foreach $i (%label)
-			{ if ($label{$i} eq $_[0]) { $pre=''; last; } }
-			&out1("call",$pre.$_[0]);
-		}
-sub main'call_ptr { &out1p("call",@_); }
-sub main'ret	{ &out0("ret"); }
-sub main'nop	{ &out0("nop"); }
-sub main'test	{ &out2("testl",@_); }
-sub main'bt	{ &out2("btl",@_); }
-sub main'leave	{ &out0("leave"); }
-sub main'cpuid	{ &out0(".byte\t0x0f,0xa2"); }
-sub main'rdtsc	{ &out0(".byte\t0x0f,0x31"); }
-sub main'halt	{ &out0("hlt"); }
-sub main'movz	{ &out2("movzbl",@_); }
-sub main'neg	{ &out1("negl",@_); }
-sub main'cld	{ &out0("cld"); }
 
-# SSE2
+*::bswap = sub	{ &::emit("bswap","%$_[0]");	} if (!$::i386);
-sub main'emms	{ &out0("emms"); }
+
-sub main'movd	{ &out2("movd",@_); }
+# chosen SSE instructions
-sub main'movdqu	{ &out2("movdqu",@_); }
+sub ::movq
-sub main'movdqa	{ &out2("movdqa",@_); }
+{ my($p1,$p2,$optimize)=@_;
-sub main'movdq2q{ &out2("movdq2q",@_); }
-sub main'movq2dq{ &out2("movq2dq",@_); }
-sub main'paddq	{ &out2("paddq",@_); }
-sub main'pmuludq{ &out2("pmuludq",@_); }
-sub main'psrlq	{ &out2("psrlq",@_); }
-sub main'psllq	{ &out2("psllq",@_); }
-sub main'pxor	{ &out2("pxor",@_); }
-sub main'por	{ &out2("por",@_); }
-sub main'pand	{ &out2("pand",@_); }
-sub main'movq	{
-	local($p1,$p2,$optimize)=@_;
     if ($optimize && $p1=~/^mm[0-7]$/ && $p2=~/^mm[0-7]$/)
     # movq between mmx registers can sink Intel CPUs
-		{	push(@out,"\tpshufw\t\$0xe4,%$p2,%$p1\n");	}
+    {	&::pshufw($p1,$p2,0xe4);	}
-	else	{	&out2("movq",@_);				}
-	}
-
-# The bswapl instruction is new for the 486. Emulate if i386.
-sub main'bswap
-	{
-	if ($main'i386)
-		{
-		&main'comment("bswapl @_");
-		&main'exch(main'HB(@_),main'LB(@_));
-		&main'rotr(@_,16);
-		&main'exch(main'HB(@_),main'LB(@_));
-		}
     else
-		{
+    {	&::generic("movq",@_);	}
-		&out1("bswapl",@_);
+}
-		}
+sub ::pshufw
+{ my($dst,$src,$magic)=@_;
+    &::emit("pshufw","\$$magic","%$src","%$dst");
+}
+
+sub ::DWP
+{ my($addr,$reg1,$reg2,$idx)=@_;
+  my $ret="";
+
+    $addr =~ s/^\s+//;
+    # prepend global references with optional underscore
+    $addr =~ s/^([^\+\-0-9][^\+\-]*)/islabel($1) or "$under$1"/ige;
+
+    $reg1 = "%$reg1" if ($reg1);
+    $reg2 = "%$reg2" if ($reg2);
+
+    $ret .= $addr if (($addr ne "") && ($addr ne 0));
+
+    if ($reg2)
+    {	$idx!= 0 or $idx=1;
+	$ret .= "($reg1,$reg2,$idx)";
     }
+    elsif ($reg1)
+    {	$ret .= "($reg1)";	}
 
-sub out2
+  $ret;
-	{
+}
-	local($name,$p1,$p2)=@_;
+sub ::QWP	{ &::DWP(@_);	}
-	local($l,$ll,$t);
+sub ::BP	{ &::DWP(@_);	}
-	local(%special)=(	"roll",0xD1C0,"rorl",0xD1C8,
+sub ::BC	{ @_;		}
-				"rcll",0xD1D0,"rcrl",0xD1D8,
+sub ::DWC	{ @_;		}
-				"shll",0xD1E0,"shrl",0xD1E8,
-				"sarl",0xD1F8);
 
-	if ((defined($special{$name})) && defined($regs{$p1}) && ($p2 == 1))
+sub ::file
-		{
+{   push(@out,".file\t\"$_[0].s\"\n");	}
-		$op=$special{$name}|$reg_val{$p1};
-		$tmp1=sprintf(".byte %d\n",($op>>8)&0xff);
-		$tmp2=sprintf(".byte %d\t",$op     &0xff);
-		push(@out,$tmp1);
-		push(@out,$tmp2);
 
-		$p2=&conv($p2);
+sub ::function_begin_B
-		$p1=&conv($p1);
+{ my($func,$extra)=@_;
-		&main'comment("$name $p2 $p1");
+  my $tmp;
-		return;
-		}
 
-	push(@out,"\t$name\t");
+    &::external_label($func);
-	$t=&conv($p2).",";
-	$l=length($t);
-	push(@out,$t);
-	$ll=4-($l+9)/8;
-	$tmp1=sprintf("\t" x $ll);
-	push(@out,$tmp1);
-	push(@out,&conv($p1)."\n");
-	}
-
-sub out1
-	{
-	local($name,$p1)=@_;
-	local($l,$t);
-	local(%special)=("bswapl",0x0FC8);
-
-	if ((defined($special{$name})) && defined($regs{$p1}))
-		{
-		$op=$special{$name}|$reg_val{$p1};
-		$tmp1=sprintf(".byte %d\n",($op>>8)&0xff);
-		$tmp2=sprintf(".byte %d\t",$op     &0xff);
-		push(@out,$tmp1);
-		push(@out,$tmp2);
-
-		$p2=&conv($p2);
-		$p1=&conv($p1);
-		&main'comment("$name $p2 $p1");
-		return;
-		}
-
-	push(@out,"\t$name\t".&conv($p1)."\n");
-	}
-
-sub out1p
-	{
-	local($name,$p1)=@_;
-	local($l,$t);
-
-	push(@out,"\t$name\t*".&conv($p1)."\n");
-	}
-
-sub out0
-	{
-	push(@out,"\t$_[0]\n");
-	}
-
-sub conv
-	{
-	local($p)=@_;
-
-#	$p =~ s/0x([0-9A-Fa-f]+)/0$1h/;
-
-	$p=$regs{$p} if (defined($regs{$p}));
-
-	$p =~ s/^(-{0,1}[0-9A-Fa-f]+)$/\$$1/;
-	$p =~ s/^(0x[0-9A-Fa-f]+)$/\$$1/;
-	return $p;
-	}
-
-sub main'file
-	{
-	local($file)=@_;
-
-	local($tmp)=<<"EOF";
-	.file	"$file.s"
-EOF
-	push(@out,$tmp);
-	}
-
-sub main'function_begin
-	{
-	local($func)=@_;
-
-	&main'external_label($func);
     $func=$under.$func;
 
-	local($tmp)=<<"EOF";
+    push(@out,".text\n.globl\t$func\n");
-.text
+    if ($::coff)
-.globl	$func
+    {	push(@out,".def\t$func;\t.scl\t2;\t.type\t32;\t.endef\n"); }
-EOF
+    elsif ($::aout and !$::pic)
-	push(@out,$tmp);
-	if ($main'cpp)
-		{ $tmp=push(@out,"TYPE($func,\@function)\n"); }
-	elsif ($main'coff)
-		{ $tmp=push(@out,".def\t$func;\t.scl\t2;\t.type\t32;\t.endef\n"); }
-	elsif ($main'aout and !$main'pic)
     { }
-	else	{ $tmp=push(@out,".type\t$func,\@function\n"); }
+    else
+    {	push(@out,".type	$func,\@function\n"); }
     push(@out,".align\t$align\n");
     push(@out,"$func:\n");
-	$tmp=<<"EOF";
+    $::stack=4;
-	pushl	%ebp
+}
-	pushl	%ebx
-	pushl	%esi
-	pushl	%edi
 
-EOF
+sub ::function_end_B
-	push(@out,$tmp);
+{ my($func)=@_;
-	$stack=20;
-	}
-
-sub main'function_begin_B
-	{
-	local($func,$extra)=@_;
-
-	&main'external_label($func);
-	$func=$under.$func;
-
-	local($tmp)=<<"EOF";
-.text
-.globl	$func
-EOF
-	push(@out,$tmp);
-	if ($main'cpp)
-		{ push(@out,"TYPE($func,\@function)\n"); }
-	elsif ($main'coff)
-		{ $tmp=push(@out,".def\t$func;\t.scl\t2;\t.type\t32;\t.endef\n"); }
-	elsif ($main'aout and !$main'pic)
-		{ }
-	else	{ push(@out,".type	$func,\@function\n"); }
-	push(@out,".align\t$align\n");
-	push(@out,"$func:\n");
-	$stack=4;
-	}
-
-sub main'function_end
-	{
-	local($func)=@_;
 
     $func=$under.$func;
-
-	local($tmp)=<<"EOF";
-	popl	%edi
-	popl	%esi
-	popl	%ebx
-	popl	%ebp
-	ret
-${dot}L_${func}_end:
-EOF
-	push(@out,$tmp);
-
-	if ($main'cpp)
-		{ push(@out,"SIZE($func,${dot}L_${func}_end-$func)\n"); }
-	elsif ($main'coff or $main'aout)
-                { }
-	else	{ push(@out,".size\t$func,${dot}L_${func}_end-$func\n"); }
-	push(@out,".ident	\"$func\"\n");
-	$stack=0;
-	%label=();
-	}
-
-sub main'function_end_A
-	{
-	local($func)=@_;
-
-	local($tmp)=<<"EOF";
-	popl	%edi
-	popl	%esi
-	popl	%ebx
-	popl	%ebp
-	ret
-EOF
-	push(@out,$tmp);
-	}
-
-sub main'function_end_B
-	{
-	local($func)=@_;
-
-	$func=$under.$func;
-
     push(@out,"${dot}L_${func}_end:\n");
-	if ($main'cpp)
+    if ($::elf)
-		{ push(@out,"SIZE($func,${dot}L_${func}_end-$func)\n"); }
+    {	push(@out,".size\t$func,${dot}L_${func}_end-$func\n"); }
-        elsif ($main'coff or $main'aout)
+    $::stack=0;
-                { }
-	else	{ push(@out,".size\t$func,${dot}L_${func}_end-$func\n"); }
-	push(@out,".ident	\"$func\"\n");
-	$stack=0;
     %label=();
-	}
+}
 
-sub main'wparam
+sub ::comment
 	{
-	local($num)=@_;
+	if (!defined($com_start) or $::elf)
-
+		{	# Regarding $::elf above...
-	return(&main'DWP($stack+$num*4,"esp","",0));
-	}
-
-sub main'stack_push
-	{
-	local($num)=@_;
-	$stack+=$num*4;
-	&main'sub("esp",$num*4);
-	}
-
-sub main'stack_pop
-	{
-	local($num)=@_;
-	$stack-=$num*4;
-	&main'add("esp",$num*4);
-	}
-
-sub main'swtmp
-	{
-	return(&main'DWP($_[0]*4,"esp","",0));
-	}
-
-# Should use swtmp, which is above esp.  Linix can trash the stack above esp
-#sub main'wtmp
-#	{
-#	local($num)=@_;
-#
-#	return(&main'DWP(-($num+1)*4,"esp","",0));
-#	}
-
-sub main'comment
-	{
-	if (!defined($com_start) or $main'elf)
-		{	# Regarding $main'elf above...
 			# GNU and SVR4 as'es use different comment delimiters,
 		push(@out,"\n");	# so we just skip ELF comments...
 		return;
@@ -507,58 +160,47 @@ sub main'comment
 		}
 	}
 
-sub main'public_label
+sub islabel	# see is argument is a known label
-	{
+{ my $i;
-	$label{$_[0]}="${under}${_[0]}"	if (!defined($label{$_[0]}));
+    foreach $i (%label) { return $label{$i} if ($label{$i} eq $_[0]); }
+  undef;
+}
+
+sub ::external_label { push(@labels,@_); }
+
+sub ::public_label
+{   $label{$_[0]}="${under}${_[0]}"	if (!defined($label{$_[0]}));
     push(@out,".globl\t$label{$_[0]}\n");
-	}
+}
 
-sub main'label
+sub ::label
-	{
+{   if (!defined($label{$_[0]}))
-	if (!defined($label{$_[0]}))
+    {	$label{$_[0]}="${dot}${label}${_[0]}"; $label++;   }
-		{
+  $label{$_[0]};
-		$label{$_[0]}="${dot}${label}${_[0]}";
+}
-		$label++;
-		}
-	return($label{$_[0]});
-	}
 
-sub main'set_label
+sub ::set_label
-	{
+{ my $label=&::label($_[0]);
-	if (!defined($label{$_[0]}))
+    &::align($_[1]) if ($_[1]>1);
-		{
+    push(@out,"$label:\n");
-		$label{$_[0]}="${dot}${label}${_[0]}";
+}
-		$label++;
-		}
-	if ($_[1]!=0)
-		{
-		if ($_[1]>1)	{ main'align($_[1]);		}
-		else		{ push(@out,".align $align\n");	}
-		}
-	push(@out,"$label{$_[0]}:\n");
-	}
 
-sub main'file_end
+sub ::file_end
-	{
+{   # try to detect if SSE2 or MMX extensions were used on ELF platform...
-	# try to detect if SSE2 or MMX extensions were used on ELF platform...
+    if ($::elf && grep {/\b%[x]?mm[0-7]\b|OPENSSL_ia32cap_P\b/i} @out) {
-	if ($main'elf && grep {/%[x]*mm[0-7]/i} @out) {
-		local($tmp);
 
 	push (@out,"\n.section\t.bss\n");
 	push (@out,".comm\t${under}OPENSSL_ia32cap_P,4,4\n");
 
+	return;	# below is not needed in OpenSSL context
+
 	push (@out,".section\t.init\n");
-		# One can argue that it's wasteful to craft every
+	&::picmeup("edx","OPENSSL_ia32cap_P");
-		# SSE/MMX module with this snippet... Well, it's 72
-		# bytes long and for the moment we have two modules.
-		# Let's argue when we have 7 modules or so...
-		#
 	# $1<<10 sets a reserved bit to signal that variable
 	# was initialized already...
-		&main'picmeup("edx","OPENSSL_ia32cap_P");
+	my $code=<<___;
-		$tmp=<<___;
 	cmpl	\$0,(%edx)
-		jne	1f
+	jne	3f
 	movl	\$1<<10,(%edx)
 	pushf
 	popl	%eax
@@ -570,165 +212,90 @@ sub main'file_end
 	popl	%eax
 	xorl	%ecx,%eax
 	btl	\$21,%eax
-		jnc	1f
+	jnc	3f
+	pushl	%ebp
 	pushl	%edi
 	pushl	%ebx
 	movl	%edx,%edi
-		movl	\$1,%eax
+	xor	%eax,%eax
 	.byte	0x0f,0xa2
-		orl	\$1<<10,%edx
+	xorl	%eax,%eax
+	cmpl	$1970169159,%ebx
+	setne	%al
+	movl	%eax,%ebp
+	cmpl	$1231384169,%edx
+	setne	%al
+	orl	%eax,%ebp
+	cmpl	$1818588270,%ecx
+	setne	%al
+	orl	%eax,%ebp
+	movl	$1,%eax
+	.byte	0x0f,0xa2
+	cmpl	$0,%ebp
+	jne	1f
+	andb	$15,%ah
+	cmpb	$15,%ah
+	jne	1f
+	orl	$1048576,%edx
+1:	btl	$28,%edx
+	jnc	2f
+	shrl	$16,%ebx
+	cmpb	$1,%bl
+	ja	2f
+	andl	$4026531839,%edx
+2:	orl	\$1<<10,%edx
 	movl	%edx,0(%edi)
 	popl	%ebx
 	popl	%edi
-		jmp	1f
+	popl	%ebp
+	jmp	3f
 	.align	$align
-	1:
+	3:
 ___
-		push (@out,$tmp);
+	push (@out,$code);
     }
+}
 
-	if ($const ne "")
+sub ::data_byte	{   push(@out,".byte\t".join(',',@_)."\n");   }
-		{
+sub ::data_word {   push(@out,".long\t".join(',',@_)."\n");   }
-		push(@out,".section .rodata\n");
-		push(@out,$const);
-		$const="";
-		}
-	}
 
-sub main'data_byte
+sub ::align
-	{
+{ my $val=$_[0],$p2,$i;
-	push(@out,"\t.byte\t".join(',',@_)."\n");
+    if ($::aout)
-	}
+    {	for ($p2=0;$val!=0;$val>>=1) { $p2++; }
-
-sub main'data_word
-	{
-	push(@out,"\t.long\t".join(',',@_)."\n");
-	}
-
-sub main'align
-	{
-	my $val=$_[0],$p2,$i;
-	if ($main'aout) {
-		for ($p2=0;$val!=0;$val>>=1) { $p2++; }
 	$val=$p2-1;
 	$val.=",0x90";
     }
     push(@out,".align\t$val\n");
-	}
+}
 
-# debug output functions: puts, putx, printf
+sub ::picmeup
+{ my($dst,$sym,$base,$reflabel)=@_;
 
-sub main'puts
+    if ($::pic && ($::elf || $::aout))
-	{
+    {	if (!defined($base))
-	&pushvars();
+	{   &::call(&::label("PIC_me_up"));
-	&main'push('$Lstring' . ++$constl);
+	    &::set_label("PIC_me_up");
-	&main'call('puts');
+	    &::blindpop($dst);
-	$stack-=4;
+	    &::add($dst,"\$${under}_GLOBAL_OFFSET_TABLE_+[.-".
-	&main'add("esp",4);
+			    &::label("PIC_me_up") . "]");
-	&popvars();
-
-	$const .= "Lstring$constl:\n\t.string \"@_[0]\"\n";
-	}
-
-sub main'putx
-	{
-	&pushvars();
-	&main'push($_[0]);
-	&main'push('$Lstring' . ++$constl);
-	&main'call('printf');
-	&main'add("esp",8);
-	$stack-=8;
-	&popvars();
-
-	$const .= "Lstring$constl:\n\t.string \"\%X\"\n";
-	}
-
-sub main'printf
-	{
-	$ostack = $stack;
-	&pushvars();
-	for ($i = @_ - 1; $i >= 0; $i--)
-		{
-		if ($i == 0) # change this to support %s format strings
-			{
-			&main'push('$Lstring' . ++$constl);
-			$const .= "Lstring$constl:\n\t.string \"@_[$i]\"\n";
 	}
 	else
-			{
+	{   &::lea($dst,&::DWP("${under}_GLOBAL_OFFSET_TABLE_+[.-$reflabel]",
-			if ($_[$i] =~ /([0-9]*)\(%esp\)/)
+			    $base));
-				{
+	}
-				&main'push(($1 + $stack - $ostack) . '(%esp)');
+	&::mov($dst,&::DWP($under.$sym."\@GOT",$dst));
     }
     else
-				{
+    {	&::lea($dst,&::DWP($sym));	}
-				&main'push($_[$i]);
+}
-				}
-			}
-		}
-	&main'call('printf');
-	$stack-=4*@_;
-	&main'add("esp",4*@_);
-	&popvars();
-	}
 
-sub pushvars
+sub ::initseg
-	{
+{ my($f)=@_;
-	&main'pushf();
+  my($tmp,$ctor);
-	&main'push("edx");
-	&main'push("ecx");
-	&main'push("eax");
-	}
 
-sub popvars
+    if ($::elf)
-	{
+    {	$tmp=<<___;
-	&main'pop("eax");
-	&main'pop("ecx");
-	&main'pop("edx");
-	&main'popf();
-	}
-
-sub main'picmeup
-	{
-	local($dst,$sym)=@_;
-	if ($main'cpp)
-		{
-		local($tmp)=<<___;
-#if (defined(ELF) || defined(SOL)) && defined(PIC)
-	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);
-		}
-	elsif ($main'pic && ($main'elf || $main'aout))
-		{
-		&main'call(&main'label("PIC_me_up"));
-		&main'set_label("PIC_me_up");
-		&main'blindpop($dst);
-		&main'add($dst,"\$${under}_GLOBAL_OFFSET_TABLE_+[.-".
-				&main'label("PIC_me_up") . "]");
-		&main'mov($dst,&main'DWP($under.$sym."\@GOT",$dst));
-		}
-	else
-		{
-		&main'lea($dst,&main'DWP($sym));
-		}
-	}
-
-sub main'blindpop { &out1("popl",@_); }
-
-sub main'initseg
-	{
-	local($f)=@_;
-	local($tmp);
-	if ($main'elf)
-		{
-		$tmp=<<___;
 .section	.init
 	call	$under$f
 	jmp	.Linitalign
@@ -736,18 +303,16 @@ sub main'initseg
 .Linitalign:
 ___
     }
-	elsif ($main'coff)
+    elsif ($::coff)
-		{
+    {   $tmp=<<___;	# applies to both Cygwin and Mingw
-		$tmp=<<___;	# applies to both Cygwin and Mingw
 .section	.ctors
 .long	$under$f
 ___
     }
-	elsif ($main'aout)
+    elsif ($::aout)
-		{
+    {	$ctor="${under}_GLOBAL_\$I\$$f";
-		local($ctor)="${under}_GLOBAL_\$I\$$f";
 	$tmp=".text\n";
-		$tmp.=".type	$ctor,\@function\n" if ($main'pic);
+	$tmp.=".type	$ctor,\@function\n" if ($::pic);
 	$tmp.=<<___;	# OpenBSD way...
 .globl	$ctor
 .align	2
@@ -756,6 +321,6 @@ $ctor:
 ___
     }
     push(@out,$tmp) if ($tmp);
-	}
+}
 
 1;

crypto/pqueue/pq_compat.h

@@ -57,7 +57,7 @@
  *
  */
 
-#include "opensslconf.h"
+#include <openssl/opensslconf.h>
 #include <openssl/bn.h>
 
 /* 

crypto/rand/rand.h

@@ -137,9 +137,13 @@ void ERR_load_RAND_strings(void);
 /* Error codes for the RAND functions. */
 
 /* Function codes. */
+#define RAND_F_ENG_RAND_GET_RAND_METHOD			 108
 #define RAND_F_FIPS_RAND				 103
 #define RAND_F_FIPS_RAND_BYTES				 102
+#define RAND_F_FIPS_RAND_GET_RAND_METHOD		 109
+#define RAND_F_FIPS_RAND_SET_DT				 106
 #define RAND_F_FIPS_SET_DT				 104
+#define RAND_F_FIPS_SET_PRNG_SEED			 107
 #define RAND_F_FIPS_SET_TEST_MODE			 105
 #define RAND_F_RAND_GET_RAND_METHOD			 101
 #define RAND_F_SSLEAY_RAND_BYTES			 100
@@ -154,6 +158,7 @@ void ERR_load_RAND_strings(void);
 #define RAND_R_PRNG_NOT_REKEYED				 102
 #define RAND_R_PRNG_NOT_RESEEDED			 103
 #define RAND_R_PRNG_NOT_SEEDED				 100
+#define RAND_R_PRNG_SEED_MUST_NOT_MATCH_KEY		 110
 #define RAND_R_PRNG_STUCK				 104
 
 #ifdef  __cplusplus

crypto/rand/rand_eng.c

@@ -109,7 +109,7 @@ const RAND_METHOD *eng_RAND_get_rand_method(const RAND_METHOD **pmeth)
 	if(FIPS_mode()
 		&& *pmeth != FIPS_rand_check())
 	    {
-	    RANDerr(RAND_F_RAND_GET_RAND_METHOD,RAND_R_NON_FIPS_METHOD);
+	    RANDerr(RAND_F_ENG_RAND_GET_RAND_METHOD,RAND_R_NON_FIPS_METHOD);
 	    return 0;
 	    }
 

crypto/rand/rand_err.c

@@ -1,6 +1,6 @@
 /* crypto/rand/rand_err.c */
 /* ====================================================================
- * Copyright (c) 1999-2005 The OpenSSL Project.  All rights reserved.
+ * Copyright (c) 1999-2007 The OpenSSL Project.  All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -70,9 +70,13 @@
 
 static ERR_STRING_DATA RAND_str_functs[]=
 	{
+{ERR_FUNC(RAND_F_ENG_RAND_GET_RAND_METHOD),	"ENG_RAND_GET_RAND_METHOD"},
 {ERR_FUNC(RAND_F_FIPS_RAND),	"FIPS_RAND"},
 {ERR_FUNC(RAND_F_FIPS_RAND_BYTES),	"FIPS_RAND_BYTES"},
+{ERR_FUNC(RAND_F_FIPS_RAND_GET_RAND_METHOD),	"FIPS_RAND_GET_RAND_METHOD"},
+{ERR_FUNC(RAND_F_FIPS_RAND_SET_DT),	"FIPS_RAND_SET_DT"},
 {ERR_FUNC(RAND_F_FIPS_SET_DT),	"FIPS_SET_DT"},
+{ERR_FUNC(RAND_F_FIPS_SET_PRNG_SEED),	"FIPS_SET_PRNG_SEED"},
 {ERR_FUNC(RAND_F_FIPS_SET_TEST_MODE),	"FIPS_SET_TEST_MODE"},
 {ERR_FUNC(RAND_F_RAND_GET_RAND_METHOD),	"RAND_get_rand_method"},
 {ERR_FUNC(RAND_F_SSLEAY_RAND_BYTES),	"SSLEAY_RAND_BYTES"},
@@ -90,6 +94,7 @@ static ERR_STRING_DATA RAND_str_reasons[]=
 {ERR_REASON(RAND_R_PRNG_NOT_REKEYED)     ,"prng not rekeyed"},
 {ERR_REASON(RAND_R_PRNG_NOT_RESEEDED)    ,"prng not reseeded"},
 {ERR_REASON(RAND_R_PRNG_NOT_SEEDED)      ,"PRNG not seeded"},
+{ERR_REASON(RAND_R_PRNG_SEED_MUST_NOT_MATCH_KEY),"prng seed must not match key"},
 {ERR_REASON(RAND_R_PRNG_STUCK)           ,"prng stuck"},
 {0,NULL}
 	};

crypto/rand/rand_lib.c

@@ -91,7 +91,7 @@ static const RAND_METHOD *fips_RAND_get_rand_method(const RAND_METHOD **pmeth)
 	if(FIPS_mode()
 		&& *pmeth != FIPS_rand_check())
 	    {
-	    RANDerr(RAND_F_RAND_GET_RAND_METHOD,RAND_R_NON_FIPS_METHOD);
+	    RANDerr(RAND_F_FIPS_RAND_GET_RAND_METHOD,RAND_R_NON_FIPS_METHOD);
 	    return 0;
 	    }
 

crypto/rc4/rc4_skey.c

@@ -128,7 +128,12 @@ FIPS_NON_FIPS_VCIPHER_Init(RC4)
 		 * module...
 		 *				<appro@fy.chalmers.se>
 		 */
+#ifdef OPENSSL_FIPS
+		unsigned long *ia32cap_ptr = OPENSSL_ia32cap_loc();
+		if (ia32cap_ptr && (*ia32cap_ptr & (1<<28))) {
+#else
 		if (OPENSSL_ia32cap_P & (1<<28)) {
+#endif
 			unsigned char *cp=(unsigned char *)d;
 
 			for (i=0;i<256;i++) cp[i]=i;

crypto/ripemd/asm/rmd-586.pl

@@ -1,7 +1,7 @@
 #!/usr/local/bin/perl
 
 # Normal is the
-# ripemd160_block_asm_host_order(RIPEMD160_CTX *c, ULONG *X,int blocks);
+# ripemd160_block_asm_data_order(RIPEMD160_CTX *c, ULONG *X,int blocks);
 
 $normal=0;
 
@@ -56,7 +56,7 @@ $KR3=0x7A6D76E9;
 	 8, 5,12, 9,12, 5,14, 6, 8,13, 6, 5,15,13,11,11,
  	);
 
-&ripemd160_block("ripemd160_block_asm_host_order");
+&ripemd160_block("ripemd160_block_asm_data_order");
 &asm_finish();
 
 sub Xv

crypto/ripemd/rmd_dgst.c

@@ -84,207 +84,6 @@ FIPS_NON_FIPS_MD_Init(RIPEMD160)
 	return 1;
 	}
 
-#ifndef ripemd160_block_host_order
-#ifdef X
-#undef X
-#endif
-#define X(i)	XX[i]
-void ripemd160_block_host_order (RIPEMD160_CTX *ctx, const void *p, size_t num)
-	{
-	const RIPEMD160_LONG *XX=p;
-	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)
-		{
-
-	A=ctx->A; B=ctx->B; C=ctx->C; D=ctx->D; E=ctx->E;
-
-	RIP1(A,B,C,D,E,WL00,SL00);
-	RIP1(E,A,B,C,D,WL01,SL01);
-	RIP1(D,E,A,B,C,WL02,SL02);
-	RIP1(C,D,E,A,B,WL03,SL03);
-	RIP1(B,C,D,E,A,WL04,SL04);
-	RIP1(A,B,C,D,E,WL05,SL05);
-	RIP1(E,A,B,C,D,WL06,SL06);
-	RIP1(D,E,A,B,C,WL07,SL07);
-	RIP1(C,D,E,A,B,WL08,SL08);
-	RIP1(B,C,D,E,A,WL09,SL09);
-	RIP1(A,B,C,D,E,WL10,SL10);
-	RIP1(E,A,B,C,D,WL11,SL11);
-	RIP1(D,E,A,B,C,WL12,SL12);
-	RIP1(C,D,E,A,B,WL13,SL13);
-	RIP1(B,C,D,E,A,WL14,SL14);
-	RIP1(A,B,C,D,E,WL15,SL15);
-
-	RIP2(E,A,B,C,D,WL16,SL16,KL1);
-	RIP2(D,E,A,B,C,WL17,SL17,KL1);
-	RIP2(C,D,E,A,B,WL18,SL18,KL1);
-	RIP2(B,C,D,E,A,WL19,SL19,KL1);
-	RIP2(A,B,C,D,E,WL20,SL20,KL1);
-	RIP2(E,A,B,C,D,WL21,SL21,KL1);
-	RIP2(D,E,A,B,C,WL22,SL22,KL1);
-	RIP2(C,D,E,A,B,WL23,SL23,KL1);
-	RIP2(B,C,D,E,A,WL24,SL24,KL1);
-	RIP2(A,B,C,D,E,WL25,SL25,KL1);
-	RIP2(E,A,B,C,D,WL26,SL26,KL1);
-	RIP2(D,E,A,B,C,WL27,SL27,KL1);
-	RIP2(C,D,E,A,B,WL28,SL28,KL1);
-	RIP2(B,C,D,E,A,WL29,SL29,KL1);
-	RIP2(A,B,C,D,E,WL30,SL30,KL1);
-	RIP2(E,A,B,C,D,WL31,SL31,KL1);
-
-	RIP3(D,E,A,B,C,WL32,SL32,KL2);
-	RIP3(C,D,E,A,B,WL33,SL33,KL2);
-	RIP3(B,C,D,E,A,WL34,SL34,KL2);
-	RIP3(A,B,C,D,E,WL35,SL35,KL2);
-	RIP3(E,A,B,C,D,WL36,SL36,KL2);
-	RIP3(D,E,A,B,C,WL37,SL37,KL2);
-	RIP3(C,D,E,A,B,WL38,SL38,KL2);
-	RIP3(B,C,D,E,A,WL39,SL39,KL2);
-	RIP3(A,B,C,D,E,WL40,SL40,KL2);
-	RIP3(E,A,B,C,D,WL41,SL41,KL2);
-	RIP3(D,E,A,B,C,WL42,SL42,KL2);
-	RIP3(C,D,E,A,B,WL43,SL43,KL2);
-	RIP3(B,C,D,E,A,WL44,SL44,KL2);
-	RIP3(A,B,C,D,E,WL45,SL45,KL2);
-	RIP3(E,A,B,C,D,WL46,SL46,KL2);
-	RIP3(D,E,A,B,C,WL47,SL47,KL2);
-
-	RIP4(C,D,E,A,B,WL48,SL48,KL3);
-	RIP4(B,C,D,E,A,WL49,SL49,KL3);
-	RIP4(A,B,C,D,E,WL50,SL50,KL3);
-	RIP4(E,A,B,C,D,WL51,SL51,KL3);
-	RIP4(D,E,A,B,C,WL52,SL52,KL3);
-	RIP4(C,D,E,A,B,WL53,SL53,KL3);
-	RIP4(B,C,D,E,A,WL54,SL54,KL3);
-	RIP4(A,B,C,D,E,WL55,SL55,KL3);
-	RIP4(E,A,B,C,D,WL56,SL56,KL3);
-	RIP4(D,E,A,B,C,WL57,SL57,KL3);
-	RIP4(C,D,E,A,B,WL58,SL58,KL3);
-	RIP4(B,C,D,E,A,WL59,SL59,KL3);
-	RIP4(A,B,C,D,E,WL60,SL60,KL3);
-	RIP4(E,A,B,C,D,WL61,SL61,KL3);
-	RIP4(D,E,A,B,C,WL62,SL62,KL3);
-	RIP4(C,D,E,A,B,WL63,SL63,KL3);
-
-	RIP5(B,C,D,E,A,WL64,SL64,KL4);
-	RIP5(A,B,C,D,E,WL65,SL65,KL4);
-	RIP5(E,A,B,C,D,WL66,SL66,KL4);
-	RIP5(D,E,A,B,C,WL67,SL67,KL4);
-	RIP5(C,D,E,A,B,WL68,SL68,KL4);
-	RIP5(B,C,D,E,A,WL69,SL69,KL4);
-	RIP5(A,B,C,D,E,WL70,SL70,KL4);
-	RIP5(E,A,B,C,D,WL71,SL71,KL4);
-	RIP5(D,E,A,B,C,WL72,SL72,KL4);
-	RIP5(C,D,E,A,B,WL73,SL73,KL4);
-	RIP5(B,C,D,E,A,WL74,SL74,KL4);
-	RIP5(A,B,C,D,E,WL75,SL75,KL4);
-	RIP5(E,A,B,C,D,WL76,SL76,KL4);
-	RIP5(D,E,A,B,C,WL77,SL77,KL4);
-	RIP5(C,D,E,A,B,WL78,SL78,KL4);
-	RIP5(B,C,D,E,A,WL79,SL79,KL4);
-
-	a=A; b=B; c=C; d=D; e=E;
-	/* Do other half */
-	A=ctx->A; B=ctx->B; C=ctx->C; D=ctx->D; E=ctx->E;
-
-	RIP5(A,B,C,D,E,WR00,SR00,KR0);
-	RIP5(E,A,B,C,D,WR01,SR01,KR0);
-	RIP5(D,E,A,B,C,WR02,SR02,KR0);
-	RIP5(C,D,E,A,B,WR03,SR03,KR0);
-	RIP5(B,C,D,E,A,WR04,SR04,KR0);
-	RIP5(A,B,C,D,E,WR05,SR05,KR0);
-	RIP5(E,A,B,C,D,WR06,SR06,KR0);
-	RIP5(D,E,A,B,C,WR07,SR07,KR0);
-	RIP5(C,D,E,A,B,WR08,SR08,KR0);
-	RIP5(B,C,D,E,A,WR09,SR09,KR0);
-	RIP5(A,B,C,D,E,WR10,SR10,KR0);
-	RIP5(E,A,B,C,D,WR11,SR11,KR0);
-	RIP5(D,E,A,B,C,WR12,SR12,KR0);
-	RIP5(C,D,E,A,B,WR13,SR13,KR0);
-	RIP5(B,C,D,E,A,WR14,SR14,KR0);
-	RIP5(A,B,C,D,E,WR15,SR15,KR0);
-
-	RIP4(E,A,B,C,D,WR16,SR16,KR1);
-	RIP4(D,E,A,B,C,WR17,SR17,KR1);
-	RIP4(C,D,E,A,B,WR18,SR18,KR1);
-	RIP4(B,C,D,E,A,WR19,SR19,KR1);
-	RIP4(A,B,C,D,E,WR20,SR20,KR1);
-	RIP4(E,A,B,C,D,WR21,SR21,KR1);
-	RIP4(D,E,A,B,C,WR22,SR22,KR1);
-	RIP4(C,D,E,A,B,WR23,SR23,KR1);
-	RIP4(B,C,D,E,A,WR24,SR24,KR1);
-	RIP4(A,B,C,D,E,WR25,SR25,KR1);
-	RIP4(E,A,B,C,D,WR26,SR26,KR1);
-	RIP4(D,E,A,B,C,WR27,SR27,KR1);
-	RIP4(C,D,E,A,B,WR28,SR28,KR1);
-	RIP4(B,C,D,E,A,WR29,SR29,KR1);
-	RIP4(A,B,C,D,E,WR30,SR30,KR1);
-	RIP4(E,A,B,C,D,WR31,SR31,KR1);
-
-	RIP3(D,E,A,B,C,WR32,SR32,KR2);
-	RIP3(C,D,E,A,B,WR33,SR33,KR2);
-	RIP3(B,C,D,E,A,WR34,SR34,KR2);
-	RIP3(A,B,C,D,E,WR35,SR35,KR2);
-	RIP3(E,A,B,C,D,WR36,SR36,KR2);
-	RIP3(D,E,A,B,C,WR37,SR37,KR2);
-	RIP3(C,D,E,A,B,WR38,SR38,KR2);
-	RIP3(B,C,D,E,A,WR39,SR39,KR2);
-	RIP3(A,B,C,D,E,WR40,SR40,KR2);
-	RIP3(E,A,B,C,D,WR41,SR41,KR2);
-	RIP3(D,E,A,B,C,WR42,SR42,KR2);
-	RIP3(C,D,E,A,B,WR43,SR43,KR2);
-	RIP3(B,C,D,E,A,WR44,SR44,KR2);
-	RIP3(A,B,C,D,E,WR45,SR45,KR2);
-	RIP3(E,A,B,C,D,WR46,SR46,KR2);
-	RIP3(D,E,A,B,C,WR47,SR47,KR2);
-
-	RIP2(C,D,E,A,B,WR48,SR48,KR3);
-	RIP2(B,C,D,E,A,WR49,SR49,KR3);
-	RIP2(A,B,C,D,E,WR50,SR50,KR3);
-	RIP2(E,A,B,C,D,WR51,SR51,KR3);
-	RIP2(D,E,A,B,C,WR52,SR52,KR3);
-	RIP2(C,D,E,A,B,WR53,SR53,KR3);
-	RIP2(B,C,D,E,A,WR54,SR54,KR3);
-	RIP2(A,B,C,D,E,WR55,SR55,KR3);
-	RIP2(E,A,B,C,D,WR56,SR56,KR3);
-	RIP2(D,E,A,B,C,WR57,SR57,KR3);
-	RIP2(C,D,E,A,B,WR58,SR58,KR3);
-	RIP2(B,C,D,E,A,WR59,SR59,KR3);
-	RIP2(A,B,C,D,E,WR60,SR60,KR3);
-	RIP2(E,A,B,C,D,WR61,SR61,KR3);
-	RIP2(D,E,A,B,C,WR62,SR62,KR3);
-	RIP2(C,D,E,A,B,WR63,SR63,KR3);
-
-	RIP1(B,C,D,E,A,WR64,SR64);
-	RIP1(A,B,C,D,E,WR65,SR65);
-	RIP1(E,A,B,C,D,WR66,SR66);
-	RIP1(D,E,A,B,C,WR67,SR67);
-	RIP1(C,D,E,A,B,WR68,SR68);
-	RIP1(B,C,D,E,A,WR69,SR69);
-	RIP1(A,B,C,D,E,WR70,SR70);
-	RIP1(E,A,B,C,D,WR71,SR71);
-	RIP1(D,E,A,B,C,WR72,SR72);
-	RIP1(C,D,E,A,B,WR73,SR73);
-	RIP1(B,C,D,E,A,WR74,SR74);
-	RIP1(A,B,C,D,E,WR75,SR75);
-	RIP1(E,A,B,C,D,WR76,SR76);
-	RIP1(D,E,A,B,C,WR77,SR77);
-	RIP1(C,D,E,A,B,WR78,SR78);
-	RIP1(B,C,D,E,A,WR79,SR79);
-
-	D     =ctx->B+c+D;
-	ctx->B=ctx->C+d+E;
-	ctx->C=ctx->D+e+A;
-	ctx->D=ctx->E+a+B;
-	ctx->E=ctx->A+b+C;
-	ctx->A=D;
-
-		}
-	}
-#endif
-
 #ifndef ripemd160_block_data_order
 #ifdef X
 #undef X

crypto/ripemd/rmd_locl.h

@@ -72,32 +72,20 @@
  */
 #ifdef RMD160_ASM
 # if defined(__i386) || defined(__i386__) || defined(_M_IX86) || defined(__INTEL__)
-#  if !defined(B_ENDIAN)
+#  define ripemd160_block_host_order ripemd160_block_asm_data_order
-#   define ripemd160_block_host_order ripemd160_block_asm_host_order
-#  endif
 # endif
 #endif
 
-void ripemd160_block_host_order (RIPEMD160_CTX *c, const void *p,size_t num);
 void ripemd160_block_data_order (RIPEMD160_CTX *c, const void *p,size_t num);
 
-#if defined(__i386) || defined(__i386__) || defined(_M_IX86) || defined(__INTEL__)
-# if !defined(B_ENDIAN)
-#  define ripemd160_block_data_order ripemd160_block_host_order
-# endif
-#endif
-
 #define DATA_ORDER_IS_LITTLE_ENDIAN
 
 #define HASH_LONG               RIPEMD160_LONG
-#define HASH_LONG_LOG2          RIPEMD160_LONG_LOG2
 #define HASH_CTX                RIPEMD160_CTX
 #define HASH_CBLOCK             RIPEMD160_CBLOCK
-#define HASH_LBLOCK             RIPEMD160_LBLOCK
 #define HASH_UPDATE             RIPEMD160_Update
 #define HASH_TRANSFORM          RIPEMD160_Transform
 #define HASH_FINAL              RIPEMD160_Final
-#define HASH_BLOCK_HOST_ORDER   ripemd160_block_host_order
 #define	HASH_MAKE_STRING(c,s)	do {	\
 	unsigned long ll;		\
 	ll=(c)->A; HOST_l2c(ll,(s));	\
@@ -106,9 +94,7 @@ void ripemd160_block_data_order (RIPEMD160_CTX *c, const void *p,size_t num);
 	ll=(c)->D; HOST_l2c(ll,(s));	\
 	ll=(c)->E; HOST_l2c(ll,(s));	\
 	} while (0)
-#if !defined(L_ENDIAN) || defined(ripemd160_block_data_order)
 #define HASH_BLOCK_DATA_ORDER   ripemd160_block_data_order
-#endif
 
 #include "md32_common.h"
 

crypto/rsa/rsa.h

@@ -182,6 +182,8 @@ struct rsa_st
 # define OPENSSL_RSA_MAX_MODULUS_BITS	16384
 #endif
 
+#define OPENSSL_RSA_FIPS_MIN_MODULUS_BITS 1024
+
 #ifndef OPENSSL_RSA_SMALL_MODULUS_BITS
 # define OPENSSL_RSA_SMALL_MODULUS_BITS	3072
 #endif
@@ -404,6 +406,8 @@ void ERR_load_RSA_strings(void);
 /* Error codes for the RSA functions. */
 
 /* Function codes. */
+#define RSA_F_FIPS_RSA_SIGN				 140
+#define RSA_F_FIPS_RSA_VERIFY				 141
 #define RSA_F_MEMORY_LOCK				 100
 #define RSA_F_RSA_BUILTIN_KEYGEN			 129
 #define RSA_F_RSA_CHECK_KEY				 123
@@ -439,6 +443,7 @@ void ERR_load_RSA_strings(void);
 #define RSA_F_RSA_PUBLIC_DECRYPT			 138
 #define RSA_F_RSA_SETUP_BLINDING			 136
 #define RSA_F_RSA_SET_DEFAULT_METHOD			 139
+#define RSA_F_RSA_SET_METHOD				 142
 #define RSA_F_RSA_SIGN					 117
 #define RSA_F_RSA_SIGN_ASN1_OCTET_STRING		 118
 #define RSA_F_RSA_VERIFY				 119

crypto/rsa/rsa_eng.c

@@ -121,7 +121,7 @@ int RSA_set_method(RSA *rsa, const RSA_METHOD *meth)
 #ifdef OPENSSL_FIPS
 	if (FIPS_mode() && !(meth->flags & RSA_FLAG_FIPS_METHOD))
 		{
-		RSAerr(RSA_F_RSA_SET_DEFAULT_METHOD, RSA_R_NON_FIPS_METHOD);
+		RSAerr(RSA_F_RSA_SET_METHOD, RSA_R_NON_FIPS_METHOD);
 		return 0;
 		}
 #endif
@@ -180,7 +180,7 @@ RSA *RSA_new_method(ENGINE *engine)
 #ifdef OPENSSL_FIPS
 	if (FIPS_mode() && !(ret->meth->flags & RSA_FLAG_FIPS_METHOD))
 		{
-		RSAerr(RSA_F_RSA_SET_DEFAULT_METHOD, RSA_R_NON_FIPS_METHOD);
+		RSAerr(RSA_F_RSA_NEW_METHOD, RSA_R_NON_FIPS_METHOD);
 #ifndef OPENSSL_NO_ENGINE
 		if (ret->engine)
 			ENGINE_finish(ret->engine);

crypto/rsa/rsa_err.c

@@ -70,6 +70,8 @@
 
 static ERR_STRING_DATA RSA_str_functs[]=
 	{
+{ERR_FUNC(RSA_F_FIPS_RSA_SIGN),	"FIPS_RSA_SIGN"},
+{ERR_FUNC(RSA_F_FIPS_RSA_VERIFY),	"FIPS_RSA_VERIFY"},
 {ERR_FUNC(RSA_F_MEMORY_LOCK),	"MEMORY_LOCK"},
 {ERR_FUNC(RSA_F_RSA_BUILTIN_KEYGEN),	"RSA_BUILTIN_KEYGEN"},
 {ERR_FUNC(RSA_F_RSA_CHECK_KEY),	"RSA_check_key"},
@@ -105,6 +107,7 @@ static ERR_STRING_DATA RSA_str_functs[]=
 {ERR_FUNC(RSA_F_RSA_PUBLIC_DECRYPT),	"RSA_public_decrypt"},
 {ERR_FUNC(RSA_F_RSA_SETUP_BLINDING),	"RSA_setup_blinding"},
 {ERR_FUNC(RSA_F_RSA_SET_DEFAULT_METHOD),	"RSA_set_default_method"},
+{ERR_FUNC(RSA_F_RSA_SET_METHOD),	"RSA_set_method"},
 {ERR_FUNC(RSA_F_RSA_SIGN),	"RSA_sign"},
 {ERR_FUNC(RSA_F_RSA_SIGN_ASN1_OCTET_STRING),	"RSA_sign_ASN1_OCTET_STRING"},
 {ERR_FUNC(RSA_F_RSA_VERIFY),	"RSA_verify"},

crypto/sha/Makefile

@@ -65,6 +65,15 @@ sha256-ia64.s: asm/sha512-ia64.pl
 sha512-ia64.s: asm/sha512-ia64.pl
 	(cd asm; $(PERL) sha512-ia64.pl ../$@ $(CFLAGS))
 
+# Solaris make has to be explicitly told
+sha1-x86_64.s:	asm/sha1-x86_64.pl;	$(PERL) asm/sha1-x86_64.pl $@
+sha256-x86_64.s:asm/sha512-x86_64.pl;	$(PERL) asm/sha512-x86_64.pl $@
+sha512-x86_64.s:asm/sha512-x86_64.pl;	$(PERL) asm/sha512-x86_64.pl $@
+# GNU make "catch all"
+sha1-%.s:	asm/sha1-%.pl;		$(PERL) $< $@
+sha256-%.s:	asm/sha512-%.pl;	$(PERL) $< $@
+sha512-%.s:	asm/sha512-%.pl;	$(PERL) $< $@
+
 files:
 	$(PERL) $(TOP)/util/files.pl Makefile >> $(TOP)/MINFO
 
@@ -108,16 +117,23 @@ sha1_one.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 sha1_one.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
 sha1_one.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
 sha1_one.o: sha1_one.c
-sha1dgst.o: ../../include/openssl/opensslconf.h
+sha1dgst.o: ../../include/openssl/e_os2.h ../../include/openssl/opensslconf.h
-sha1dgst.o: ../../include/openssl/opensslv.h sha1dgst.c
+sha1dgst.o: ../../include/openssl/opensslv.h ../../include/openssl/sha.h
+sha1dgst.o: ../md32_common.h sha1dgst.c sha_locl.h
 sha256.o: ../../include/openssl/crypto.h ../../include/openssl/e_os2.h
 sha256.o: ../../include/openssl/fips.h ../../include/openssl/opensslconf.h
 sha256.o: ../../include/openssl/opensslv.h ../../include/openssl/ossl_typ.h
 sha256.o: ../../include/openssl/safestack.h ../../include/openssl/sha.h
 sha256.o: ../../include/openssl/stack.h ../../include/openssl/symhacks.h
-sha256.o: sha256.c
+sha256.o: ../md32_common.h sha256.c
-sha512.o: ../../include/openssl/fips.h ../../include/openssl/opensslconf.h
+sha512.o: ../../e_os.h ../../include/openssl/bio.h
-sha512.o: sha512.c
+sha512.o: ../../include/openssl/buffer.h ../../include/openssl/crypto.h
+sha512.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
+sha512.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h
+sha512.o: ../../include/openssl/opensslconf.h ../../include/openssl/opensslv.h
+sha512.o: ../../include/openssl/ossl_typ.h ../../include/openssl/safestack.h
+sha512.o: ../../include/openssl/sha.h ../../include/openssl/stack.h
+sha512.o: ../../include/openssl/symhacks.h ../cryptlib.h sha512.c
 sha_dgst.o: ../../include/openssl/bio.h ../../include/openssl/crypto.h
 sha_dgst.o: ../../include/openssl/e_os2.h ../../include/openssl/err.h
 sha_dgst.o: ../../include/openssl/fips.h ../../include/openssl/lhash.h

crypto/sha/asm/sha1-586.pl

@@ -1,4 +1,16 @@
-#!/usr/local/bin/perl
+#!/usr/bin/env perl
+
+# ====================================================================
+# [Re]written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# "[Re]written" was achieved in two major overhauls. In 2004 BODY_*
+# functions were re-implemented to address P4 performance issue [see
+# commentary below], and in 2006 the rest was rewritten in order to
+# gain freedom to liberate licensing terms.
 
 # It was noted that Intel IA-32 C compiler generates code which
 # performs ~30% *faster* on P4 CPU than original *hand-coded*
@@ -17,90 +29,27 @@
 # improvement on P4 outweights the loss and incorporate this
 # re-tuned code to 0.9.7 and later.
 # ----------------------------------------------------------------
-# Those who for any particular reason absolutely must score on
-# Pentium can replace this module with one from 0.9.6 distribution.
-# This "offer" shall be revoked the moment programming interface to
-# this module is changed, in which case this paragraph should be
-# removed.
-# ----------------------------------------------------------------
 #					<appro@fy.chalmers.se>
 
-$normal=0;
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
-
+push(@INC,"${dir}","${dir}../../perlasm");
-push(@INC,"perlasm","../../perlasm");
 require "x86asm.pl";
 
 &asm_init($ARGV[0],"sha1-586.pl",$ARGV[$#ARGV] eq "386");
 
 $A="eax";
-$B="ecx";
+$B="ebx";
-$C="ebx";
+$C="ecx";
 $D="edx";
 $E="edi";
 $T="esi";
 $tmp1="ebp";
 
-$off=9*4;
+@V=($A,$B,$C,$D,$E,$T);
-
-@K=(0x5a827999,0x6ed9eba1,0x8f1bbcdc,0xca62c1d6);
-
-&sha1_block_data("sha1_block_asm_data_order");
-
-&asm_finish();
-
-sub Nn
-	{
-	local($p)=@_;
-	local(%n)=($A,$T,$B,$A,$C,$B,$D,$C,$E,$D,$T,$E);
-	return($n{$p});
-	}
-
-sub Np
-	{
-	local($p)=@_;
-	local(%n)=($A,$T,$B,$A,$C,$B,$D,$C,$E,$D,$T,$E);
-	local(%n)=($A,$B,$B,$C,$C,$D,$D,$E,$E,$T,$T,$A);
-	return($n{$p});
-	}
-
-sub Na
-	{
-	local($n)=@_;
-	return( (($n   )&0x0f),
-		(($n+ 2)&0x0f),
-		(($n+ 8)&0x0f),
-		(($n+13)&0x0f),
-		(($n+ 1)&0x0f));
-	}
-
-sub X_expand
-	{
-	local($in)=@_;
-
-	&comment("First, load the words onto the stack in network byte order");
-	for ($i=0; $i<16; $i+=2)
-		{
-		&mov($A,&DWP(($i+0)*4,$in,"",0));# unless $i == 0;
-		 &mov($B,&DWP(($i+1)*4,$in,"",0));
-		&bswap($A);
-		 &bswap($B);
-		&mov(&swtmp($i+0),$A);
-		 &mov(&swtmp($i+1),$B);
-		}
-
-	&comment("We now have the X array on the stack");
-	&comment("starting at sp-4");
-	}
-
-# Rules of engagement
-# F is always trashable at the start, the running total.
-# E becomes the next F so it can be trashed after it has been 'accumulated'
-# F becomes A in the next round.  We don't need to access it much.
-# During the X update part, the result ends up in $X[$n0].
 
 sub BODY_00_15
 	{
-	local($pos,$K,$X,$n,$a,$b,$c,$d,$e,$f)=@_;
+	local($n,$a,$b,$c,$d,$e,$f)=@_;
 
 	&comment("00_15 $n");
 
@@ -109,37 +58,37 @@ sub BODY_00_15
 	 else        { &mov($a,$tmp1); }
 	&rotl($tmp1,5);			# tmp1=ROTATE(a,5)
 	 &xor($f,$d);
-	&and($f,$b);
 	&add($tmp1,$e);			# tmp1+=e;
-	&mov($e,&swtmp($n));		# e becomes volatile and
+	 &and($f,$b);
-	 				# is loaded with xi
+	&mov($e,&swtmp($n%16));		# e becomes volatile and is loaded
+	 				# with xi, also note that e becomes
+					# f in next round...
 	 &xor($f,$d);			# f holds F_00_19(b,c,d)
 	&rotr($b,2);			# b=ROTATE(b,30)
-	 &lea($tmp1,&DWP($K,$tmp1,$e,1));# tmp1+=K_00_19+xi
+	 &lea($tmp1,&DWP(0x5a827999,$tmp1,$e));	# tmp1+=K_00_19+xi
 
 	if ($n==15) { &add($f,$tmp1); }	# f+=tmp1
-	else        { &add($tmp1,$f); }
+	else        { &add($tmp1,$f); }	# f becomes a in next round
 	}
 
 sub BODY_16_19
 	{
-	local($pos,$K,$X,$n,$a,$b,$c,$d,$e,$f)=@_;
+	local($n,$a,$b,$c,$d,$e,$f)=@_;
-	local($n0,$n1,$n2,$n3,$np)=&Na($n);
 
 	&comment("16_19 $n");
 
-	&mov($f,&swtmp($n1));		# f to hold Xupdate(xi,xa,xb,xc,xd)
+	&mov($f,&swtmp($n%16));		# f to hold Xupdate(xi,xa,xb,xc,xd)
 	 &mov($tmp1,$c);		# tmp1 to hold F_00_19(b,c,d)
-	&xor($f,&swtmp($n0));
+	&xor($f,&swtmp(($n+2)%16));
 	 &xor($tmp1,$d);
-	&xor($f,&swtmp($n2));
+	&xor($f,&swtmp(($n+8)%16));
 	 &and($tmp1,$b);		# tmp1 holds F_00_19(b,c,d)
 	&rotr($b,2);			# b=ROTATE(b,30)
-	 &xor($f,&swtmp($n3));		# f holds xa^xb^xc^xd
+	 &xor($f,&swtmp(($n+13)%16));	# f holds xa^xb^xc^xd
-	&rotl($f,1);			# f=ROATE(f,1)
+	&rotl($f,1);			# f=ROTATE(f,1)
 	 &xor($tmp1,$d);		# tmp1=F_00_19(b,c,d)
-	&mov(&swtmp($n0),$f);		# xi=f
+	&mov(&swtmp($n%16),$f);		# xi=f
-	&lea($f,&DWP($K,$f,$e,1));	# f+=K_00_19+e
+	&lea($f,&DWP(0x5a827999,$f,$e));# f+=K_00_19+e
 	 &mov($e,$a);			# e becomes volatile
 	&rotl($e,5);			# e=ROTATE(a,5)
 	 &add($f,$tmp1);		# f+=F_00_19(b,c,d)
@@ -148,48 +97,47 @@ sub BODY_16_19
 
 sub BODY_20_39
 	{
-	local($pos,$K,$X,$n,$a,$b,$c,$d,$e,$f)=@_;
+	local($n,$a,$b,$c,$d,$e,$f)=@_;
+	local $K=($n<40)?0x6ed9eba1:0xca62c1d6;
 
 	&comment("20_39 $n");
-	local($n0,$n1,$n2,$n3,$np)=&Na($n);
 
 	&mov($tmp1,$b);			# tmp1 to hold F_20_39(b,c,d)
-	 &mov($f,&swtmp($n0));		# f to hold Xupdate(xi,xa,xb,xc,xd)
+	 &mov($f,&swtmp($n%16));	# f to hold Xupdate(xi,xa,xb,xc,xd)
 	&rotr($b,2);			# b=ROTATE(b,30)
-	 &xor($f,&swtmp($n1));
+	 &xor($f,&swtmp(($n+2)%16));
 	&xor($tmp1,$c);
-	 &xor($f,&swtmp($n2));
+	 &xor($f,&swtmp(($n+8)%16));
 	&xor($tmp1,$d);			# tmp1 holds F_20_39(b,c,d)
-	 &xor($f,&swtmp($n3));		# f holds xa^xb^xc^xd
+	 &xor($f,&swtmp(($n+13)%16));	# f holds xa^xb^xc^xd
 	&rotl($f,1);			# f=ROTATE(f,1)
 	 &add($tmp1,$e);
-	&mov(&swtmp($n0),$f);		# xi=f
+	&mov(&swtmp($n%16),$f);		# xi=f
 	 &mov($e,$a);			# e becomes volatile
 	&rotl($e,5);			# e=ROTATE(a,5)
-	 &lea($f,&DWP($K,$f,$tmp1,1));	# f+=K_20_39+e
+	 &lea($f,&DWP($K,$f,$tmp1));	# f+=K_20_39+e
 	&add($f,$e);			# f+=ROTATE(a,5)
 	}
 
 sub BODY_40_59
 	{
-	local($pos,$K,$X,$n,$a,$b,$c,$d,$e,$f)=@_;
+	local($n,$a,$b,$c,$d,$e,$f)=@_;
 
 	&comment("40_59 $n");
-	local($n0,$n1,$n2,$n3,$np)=&Na($n);
 
-	&mov($f,&swtmp($n0));		# f to hold Xupdate(xi,xa,xb,xc,xd)
+	&mov($f,&swtmp($n%16));		# f to hold Xupdate(xi,xa,xb,xc,xd)
-	 &mov($tmp1,&swtmp($n1));
+	 &mov($tmp1,&swtmp(($n+2)%16));
 	&xor($f,$tmp1);
-	 &mov($tmp1,&swtmp($n2));
+	 &mov($tmp1,&swtmp(($n+8)%16));
 	&xor($f,$tmp1);
-	 &mov($tmp1,&swtmp($n3));
+	 &mov($tmp1,&swtmp(($n+13)%16));
 	&xor($f,$tmp1);			# f holds xa^xb^xc^xd
 	 &mov($tmp1,$b);		# tmp1 to hold F_40_59(b,c,d)
 	&rotl($f,1);			# f=ROTATE(f,1)
 	 &or($tmp1,$c);
-	&mov(&swtmp($n0),$f);		# xi=f
+	&mov(&swtmp($n%16),$f);		# xi=f
 	 &and($tmp1,$d);
-	&lea($f,&DWP($K,$f,$e,1));	# f+=K_40_59+e
+	&lea($f,&DWP(0x8f1bbcdc,$f,$e));# f+=K_40_59+e
 	 &mov($e,$b);			# e becomes volatile and is used
 					# to calculate F_40_59(b,c,d)
 	&rotr($b,2);			# b=ROTATE(b,30)
@@ -201,230 +149,71 @@ sub BODY_40_59
 	&add($f,$e);			# f+=ROTATE(a,5)
 	}
 
-sub BODY_60_79
+&function_begin("sha1_block_data_order",16);
+	&mov($tmp1,&wparam(0));	# SHA_CTX *c
+	&mov($T,&wparam(1));	# const void *input
+	&mov($A,&wparam(2));	# size_t num
+	&stack_push(16);	# allocate X[16]
+	&shl($A,6);
+	&add($A,$T);
+	&mov(&wparam(2),$A);	# pointer beyond the end of input
+	&mov($E,&DWP(16,$tmp1));# pre-load E
+
+	&set_label("loop",16);
+
+	# copy input chunk to X, but reversing byte order!
+	for ($i=0; $i<16; $i+=4)
 		{
-	&BODY_20_39(@_);
+		&mov($A,&DWP(4*($i+0),$T));
-	}
+		&mov($B,&DWP(4*($i+1),$T));
-
+		&mov($C,&DWP(4*($i+2),$T));
-sub sha1_block_host
+		&mov($D,&DWP(4*($i+3),$T));
-	{
+		&bswap($A);
-	local($name, $sclabel)=@_;
+		&bswap($B);
-
+		&bswap($C);
-	&function_begin_B($name,"");
+		&bswap($D);
-
-	# parameter 1 is the MD5_CTX structure.
-	# A	0
-	# B	4
-	# C	8
-	# D 	12
-	# E 	16
-
-	&mov("ecx",	&wparam(2));
-	 &push("esi");
-	&shl("ecx",6);
-	 &mov("esi",	&wparam(1));
-	&push("ebp");
-	 &add("ecx","esi");	# offset to leave on
-	&push("ebx");
-	 &mov("ebp",	&wparam(0));
-	&push("edi");
-	 &mov($D,	&DWP(12,"ebp","",0));
-	&stack_push(18+9);
-	 &mov($E,	&DWP(16,"ebp","",0));
-	&mov($C,	&DWP( 8,"ebp","",0));
-	 &mov(&swtmp(17),"ecx");
-
-	&comment("First we need to setup the X array");
-
-	for ($i=0; $i<16; $i+=2)
-		{
-		&mov($A,&DWP(($i+0)*4,"esi","",0));# unless $i == 0;
-		 &mov($B,&DWP(($i+1)*4,"esi","",0));
 		&mov(&swtmp($i+0),$A);
 		&mov(&swtmp($i+1),$B);
+		&mov(&swtmp($i+2),$C);
+		&mov(&swtmp($i+3),$D);
 		}
-	&jmp($sclabel);
+	&mov(&wparam(1),$T);	# redundant in 1st spin
-	&function_end_B($name);
-	}
 
+	&mov($A,&DWP(0,$tmp1));	# load SHA_CTX
+	&mov($B,&DWP(4,$tmp1));
+	&mov($C,&DWP(8,$tmp1));
+	&mov($D,&DWP(12,$tmp1));
+	# E is pre-loaded
 
-sub sha1_block_data
+	for($i=0;$i<16;$i++)	{ &BODY_00_15($i,@V); unshift(@V,pop(@V)); }
-	{
+	for(;$i<20;$i++)	{ &BODY_16_19($i,@V); unshift(@V,pop(@V)); }
-	local($name)=@_;
+	for(;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+	for(;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+	for(;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
 
-	&function_begin_B($name,"");
+	(($V[5] eq $D) and ($V[0] eq $E)) or die;	# double-check
 
-	# parameter 1 is the MD5_CTX structure.
+	&mov($tmp1,&wparam(0));	# re-load SHA_CTX*
-	# A	0
+	&mov($D,&wparam(1));	# D is last "T" and is discarded
-	# B	4
-	# C	8
-	# D 	12
-	# E 	16
 
-	&mov("ecx",	&wparam(2));
+	&add($E,&DWP(0,$tmp1));	# E is last "A"...
-	 &push("esi");
+	&add($T,&DWP(4,$tmp1));
-	&shl("ecx",6);
+	&add($A,&DWP(8,$tmp1));
-	 &mov("esi",	&wparam(1));
+	&add($B,&DWP(12,$tmp1));
-	&push("ebp");
+	&add($C,&DWP(16,$tmp1));
-	 &add("ecx","esi");	# offset to leave on
-	&push("ebx");
-	 &mov("ebp",	&wparam(0));
-	&push("edi");
-	 &mov($D,	&DWP(12,"ebp","",0));
-	&stack_push(18+9);
-	 &mov($E,	&DWP(16,"ebp","",0));
-	&mov($C,	&DWP( 8,"ebp","",0));
-	 &mov(&swtmp(17),"ecx");
 
-	&comment("First we need to setup the X array");
+	&mov(&DWP(0,$tmp1),$E);	# update SHA_CTX
+	 &add($D,64);		# advance input pointer
+	&mov(&DWP(4,$tmp1),$T);
+	 &cmp($D,&wparam(2));	# have we reached the end yet?
+	&mov(&DWP(8,$tmp1),$A);
+	 &mov($E,$C);		# C is last "E" which needs to be "pre-loaded"
+	&mov(&DWP(12,$tmp1),$B);
+	 &mov($T,$D);		# input pointer
+	&mov(&DWP(16,$tmp1),$C);
+	&jb(&label("loop"));
 
-	&set_label("start") unless $normal;
+	&stack_pop(16);
-
+&function_end("sha1_block_data_order");
-	&X_expand("esi");
-	 &mov(&wparam(1),"esi");
-
-	&set_label("shortcut", 0, 1);
-	&comment("");
-	&comment("Start processing");
-
-	# odd start
-	&mov($A,	&DWP( 0,"ebp","",0));
-	 &mov($B,	&DWP( 4,"ebp","",0));
-	$X="esp";
-	&BODY_00_15(-2,$K[0],$X, 0,$A,$B,$C,$D,$E,$T);
-	&BODY_00_15( 0,$K[0],$X, 1,$T,$A,$B,$C,$D,$E);
-	&BODY_00_15( 0,$K[0],$X, 2,$E,$T,$A,$B,$C,$D);
-	&BODY_00_15( 0,$K[0],$X, 3,$D,$E,$T,$A,$B,$C);
-	&BODY_00_15( 0,$K[0],$X, 4,$C,$D,$E,$T,$A,$B);
-	&BODY_00_15( 0,$K[0],$X, 5,$B,$C,$D,$E,$T,$A);
-	&BODY_00_15( 0,$K[0],$X, 6,$A,$B,$C,$D,$E,$T);
-	&BODY_00_15( 0,$K[0],$X, 7,$T,$A,$B,$C,$D,$E);
-	&BODY_00_15( 0,$K[0],$X, 8,$E,$T,$A,$B,$C,$D);
-	&BODY_00_15( 0,$K[0],$X, 9,$D,$E,$T,$A,$B,$C);
-	&BODY_00_15( 0,$K[0],$X,10,$C,$D,$E,$T,$A,$B);
-	&BODY_00_15( 0,$K[0],$X,11,$B,$C,$D,$E,$T,$A);
-	&BODY_00_15( 0,$K[0],$X,12,$A,$B,$C,$D,$E,$T);
-	&BODY_00_15( 0,$K[0],$X,13,$T,$A,$B,$C,$D,$E);
-	&BODY_00_15( 0,$K[0],$X,14,$E,$T,$A,$B,$C,$D);
-	&BODY_00_15( 1,$K[0],$X,15,$D,$E,$T,$A,$B,$C);
-	&BODY_16_19(-1,$K[0],$X,16,$C,$D,$E,$T,$A,$B);
-	&BODY_16_19( 0,$K[0],$X,17,$B,$C,$D,$E,$T,$A);
-	&BODY_16_19( 0,$K[0],$X,18,$A,$B,$C,$D,$E,$T);
-	&BODY_16_19( 1,$K[0],$X,19,$T,$A,$B,$C,$D,$E);
-
-	&BODY_20_39(-1,$K[1],$X,20,$E,$T,$A,$B,$C,$D);
-	&BODY_20_39( 0,$K[1],$X,21,$D,$E,$T,$A,$B,$C);
-	&BODY_20_39( 0,$K[1],$X,22,$C,$D,$E,$T,$A,$B);
-	&BODY_20_39( 0,$K[1],$X,23,$B,$C,$D,$E,$T,$A);
-	&BODY_20_39( 0,$K[1],$X,24,$A,$B,$C,$D,$E,$T);
-	&BODY_20_39( 0,$K[1],$X,25,$T,$A,$B,$C,$D,$E);
-	&BODY_20_39( 0,$K[1],$X,26,$E,$T,$A,$B,$C,$D);
-	&BODY_20_39( 0,$K[1],$X,27,$D,$E,$T,$A,$B,$C);
-	&BODY_20_39( 0,$K[1],$X,28,$C,$D,$E,$T,$A,$B);
-	&BODY_20_39( 0,$K[1],$X,29,$B,$C,$D,$E,$T,$A);
-	&BODY_20_39( 0,$K[1],$X,30,$A,$B,$C,$D,$E,$T);
-	&BODY_20_39( 0,$K[1],$X,31,$T,$A,$B,$C,$D,$E);
-	&BODY_20_39( 0,$K[1],$X,32,$E,$T,$A,$B,$C,$D);
-	&BODY_20_39( 0,$K[1],$X,33,$D,$E,$T,$A,$B,$C);
-	&BODY_20_39( 0,$K[1],$X,34,$C,$D,$E,$T,$A,$B);
-	&BODY_20_39( 0,$K[1],$X,35,$B,$C,$D,$E,$T,$A);
-	&BODY_20_39( 0,$K[1],$X,36,$A,$B,$C,$D,$E,$T);
-	&BODY_20_39( 0,$K[1],$X,37,$T,$A,$B,$C,$D,$E);
-	&BODY_20_39( 0,$K[1],$X,38,$E,$T,$A,$B,$C,$D);
-	&BODY_20_39( 1,$K[1],$X,39,$D,$E,$T,$A,$B,$C);
-
-	&BODY_40_59(-1,$K[2],$X,40,$C,$D,$E,$T,$A,$B);
-	&BODY_40_59( 0,$K[2],$X,41,$B,$C,$D,$E,$T,$A);
-	&BODY_40_59( 0,$K[2],$X,42,$A,$B,$C,$D,$E,$T);
-	&BODY_40_59( 0,$K[2],$X,43,$T,$A,$B,$C,$D,$E);
-	&BODY_40_59( 0,$K[2],$X,44,$E,$T,$A,$B,$C,$D);
-	&BODY_40_59( 0,$K[2],$X,45,$D,$E,$T,$A,$B,$C);
-	&BODY_40_59( 0,$K[2],$X,46,$C,$D,$E,$T,$A,$B);
-	&BODY_40_59( 0,$K[2],$X,47,$B,$C,$D,$E,$T,$A);
-	&BODY_40_59( 0,$K[2],$X,48,$A,$B,$C,$D,$E,$T);
-	&BODY_40_59( 0,$K[2],$X,49,$T,$A,$B,$C,$D,$E);
-	&BODY_40_59( 0,$K[2],$X,50,$E,$T,$A,$B,$C,$D);
-	&BODY_40_59( 0,$K[2],$X,51,$D,$E,$T,$A,$B,$C);
-	&BODY_40_59( 0,$K[2],$X,52,$C,$D,$E,$T,$A,$B);
-	&BODY_40_59( 0,$K[2],$X,53,$B,$C,$D,$E,$T,$A);
-	&BODY_40_59( 0,$K[2],$X,54,$A,$B,$C,$D,$E,$T);
-	&BODY_40_59( 0,$K[2],$X,55,$T,$A,$B,$C,$D,$E);
-	&BODY_40_59( 0,$K[2],$X,56,$E,$T,$A,$B,$C,$D);
-	&BODY_40_59( 0,$K[2],$X,57,$D,$E,$T,$A,$B,$C);
-	&BODY_40_59( 0,$K[2],$X,58,$C,$D,$E,$T,$A,$B);
-	&BODY_40_59( 1,$K[2],$X,59,$B,$C,$D,$E,$T,$A);
-
-	&BODY_60_79(-1,$K[3],$X,60,$A,$B,$C,$D,$E,$T);
-	&BODY_60_79( 0,$K[3],$X,61,$T,$A,$B,$C,$D,$E);
-	&BODY_60_79( 0,$K[3],$X,62,$E,$T,$A,$B,$C,$D);
-	&BODY_60_79( 0,$K[3],$X,63,$D,$E,$T,$A,$B,$C);
-	&BODY_60_79( 0,$K[3],$X,64,$C,$D,$E,$T,$A,$B);
-	&BODY_60_79( 0,$K[3],$X,65,$B,$C,$D,$E,$T,$A);
-	&BODY_60_79( 0,$K[3],$X,66,$A,$B,$C,$D,$E,$T);
-	&BODY_60_79( 0,$K[3],$X,67,$T,$A,$B,$C,$D,$E);
-	&BODY_60_79( 0,$K[3],$X,68,$E,$T,$A,$B,$C,$D);
-	&BODY_60_79( 0,$K[3],$X,69,$D,$E,$T,$A,$B,$C);
-	&BODY_60_79( 0,$K[3],$X,70,$C,$D,$E,$T,$A,$B);
-	&BODY_60_79( 0,$K[3],$X,71,$B,$C,$D,$E,$T,$A);
-	&BODY_60_79( 0,$K[3],$X,72,$A,$B,$C,$D,$E,$T);
-	&BODY_60_79( 0,$K[3],$X,73,$T,$A,$B,$C,$D,$E);
-	&BODY_60_79( 0,$K[3],$X,74,$E,$T,$A,$B,$C,$D);
-	&BODY_60_79( 0,$K[3],$X,75,$D,$E,$T,$A,$B,$C);
-	&BODY_60_79( 0,$K[3],$X,76,$C,$D,$E,$T,$A,$B);
-	&BODY_60_79( 0,$K[3],$X,77,$B,$C,$D,$E,$T,$A);
-	&BODY_60_79( 0,$K[3],$X,78,$A,$B,$C,$D,$E,$T);
-	&BODY_60_79( 2,$K[3],$X,79,$T,$A,$B,$C,$D,$E);
-
-	&comment("End processing");
-	&comment("");
-	# D is the tmp value
-
-	# E -> A
-	# T -> B
-	# A -> C
-	# B -> D
-	# C -> E
-	# D -> T
-
-	&mov($tmp1,&wparam(0));
-
-	 &mov($D,	&DWP(12,$tmp1,"",0));
-	&add($D,$B);
-	 &mov($B,	&DWP( 4,$tmp1,"",0));
-	&add($B,$T);
-	 &mov($T,	$A);
-	&mov($A,	&DWP( 0,$tmp1,"",0));
-	 &mov(&DWP(12,$tmp1,"",0),$D);
-
-	&add($A,$E);
-	 &mov($E,	&DWP(16,$tmp1,"",0));
-	&add($E,$C);
-	 &mov($C,	&DWP( 8,$tmp1,"",0));
-	&add($C,$T);
-
-	 &mov(&DWP( 0,$tmp1,"",0),$A);
-	&mov("esi",&wparam(1));
-	 &mov(&DWP( 8,$tmp1,"",0),$C);
- 	&add("esi",64);
-	 &mov("eax",&swtmp(17));
-	&mov(&DWP(16,$tmp1,"",0),$E);
-	 &cmp("esi","eax");
-	&mov(&DWP( 4,$tmp1,"",0),$B);
-	 &jb(&label("start"));
-
-	&stack_pop(18+9);
-	 &pop("edi");
-	&pop("ebx");
-	 &pop("ebp");
-	&pop("esi");
-	 &ret();
-
-	# keep a note of shortcut label so it can be used outside
-	# block.
-	my $sclabel = &label("shortcut");
-
-	&function_end_B($name);
-	# Putting this here avoids problems with MASM in debugging mode
-	&sha1_block_host("sha1_block_asm_host_order", $sclabel);
-	}
 
+&asm_finish();

crypto/sha/asm/sha1-ia64.pl

@@ -2,8 +2,9 @@
 #
 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
-# project. Rights for redistribution and usage in source and binary
+# project. The module is, however, dual licensed under OpenSSL and
-# forms are granted according to the OpenSSL license.
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
 # ====================================================================
 #
 # Eternal question is what's wrong with compiler generated code? The
@@ -11,15 +12,10 @@
 # to perform rotations by maintaining copy of 32-bit value in upper
 # bits of 64-bit register. Just follow mux2 and shrp instructions...
 # Performance under big-endian OS such as HP-UX is 179MBps*1GHz, which
-# is >50% better than HP C and >2x better than gcc. As of this moment
+# is >50% better than HP C and >2x better than gcc.
-# performance under little-endian OS such as Linux and Windows will be
-# a bit lower, because data has to be picked in reverse byte-order.
-# It's possible to resolve this issue by implementing third function,
-# sha1_block_asm_data_order_aligned, which would temporarily flip
-# BE field in User Mask register...
 
 $code=<<___;
-.ident  \"sha1-ia64.s, version 1.0\"
+.ident  \"sha1-ia64.s, version 1.2\"
 .ident  \"IA-64 ISA artwork by Andy Polyakov <appro\@fy.chalmers.se>\"
 .explicit
 
@@ -55,63 +51,55 @@ else {
 
 sub BODY_00_15 {
 local	*code=shift;
-local	($i,$a,$b,$c,$d,$e,$f,$unaligned)=@_;
+local	($i,$a,$b,$c,$d,$e,$f)=@_;
 
-if ($unaligned) {
+$code.=<<___ if ($i==0);
-	$code.=<<___;
+{ .mmi;	ld1	$X[$i&0xf]=[inp],2	    // MSB
-{ .mmi;	ld1	tmp0=[inp],2		    // MSB
+	ld1	tmp2=[tmp3],2		};;
-	ld1	tmp1=[tmp3],2		};;
+{ .mmi;	ld1	tmp0=[inp],2
-{ .mmi;	ld1	tmp2=[inp],2
+	ld1	tmp4=[tmp3],2		    // LSB
-	ld1	$X[$i&0xf]=[tmp3],2	    // LSB
+	dep	$X[$i&0xf]=$X[$i&0xf],tmp2,8,8	};;
-	dep	tmp1=tmp0,tmp1,8,8	};;
-{ .mii;	cmp.ne	p16,p0=r0,r0		    // no misaligned prefetch
-	dep	$X[$i&0xf]=tmp2,$X[$i&0xf],8,8;;
-	dep	$X[$i&0xf]=tmp1,$X[$i&0xf],16,16	};;
-{ .mmi;	nop.m	0
 ___
-	}
-elsif ($i<15) {
-	$code.=<<___;
-{ .mmi;	ld4	$X[($i+1)&0xf]=[inp],4	// prefetch
-___
-	}
-else	{
-	$code.=<<___;
-{ .mmi;	nop.m	0
-___
-	}
 if ($i<15) {
 	$code.=<<___;
-	and	tmp0=$c,$b
+{ .mmi;	ld1	$X[($i+1)&0xf]=[inp],2	    // +1
-	dep.z	tmp5=$a,5,27		}   // a<<5
+	dep	tmp1=tmp0,tmp4,8,8	};;
+{ .mmi;	ld1	tmp2=[tmp3],2		    // +1
+	and	tmp4=$c,$b
+	dep	$X[$i&0xf]=$X[$i&0xf],tmp1,16,16	} //;;
 { .mmi;	andcm	tmp1=$d,$b
-	add	tmp4=$e,$K_00_19	};;
+	add	tmp0=$e,$K_00_19
-{ .mmi;	or	tmp0=tmp0,tmp1		    // F_00_19(b,c,d)=(b&c)|(~b&d)
+	dep.z	tmp5=$a,5,27		};; // a<<5
-	add	$f=tmp4,$X[$i&0xf]	    // f=xi+e+K_00_19
+{ .mmi;	or	tmp4=tmp4,tmp1		    // F_00_19(b,c,d)=(b&c)|(~b&d)
+	add	$f=tmp0,$X[$i&0xf]	    // f=xi+e+K_00_19
 	extr.u	tmp1=$a,27,5		};; // a>>27
-{ .mib;	add	$f=$f,tmp0		    // f+=F_00_19(b,c,d)
+{ .mmi;	ld1	tmp0=[inp],2		    // +1
+	add	$f=$f,tmp4		    // f+=F_00_19(b,c,d)
 	shrp	$b=tmp6,tmp6,2		}   // b=ROTATE(b,30)
-{ .mib;	or	tmp1=tmp1,tmp5		    // ROTATE(a,5)
+{ .mmi;	ld1	tmp4=[tmp3],2		    // +1
+	or	tmp5=tmp1,tmp5		    // ROTATE(a,5)
 	mux2	tmp6=$a,0x44		};; // see b in next iteration
-{ .mii;	add	$f=$f,tmp1		    // f+=ROTATE(a,5)
+{ .mii;	add	$f=$f,tmp5		    // f+=ROTATE(a,5)
-	mux2	$X[$i&0xf]=$X[$i&0xf],0x44
+	dep	$X[($i+1)&0xf]=$X[($i+1)&0xf],tmp2,8,8	// +1
-	nop.i	0			};;
+	mux2	$X[$i&0xf]=$X[$i&0xf],0x44	} //;;
 
 ___
 	}
 else	{
 	$code.=<<___;
-	and	tmp0=$c,$b
+{ .mii;	and	tmp3=$c,$b
-	dep.z	tmp5=$a,5,27		}   // a<<5 ;;?
+	dep	tmp1=tmp0,tmp4,8,8;;
+	dep	$X[$i&0xf]=$X[$i&0xf],tmp1,16,16	} //;;
 { .mmi;	andcm	tmp1=$d,$b
-	add	tmp4=$e,$K_00_19	};;
+	add	tmp0=$e,$K_00_19
-{ .mmi;	or	tmp0=tmp0,tmp1		    // F_00_19(b,c,d)=(b&c)|(~b&d)
+	dep.z	tmp5=$a,5,27		};; // a<<5
-	add	$f=tmp4,$X[$i&0xf]	    // f=xi+e+K_00_19
+{ .mmi;	or	tmp4=tmp3,tmp1		    // F_00_19(b,c,d)=(b&c)|(~b&d)
+	add	$f=tmp0,$X[$i&0xf]	    // f=xi+e+K_00_19
 	extr.u	tmp1=$a,27,5		}   // a>>27
 { .mmi;	xor	tmp2=$X[($i+0+1)&0xf],$X[($i+2+1)&0xf]	// +1
 	xor	tmp3=$X[($i+8+1)&0xf],$X[($i+13+1)&0xf] // +1
 	nop.i	0			};;
-{ .mmi;	add	$f=$f,tmp0		    // f+=F_00_19(b,c,d)
+{ .mmi;	add	$f=$f,tmp4		    // f+=F_00_19(b,c,d)
 	xor	tmp2=tmp2,tmp3		    // +1
 	shrp	$b=tmp6,tmp6,2		}   // b=ROTATE(b,30)
 { .mmi; or	tmp1=tmp1,tmp5		    // ROTATE(a,5)
@@ -190,9 +178,7 @@ $code.=<<___;
 	extr.u	tmp1=$a,27,5		}   // a>>27
 { .mib;	add	$f=$f,tmp4		    // f+=e+K_20_39
 	add	$h1=$h1,$a		};; // wrap up
-{ .mmi;
+{ .mmi;	add	$f=$f,tmp0		    // f+=F_20_39(b,c,d)
-(p16)	ld4.s	$X[0]=[inp],4		    // non-faulting prefetch
-	add	$f=$f,tmp0		    // f+=F_20_39(b,c,d)
 	shrp	$b=tmp6,tmp6,2		}   // b=ROTATE(b,30) ;;?
 { .mmi;	or	tmp1=tmp1,tmp5		    // ROTATE(a,5)
 	add	$h3=$h3,$c		};; // wrap up
@@ -245,172 +231,15 @@ tmp3=r11;
 ctx=r32;	// in0
 inp=r33;	// in1
 
-// void sha1_block_asm_host_order(SHA_CTX *c,const void *p,size_t num);
+// void sha1_block_data_order(SHA_CTX *c,const void *p,size_t num);
-.global	sha1_block_asm_host_order#
+.global	sha1_block_data_order#
-.proc	sha1_block_asm_host_order#
+.proc	sha1_block_data_order#
 .align	32
-sha1_block_asm_host_order:
+sha1_block_data_order:
 	.prologue
-	.fframe	0
-	.save	ar.pfs,r0
-	.save	ar.lc,r3
 { .mmi;	alloc	tmp1=ar.pfs,3,15,0,0
 	$ADDP	tmp0=4,ctx
-	mov	r3=ar.lc		}
-{ .mmi;	$ADDP	ctx=0,ctx
-	$ADDP	inp=0,inp
-	mov	r2=pr			};;
-tmp4=in2;
-tmp5=loc13;
-tmp6=loc14;
-	.body
-{ .mlx;	ld4	$h0=[ctx],8
-	movl	$K_00_19=0x5a827999	}
-{ .mlx;	ld4	$h1=[tmp0],8
-	movl	$K_20_39=0x6ed9eba1	};;
-{ .mlx;	ld4	$h2=[ctx],8
-	movl	$K_40_59=0x8f1bbcdc	}
-{ .mlx;	ld4	$h3=[tmp0]
-	movl	$K_60_79=0xca62c1d6	};;
-{ .mmi;	ld4	$h4=[ctx],-16
-	add	in2=-1,in2		    // adjust num for ar.lc
-	mov	ar.ec=1			};;
-{ .mmi;	ld4	$X[0]=[inp],4		    // prefetch
-	cmp.ne	p16,p0=r0,in2		    // prefecth at loop end
-	mov	ar.lc=in2		};; // brp.loop.imp: too far
-
-.Lhtop:
-{ .mmi;	mov	$A=$h0
-	mov	$B=$h1
-	mux2	tmp6=$h1,0x44		}
-{ .mmi;	mov	$C=$h2
-	mov	$D=$h3
-	mov	$E=$h4			};;
-
-___
-
-	&BODY_00_15(\$code, 0,$A,$B,$C,$D,$E,$T);
-	&BODY_00_15(\$code, 1,$T,$A,$B,$C,$D,$E);
-	&BODY_00_15(\$code, 2,$E,$T,$A,$B,$C,$D);
-	&BODY_00_15(\$code, 3,$D,$E,$T,$A,$B,$C);
-	&BODY_00_15(\$code, 4,$C,$D,$E,$T,$A,$B);
-	&BODY_00_15(\$code, 5,$B,$C,$D,$E,$T,$A);
-	&BODY_00_15(\$code, 6,$A,$B,$C,$D,$E,$T);
-	&BODY_00_15(\$code, 7,$T,$A,$B,$C,$D,$E);
-	&BODY_00_15(\$code, 8,$E,$T,$A,$B,$C,$D);
-	&BODY_00_15(\$code, 9,$D,$E,$T,$A,$B,$C);
-	&BODY_00_15(\$code,10,$C,$D,$E,$T,$A,$B);
-	&BODY_00_15(\$code,11,$B,$C,$D,$E,$T,$A);
-	&BODY_00_15(\$code,12,$A,$B,$C,$D,$E,$T);
-	&BODY_00_15(\$code,13,$T,$A,$B,$C,$D,$E);
-	&BODY_00_15(\$code,14,$E,$T,$A,$B,$C,$D);
-	&BODY_00_15(\$code,15,$D,$E,$T,$A,$B,$C);
-
-	&BODY_16_19(\$code,16,$C,$D,$E,$T,$A,$B);
-	&BODY_16_19(\$code,17,$B,$C,$D,$E,$T,$A);
-	&BODY_16_19(\$code,18,$A,$B,$C,$D,$E,$T);
-	&BODY_16_19(\$code,19,$T,$A,$B,$C,$D,$E);
-
-	&BODY_20_39(\$code,20,$E,$T,$A,$B,$C,$D);
-	&BODY_20_39(\$code,21,$D,$E,$T,$A,$B,$C);
-	&BODY_20_39(\$code,22,$C,$D,$E,$T,$A,$B);
-	&BODY_20_39(\$code,23,$B,$C,$D,$E,$T,$A);
-	&BODY_20_39(\$code,24,$A,$B,$C,$D,$E,$T);
-	&BODY_20_39(\$code,25,$T,$A,$B,$C,$D,$E);
-	&BODY_20_39(\$code,26,$E,$T,$A,$B,$C,$D);
-	&BODY_20_39(\$code,27,$D,$E,$T,$A,$B,$C);
-	&BODY_20_39(\$code,28,$C,$D,$E,$T,$A,$B);
-	&BODY_20_39(\$code,29,$B,$C,$D,$E,$T,$A);
-	&BODY_20_39(\$code,30,$A,$B,$C,$D,$E,$T);
-	&BODY_20_39(\$code,31,$T,$A,$B,$C,$D,$E);
-	&BODY_20_39(\$code,32,$E,$T,$A,$B,$C,$D);
-	&BODY_20_39(\$code,33,$D,$E,$T,$A,$B,$C);
-	&BODY_20_39(\$code,34,$C,$D,$E,$T,$A,$B);
-	&BODY_20_39(\$code,35,$B,$C,$D,$E,$T,$A);
-	&BODY_20_39(\$code,36,$A,$B,$C,$D,$E,$T);
-	&BODY_20_39(\$code,37,$T,$A,$B,$C,$D,$E);
-	&BODY_20_39(\$code,38,$E,$T,$A,$B,$C,$D);
-	&BODY_20_39(\$code,39,$D,$E,$T,$A,$B,$C);
-
-	&BODY_40_59(\$code,40,$C,$D,$E,$T,$A,$B);
-	&BODY_40_59(\$code,41,$B,$C,$D,$E,$T,$A);
-	&BODY_40_59(\$code,42,$A,$B,$C,$D,$E,$T);
-	&BODY_40_59(\$code,43,$T,$A,$B,$C,$D,$E);
-	&BODY_40_59(\$code,44,$E,$T,$A,$B,$C,$D);
-	&BODY_40_59(\$code,45,$D,$E,$T,$A,$B,$C);
-	&BODY_40_59(\$code,46,$C,$D,$E,$T,$A,$B);
-	&BODY_40_59(\$code,47,$B,$C,$D,$E,$T,$A);
-	&BODY_40_59(\$code,48,$A,$B,$C,$D,$E,$T);
-	&BODY_40_59(\$code,49,$T,$A,$B,$C,$D,$E);
-	&BODY_40_59(\$code,50,$E,$T,$A,$B,$C,$D);
-	&BODY_40_59(\$code,51,$D,$E,$T,$A,$B,$C);
-	&BODY_40_59(\$code,52,$C,$D,$E,$T,$A,$B);
-	&BODY_40_59(\$code,53,$B,$C,$D,$E,$T,$A);
-	&BODY_40_59(\$code,54,$A,$B,$C,$D,$E,$T);
-	&BODY_40_59(\$code,55,$T,$A,$B,$C,$D,$E);
-	&BODY_40_59(\$code,56,$E,$T,$A,$B,$C,$D);
-	&BODY_40_59(\$code,57,$D,$E,$T,$A,$B,$C);
-	&BODY_40_59(\$code,58,$C,$D,$E,$T,$A,$B);
-	&BODY_40_59(\$code,59,$B,$C,$D,$E,$T,$A);
-
-	&BODY_60_79(\$code,60,$A,$B,$C,$D,$E,$T);
-	&BODY_60_79(\$code,61,$T,$A,$B,$C,$D,$E);
-	&BODY_60_79(\$code,62,$E,$T,$A,$B,$C,$D);
-	&BODY_60_79(\$code,63,$D,$E,$T,$A,$B,$C);
-	&BODY_60_79(\$code,64,$C,$D,$E,$T,$A,$B);
-	&BODY_60_79(\$code,65,$B,$C,$D,$E,$T,$A);
-	&BODY_60_79(\$code,66,$A,$B,$C,$D,$E,$T);
-	&BODY_60_79(\$code,67,$T,$A,$B,$C,$D,$E);
-	&BODY_60_79(\$code,68,$E,$T,$A,$B,$C,$D);
-	&BODY_60_79(\$code,69,$D,$E,$T,$A,$B,$C);
-	&BODY_60_79(\$code,70,$C,$D,$E,$T,$A,$B);
-	&BODY_60_79(\$code,71,$B,$C,$D,$E,$T,$A);
-	&BODY_60_79(\$code,72,$A,$B,$C,$D,$E,$T);
-	&BODY_60_79(\$code,73,$T,$A,$B,$C,$D,$E);
-	&BODY_60_79(\$code,74,$E,$T,$A,$B,$C,$D);
-	&BODY_60_79(\$code,75,$D,$E,$T,$A,$B,$C);
-	&BODY_60_79(\$code,76,$C,$D,$E,$T,$A,$B);
-	&BODY_60_79(\$code,77,$B,$C,$D,$E,$T,$A);
-	&BODY_60_79(\$code,78,$A,$B,$C,$D,$E,$T);
-	&BODY_60_79(\$code,79,$T,$A,$B,$C,$D,$E);
-
-$code.=<<___;
-{ .mmb;	add	$h0=$h0,$E
-	nop.m	0
-	br.ctop.dptk.many	.Lhtop	};;
-.Lhend:
-{ .mmi;	add	tmp0=4,ctx
-	mov	ar.lc=r3		};;
-{ .mmi;	st4	[ctx]=$h0,8
-	st4	[tmp0]=$h1,8		};;
-{ .mmi;	st4	[ctx]=$h2,8
-	st4	[tmp0]=$h3		};;
-{ .mib;	st4	[ctx]=$h4,-16
-	mov	pr=r2,0x1ffff
-	br.ret.sptk.many	b0	};;
-.endp	sha1_block_asm_host_order#
-___
-
-
-$code.=<<___;
-// void sha1_block_asm_data_order(SHA_CTX *c,const void *p,size_t num);
-.global	sha1_block_asm_data_order#
-.proc	sha1_block_asm_data_order#
-.align	32
-sha1_block_asm_data_order:
-___
-$code.=<<___ if ($big_endian);
-{ .mmi;	and	r2=3,inp				};;
-{ .mib;	cmp.eq	p6,p0=r0,r2
-(p6)	br.dptk.many	sha1_block_asm_host_order	};;
-___
-$code.=<<___;
-	.prologue
-	.fframe	0
-	.save	ar.pfs,r0
 	.save	ar.lc,r3
-{ .mmi;	alloc	tmp1=ar.pfs,3,15,0,0
-	$ADDP	tmp0=4,ctx
 	mov	r3=ar.lc		}
 { .mmi;	$ADDP	ctx=0,ctx
 	$ADDP	inp=0,inp
@@ -444,90 +273,16 @@ tmp6=loc14;
 
 ___
 
-	&BODY_00_15(\$code, 0,$A,$B,$C,$D,$E,$T,1);
+{ my $i,@V=($A,$B,$C,$D,$E,$T);
-	&BODY_00_15(\$code, 1,$T,$A,$B,$C,$D,$E,1);
-	&BODY_00_15(\$code, 2,$E,$T,$A,$B,$C,$D,1);
-	&BODY_00_15(\$code, 3,$D,$E,$T,$A,$B,$C,1);
-	&BODY_00_15(\$code, 4,$C,$D,$E,$T,$A,$B,1);
-	&BODY_00_15(\$code, 5,$B,$C,$D,$E,$T,$A,1);
-	&BODY_00_15(\$code, 6,$A,$B,$C,$D,$E,$T,1);
-	&BODY_00_15(\$code, 7,$T,$A,$B,$C,$D,$E,1);
-	&BODY_00_15(\$code, 8,$E,$T,$A,$B,$C,$D,1);
-	&BODY_00_15(\$code, 9,$D,$E,$T,$A,$B,$C,1);
-	&BODY_00_15(\$code,10,$C,$D,$E,$T,$A,$B,1);
-	&BODY_00_15(\$code,11,$B,$C,$D,$E,$T,$A,1);
-	&BODY_00_15(\$code,12,$A,$B,$C,$D,$E,$T,1);
-	&BODY_00_15(\$code,13,$T,$A,$B,$C,$D,$E,1);
-	&BODY_00_15(\$code,14,$E,$T,$A,$B,$C,$D,1);
-	&BODY_00_15(\$code,15,$D,$E,$T,$A,$B,$C,1);
 
-	&BODY_16_19(\$code,16,$C,$D,$E,$T,$A,$B);
+	for($i=0;$i<16;$i++)	{ &BODY_00_15(\$code,$i,@V); unshift(@V,pop(@V)); }
-	&BODY_16_19(\$code,17,$B,$C,$D,$E,$T,$A);
+	for(;$i<20;$i++)	{ &BODY_16_19(\$code,$i,@V); unshift(@V,pop(@V)); }
-	&BODY_16_19(\$code,18,$A,$B,$C,$D,$E,$T);
+	for(;$i<40;$i++)	{ &BODY_20_39(\$code,$i,@V); unshift(@V,pop(@V)); }
-	&BODY_16_19(\$code,19,$T,$A,$B,$C,$D,$E);
+	for(;$i<60;$i++)	{ &BODY_40_59(\$code,$i,@V); unshift(@V,pop(@V)); }
+	for(;$i<80;$i++)	{ &BODY_60_79(\$code,$i,@V); unshift(@V,pop(@V)); }
 
-	&BODY_20_39(\$code,20,$E,$T,$A,$B,$C,$D);
+	(($V[5] eq $D) and ($V[0] eq $E)) or die;	# double-check
-	&BODY_20_39(\$code,21,$D,$E,$T,$A,$B,$C);
+}
-	&BODY_20_39(\$code,22,$C,$D,$E,$T,$A,$B);
-	&BODY_20_39(\$code,23,$B,$C,$D,$E,$T,$A);
-	&BODY_20_39(\$code,24,$A,$B,$C,$D,$E,$T);
-	&BODY_20_39(\$code,25,$T,$A,$B,$C,$D,$E);
-	&BODY_20_39(\$code,26,$E,$T,$A,$B,$C,$D);
-	&BODY_20_39(\$code,27,$D,$E,$T,$A,$B,$C);
-	&BODY_20_39(\$code,28,$C,$D,$E,$T,$A,$B);
-	&BODY_20_39(\$code,29,$B,$C,$D,$E,$T,$A);
-	&BODY_20_39(\$code,30,$A,$B,$C,$D,$E,$T);
-	&BODY_20_39(\$code,31,$T,$A,$B,$C,$D,$E);
-	&BODY_20_39(\$code,32,$E,$T,$A,$B,$C,$D);
-	&BODY_20_39(\$code,33,$D,$E,$T,$A,$B,$C);
-	&BODY_20_39(\$code,34,$C,$D,$E,$T,$A,$B);
-	&BODY_20_39(\$code,35,$B,$C,$D,$E,$T,$A);
-	&BODY_20_39(\$code,36,$A,$B,$C,$D,$E,$T);
-	&BODY_20_39(\$code,37,$T,$A,$B,$C,$D,$E);
-	&BODY_20_39(\$code,38,$E,$T,$A,$B,$C,$D);
-	&BODY_20_39(\$code,39,$D,$E,$T,$A,$B,$C);
-
-	&BODY_40_59(\$code,40,$C,$D,$E,$T,$A,$B);
-	&BODY_40_59(\$code,41,$B,$C,$D,$E,$T,$A);
-	&BODY_40_59(\$code,42,$A,$B,$C,$D,$E,$T);
-	&BODY_40_59(\$code,43,$T,$A,$B,$C,$D,$E);
-	&BODY_40_59(\$code,44,$E,$T,$A,$B,$C,$D);
-	&BODY_40_59(\$code,45,$D,$E,$T,$A,$B,$C);
-	&BODY_40_59(\$code,46,$C,$D,$E,$T,$A,$B);
-	&BODY_40_59(\$code,47,$B,$C,$D,$E,$T,$A);
-	&BODY_40_59(\$code,48,$A,$B,$C,$D,$E,$T);
-	&BODY_40_59(\$code,49,$T,$A,$B,$C,$D,$E);
-	&BODY_40_59(\$code,50,$E,$T,$A,$B,$C,$D);
-	&BODY_40_59(\$code,51,$D,$E,$T,$A,$B,$C);
-	&BODY_40_59(\$code,52,$C,$D,$E,$T,$A,$B);
-	&BODY_40_59(\$code,53,$B,$C,$D,$E,$T,$A);
-	&BODY_40_59(\$code,54,$A,$B,$C,$D,$E,$T);
-	&BODY_40_59(\$code,55,$T,$A,$B,$C,$D,$E);
-	&BODY_40_59(\$code,56,$E,$T,$A,$B,$C,$D);
-	&BODY_40_59(\$code,57,$D,$E,$T,$A,$B,$C);
-	&BODY_40_59(\$code,58,$C,$D,$E,$T,$A,$B);
-	&BODY_40_59(\$code,59,$B,$C,$D,$E,$T,$A);
-
-	&BODY_60_79(\$code,60,$A,$B,$C,$D,$E,$T);
-	&BODY_60_79(\$code,61,$T,$A,$B,$C,$D,$E);
-	&BODY_60_79(\$code,62,$E,$T,$A,$B,$C,$D);
-	&BODY_60_79(\$code,63,$D,$E,$T,$A,$B,$C);
-	&BODY_60_79(\$code,64,$C,$D,$E,$T,$A,$B);
-	&BODY_60_79(\$code,65,$B,$C,$D,$E,$T,$A);
-	&BODY_60_79(\$code,66,$A,$B,$C,$D,$E,$T);
-	&BODY_60_79(\$code,67,$T,$A,$B,$C,$D,$E);
-	&BODY_60_79(\$code,68,$E,$T,$A,$B,$C,$D);
-	&BODY_60_79(\$code,69,$D,$E,$T,$A,$B,$C);
-	&BODY_60_79(\$code,70,$C,$D,$E,$T,$A,$B);
-	&BODY_60_79(\$code,71,$B,$C,$D,$E,$T,$A);
-	&BODY_60_79(\$code,72,$A,$B,$C,$D,$E,$T);
-	&BODY_60_79(\$code,73,$T,$A,$B,$C,$D,$E);
-	&BODY_60_79(\$code,74,$E,$T,$A,$B,$C,$D);
-	&BODY_60_79(\$code,75,$D,$E,$T,$A,$B,$C);
-	&BODY_60_79(\$code,76,$C,$D,$E,$T,$A,$B);
-	&BODY_60_79(\$code,77,$B,$C,$D,$E,$T,$A);
-	&BODY_60_79(\$code,78,$A,$B,$C,$D,$E,$T);
-	&BODY_60_79(\$code,79,$T,$A,$B,$C,$D,$E);
 
 $code.=<<___;
 { .mmb;	add	$h0=$h0,$E
@@ -543,7 +298,8 @@ $code.=<<___;
 { .mib;	st4	[ctx]=$h4,-16
 	mov	pr=r2,0x1ffff
 	br.ret.sptk.many	b0	};;
-.endp	sha1_block_asm_data_order#
+.endp	sha1_block_data_order#
+stringz	"SHA1 block transform for IA64, CRYPTOGAMS by <appro\@openssl.org>"
 ___
 
 print $code;

crypto/sha/asm/sha1-s390x.pl

@@ -0,0 +1,223 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA1 block procedure for s390x.
+
+# April 2007.
+#
+# Performance is >30% better than gcc 3.3 generated code. But the real
+# twist is that SHA1 hardware support is detected and utilized. In
+# which case performance can reach further >4.5x for larger chunks.
+
+$kimdfunc=0;	# magic function code for kimd instruction,
+		# 0 disables hardware support
+
+$output=shift;
+open STDOUT,">$output";
+
+$t0="%r0";
+$t1="%r1";
+$ctx="%r2";
+$inp="%r3";
+$len="%r4";
+
+$A="%r5";
+$B="%r6";
+$C="%r7";
+$D="%r8";
+$E="%r9";	@V=($A,$B,$C,$D,$E);
+$K_00_19="%r10";
+$K_20_39="%r11";
+$K_40_59="%r12";
+$K_60_79="%r13";
+$Xi="%r14";
+$sp="%r15";
+
+$frame=160+16*4;
+
+sub BODY_00_15 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $xi=($i&1)?$Xi:$t1;
+
+$code.=<<___ if ($i<16 && !($i&1));
+	lg	$Xi,`$i*4`($inp)
+___
+$code.=<<___;
+	alr	$e,$K_00_19	### $i
+	rll	$t0,$a,5
+	alr	$e,$t0
+	lr	$t0,$d
+	xr	$t0,$c
+	nr	$t0,$b
+	xr	$t0,$d
+	alr	$e,$t0
+	rll	$b,$b,30
+___
+$code.=<<___ if ($i<16 && !($i&1));
+	srlg	$xi,$Xi,32
+	stg	$Xi,`160+$i*4`($sp)
+___
+$code.=<<___;
+	alr	$e,$xi
+___
+}
+
+sub Xupdate {
+my $i=shift;
+
+return if ($i&1);	# Xupdate is vectorized and executed every 2nd cycle
+$code.=<<___;
+	lg	$Xi,`160+4*($i%16)`($sp)	### Xupdate($i)
+	xg	$Xi,`160+4*(($i+2)%16)`($sp)
+	xg	$Xi,`160+4*(($i+8)%16)`($sp)
+___
+if ((($i+13)%16)==15) {
+$code.=<<___;
+	llgf	$t0,`160+4*15`($sp)
+	x	$Xi,`160+0`($sp)
+	sllg	$t0,$t0,32
+	xgr	$Xi,$t0
+___
+} else {
+$code.=<<___;
+	xg	$Xi,`160+4*(($i+13)%16)`($sp)
+___
+}
+$code.=<<___;
+	rll	$Xi,$Xi,1
+	rllg	$t1,$Xi,32
+	rll	$t1,$t1,1
+	rllg	$Xi,$t1,32
+	stg	$Xi,`160+4*($i%16)`($sp)
+___
+}
+
+sub BODY_16_19 {
+	&Xupdate(@_[0]);
+	&BODY_00_15(@_);
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $xi=($i&1)?$Xi:$t1;
+my $K_XX_XX=($i<40)?$K_20_39:$K_60_79;
+
+	&Xupdate($i);
+$code.=<<___;
+	alr	$e,$K_XX_XX	### $i
+	rll	$t0,$a,5
+	alr	$e,$t0
+	lr	$t0,$b
+	xr	$t0,$c
+	xr	$t0,$d
+	alr	$e,$t0
+	rll	$b,$b,30
+	alr	$e,$xi
+___
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e)=@_;
+my $xi=($i&1)?$Xi:$t1;
+
+	&Xupdate($i);
+$code.=<<___;
+	alr	$e,$K_40_59	### $i
+	rll	$t0,$a,5
+	alr	$e,$t0
+	lr	$t0,$b
+	or	$t0,$c
+	nr	$t0,$d
+	alr	$e,$xi
+	lr	$t1,$b
+	nr	$t1,$c
+	or	$t0,$t1
+	alr	$e,$t0
+	rll	$b,$b,30
+___
+}
+
+$code.=<<___;
+.text
+.globl	sha1_block_data_order
+.type	sha1_block_data_order,\@function
+sha1_block_data_order:
+___
+$code.=<<___ if ($kimdfunc);
+	lghi	%r0,0
+	la	%r1,16($sp)
+	.long	0xb93e0002	# kimd %r0,%r2
+	lg	%r0,16($sp)
+	tmhh	%r0,`0x8000>>$kimdfunc`
+	jz	.Lsoftware
+	lghi	%r0,$kimdfunc
+	lgr	%r1,$ctx
+	lgr	%r2,$inp
+	sllg	%r3,$len,6
+	.long	0xb93e0002	# kimd %r0,%r2
+	brc	1,.-4		# pay attention to "partial completion"
+	br	%r14
+.Lsoftware:
+___
+$code.=<<___;
+	stmg	%r6,%r15,48($sp)
+	lgr	%r0,$sp
+	aghi	$sp,-$frame
+	stg	%r0,0($sp)
+
+	sllg	$len,$len,6
+	la	$len,0($inp,$len)
+
+	llgf	$A,0($ctx)
+	llgf	$B,4($ctx)
+	llgf	$C,8($ctx)
+	llgf	$D,12($ctx)
+	llgf	$E,16($ctx)
+
+	llilh	$K_00_19,0x5a82
+	oill	$K_00_19,0x7999
+	llilh	$K_20_39,0x6ed9
+	oill	$K_20_39,0xeba1
+	llilh	$K_40_59,0x8f1b
+	oill	$K_40_59,0xbcdc
+	llilh	$K_60_79,0xca62
+	oill	$K_60_79,0xc1d6
+.Lloop:
+___
+for ($i=0;$i<16;$i++)	{ &BODY_00_15($i,@V); unshift(@V,pop(@V)); }
+for (;$i<20;$i++)	{ &BODY_16_19($i,@V); unshift(@V,pop(@V)); }
+for (;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+for (;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+for (;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+
+	al	$A,0($ctx)
+	al	$B,4($ctx)
+	al	$C,8($ctx)
+	al	$D,12($ctx)
+	al	$E,16($ctx)
+	st	$A,0($ctx)
+	st	$B,4($ctx)
+	st	$C,8($ctx)
+	st	$D,12($ctx)
+	st	$E,16($ctx)
+	la	$inp,64($inp)
+	clgr	$inp,$len
+	jne	.Lloop
+
+	lmg	%r6,%r15,`$frame+48`($sp)
+	br	%r14
+.size	sha1_block_data_order,.-sha1_block_data_order
+.string	"SHA1 block transform for s390x, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+
+print $code;
+close STDOUT;

crypto/sha/asm/sha1-x86_64.pl

@@ -0,0 +1,242 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+#
+# sha1_block procedure for x86_64.
+#
+# It was brought to my attention that on EM64T compiler-generated code
+# was far behind 32-bit assembler implementation. This is unlike on
+# Opteron where compiler-generated code was only 15% behind 32-bit
+# assembler, which originally made it hard to motivate the effort.
+# There was suggestion to mechanically translate 32-bit code, but I
+# dismissed it, reasoning that x86_64 offers enough register bank
+# capacity to fully utilize SHA-1 parallelism. Therefore this fresh
+# implementation:-) However! While 64-bit code does performs better
+# on Opteron, I failed to beat 32-bit assembler on EM64T core. Well,
+# x86_64 does offer larger *addressable* bank, but out-of-order core
+# reaches for even more registers through dynamic aliasing, and EM64T
+# core must have managed to run-time optimize even 32-bit code just as
+# good as 64-bit one. Performance improvement is summarized in the
+# following table:
+#
+#		gcc 3.4		32-bit asm	cycles/byte
+# Opteron	+45%		+20%		6.8
+# Xeon P4	+65%		+0%		9.9
+# Core2		+60%		+10%		8.8
+
+$output=shift;
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open STDOUT,"| $^X $xlate $output";
+
+$ctx="%rdi";	# 1st arg
+$inp="%rsi";	# 2nd arg
+$num="%rdx";	# 3rd arg
+
+# reassign arguments in order to produce more compact code
+$ctx="%r8";
+$inp="%r9";
+$num="%r10";
+
+$xi="%eax";
+$t0="%ebx";
+$t1="%ecx";
+$A="%edx";
+$B="%esi";
+$C="%edi";
+$D="%ebp";
+$E="%r11d";
+$T="%r12d";
+
+@V=($A,$B,$C,$D,$E,$T);
+
+sub PROLOGUE {
+my $func=shift;
+$code.=<<___;
+.globl	$func
+.type	$func,\@function,3
+.align	16
+$func:
+	push	%rbx
+	push	%rbp
+	push	%r12
+	mov	%rsp,%rax
+	mov	%rdi,$ctx	# reassigned argument
+	sub	\$`8+16*4`,%rsp
+	mov	%rsi,$inp	# reassigned argument
+	and	\$-64,%rsp
+	mov	%rdx,$num	# reassigned argument
+	mov	%rax,`16*4`(%rsp)
+
+	mov	0($ctx),$A
+	mov	4($ctx),$B
+	mov	8($ctx),$C
+	mov	12($ctx),$D
+	mov	16($ctx),$E
+___
+}
+
+sub EPILOGUE {
+my $func=shift;
+$code.=<<___;
+	mov	`16*4`(%rsp),%rsp
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+	ret
+.size	$func,.-$func
+___
+}
+
+sub BODY_00_19 {
+my ($i,$a,$b,$c,$d,$e,$f,$host)=@_;
+my $j=$i+1;
+$code.=<<___ if ($i==0);
+	mov	`4*$i`($inp),$xi	
+	`"bswap	$xi"	if(!defined($host))`
+	mov	$xi,`4*$i`(%rsp)
+___
+$code.=<<___ if ($i<15);
+	lea	0x5a827999($xi,$e),$f
+	mov	$c,$t0
+	mov	`4*$j`($inp),$xi
+	mov	$a,$e
+	xor	$d,$t0
+	`"bswap	$xi"	if(!defined($host))`	
+	rol	\$5,$e
+	and	$b,$t0
+	mov	$xi,`4*$j`(%rsp)
+	add	$e,$f
+	xor	$d,$t0
+	rol	\$30,$b
+	add	$t0,$f
+___
+$code.=<<___ if ($i>=15);
+	lea	0x5a827999($xi,$e),$f
+	mov	`4*($j%16)`(%rsp),$xi
+	mov	$c,$t0
+	mov	$a,$e
+	xor	`4*(($j+2)%16)`(%rsp),$xi
+	xor	$d,$t0
+	rol	\$5,$e
+	xor	`4*(($j+8)%16)`(%rsp),$xi
+	and	$b,$t0
+	add	$e,$f
+	xor	`4*(($j+13)%16)`(%rsp),$xi
+	xor	$d,$t0
+	rol	\$30,$b
+	add	$t0,$f
+	rol	\$1,$xi
+	mov	$xi,`4*($j%16)`(%rsp)
+___
+}
+
+sub BODY_20_39 {
+my ($i,$a,$b,$c,$d,$e,$f)=@_;
+my $j=$i+1;
+my $K=($i<40)?0x6ed9eba1:0xca62c1d6;
+$code.=<<___ if ($i<79);
+	lea	$K($xi,$e),$f
+	mov	`4*($j%16)`(%rsp),$xi
+	mov	$c,$t0
+	mov	$a,$e
+	xor	`4*(($j+2)%16)`(%rsp),$xi
+	xor	$b,$t0
+	rol	\$5,$e
+	xor	`4*(($j+8)%16)`(%rsp),$xi
+	xor	$d,$t0
+	add	$e,$f
+	xor	`4*(($j+13)%16)`(%rsp),$xi
+	rol	\$30,$b
+	add	$t0,$f
+	rol	\$1,$xi
+___
+$code.=<<___ if ($i<76);
+	mov	$xi,`4*($j%16)`(%rsp)
+___
+$code.=<<___ if ($i==79);
+	lea	$K($xi,$e),$f
+	mov	$c,$t0
+	mov	$a,$e
+	xor	$b,$t0
+	rol	\$5,$e
+	xor	$d,$t0
+	add	$e,$f
+	rol	\$30,$b
+	add	$t0,$f
+___
+}
+
+sub BODY_40_59 {
+my ($i,$a,$b,$c,$d,$e,$f)=@_;
+my $j=$i+1;
+$code.=<<___;
+	lea	0x8f1bbcdc($xi,$e),$f
+	mov	`4*($j%16)`(%rsp),$xi
+	mov	$b,$t0
+	mov	$b,$t1
+	xor	`4*(($j+2)%16)`(%rsp),$xi
+	mov	$a,$e
+	and	$c,$t0
+	xor	`4*(($j+8)%16)`(%rsp),$xi
+	or	$c,$t1
+	rol	\$5,$e
+	xor	`4*(($j+13)%16)`(%rsp),$xi
+	and	$d,$t1
+	add	$e,$f
+	rol	\$1,$xi
+	or	$t1,$t0
+	rol	\$30,$b
+	mov	$xi,`4*($j%16)`(%rsp)
+	add	$t0,$f
+___
+}
+
+$code=".text\n";
+
+&PROLOGUE("sha1_block_data_order");
+$code.=".align	4\n.Lloop:\n";
+for($i=0;$i<20;$i++)	{ &BODY_00_19($i,@V); unshift(@V,pop(@V)); }
+for(;$i<40;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+for(;$i<60;$i++)	{ &BODY_40_59($i,@V); unshift(@V,pop(@V)); }
+for(;$i<80;$i++)	{ &BODY_20_39($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	add	0($ctx),$E
+	add	4($ctx),$T
+	add	8($ctx),$A
+	add	12($ctx),$B
+	add	16($ctx),$C
+	mov	$E,0($ctx)
+	mov	$T,4($ctx)
+	mov	$A,8($ctx)
+	mov	$B,12($ctx)
+	mov	$C,16($ctx)
+
+	xchg	$E,$A	# mov	$E,$A
+	xchg	$T,$B	# mov	$T,$B
+	xchg	$E,$C	# mov	$A,$C
+	xchg	$T,$D	# mov	$B,$D
+			# mov	$C,$E
+	lea	`16*4`($inp),$inp
+	sub	\$1,$num
+	jnz	.Lloop
+___
+&EPILOGUE("sha1_block_data_order");
+$code.=<<___;
+.asciz	"SHA1 block transform for x86_64, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+####################################################################
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;

crypto/sha/asm/sha512-ia64.pl

@@ -2,8 +2,9 @@
 #
 # ====================================================================
 # Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
-# project. Rights for redistribution and usage in source and binary
+# project. The module is, however, dual licensed under OpenSSL and
-# forms are granted according to the OpenSSL license.
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
 # ====================================================================
 #
 # SHA256/512_Transform for Itanium.
@@ -71,7 +72,7 @@ if ($output =~ /512.*\.[s|asm]/) {
 	$ADD="add";
 	$SHRU="shr.u";
 	$TABLE="K512";
-	$func="sha512_block";
+	$func="sha512_block_data_order";
 	@Sigma0=(28,34,39);
 	@Sigma1=(14,18,41);
 	@sigma0=(1,  8, 7);
@@ -85,7 +86,7 @@ if ($output =~ /512.*\.[s|asm]/) {
 	$ADD="padd4";
 	$SHRU="pshr4.u";
 	$TABLE="K256";
-	$func="sha256_block";
+	$func="sha256_block_data_order";
 	@Sigma0=( 2,13,22);
 	@Sigma1=( 6,11,25);
 	@sigma0=( 7,18, 3);
@@ -105,11 +106,13 @@ if (!defined($big_endian))
              {	$big_endian=(unpack('L',pack('N',1))==1);  }
 
 $code=<<___;
-.ident  \"$output, version 1.0\"
+.ident  \"$output, version 1.1\"
 .ident  \"IA-64 ISA artwork by Andy Polyakov <appro\@fy.chalmers.se>\"
 .explicit
 .text
 
+pfssave=r2;
+lcsave=r3;
 prsave=r14;
 K=r15;
 A=r16;	B=r17;	C=r18;	D=r19;
@@ -121,6 +124,8 @@ ctx=r31;	// 1st arg
 input=r48;	// 2nd arg
 num=r49;	// 3rd arg
 sgm0=r50;	sgm1=r51;	// small constants
+A_=r54;	B_=r55;	C_=r56;	D_=r57;
+E_=r58;	F_=r59;	G_=r60;	H_=r61;
 
 // void $func (SHA_CTX *ctx, const void *in,size_t num[,int host])
 .global	$func#
@@ -128,82 +133,319 @@ sgm0=r50;	sgm1=r51;	// small constants
 .align	32
 $func:
 	.prologue
-	.fframe	0
+	.save	ar.pfs,pfssave
-	.save	ar.pfs,r2
+{ .mmi;	alloc	pfssave=ar.pfs,3,27,0,16
-	.save	ar.lc,r3
-	.save	pr,prsave
-{ .mmi;	alloc	r2=ar.pfs,3,17,0,16
 	$ADDP	ctx=0,r32		// 1st arg
-	mov	r3=ar.lc	}
+	.save	ar.lc,lcsave
+	mov	lcsave=ar.lc	}
 { .mmi;	$ADDP	input=0,r33		// 2nd arg
-	addl	Ktbl=\@ltoff($TABLE#),gp
+	mov	num=r34			// 3rd arg
+	.save	pr,prsave
 	mov	prsave=pr	};;
 
 	.body
-{ .mii;	ld8	Ktbl=[Ktbl]
-	mov	num=r34		};;	// 3rd arg
-
 { .mib;	add	r8=0*$SZ,ctx
 	add	r9=1*$SZ,ctx
-	brp.loop.imp	.L_first16,.L_first16_ctop
+	brp.loop.imp	.L_first16,.L_first16_end-16	}
-				}
 { .mib;	add	r10=2*$SZ,ctx
 	add	r11=3*$SZ,ctx
-	brp.loop.imp	.L_rest,.L_rest_ctop
+	brp.loop.imp	.L_rest,.L_rest_end-16		};;
-				};;
-// load A-H
-{ .mmi;	$LDW	A=[r8],4*$SZ
-	$LDW	B=[r9],4*$SZ
-	mov	sgm0=$sigma0[2]	}
-{ .mmi;	$LDW	C=[r10],4*$SZ
-	$LDW	D=[r11],4*$SZ
-	mov	sgm1=$sigma1[2]	};;
-{ .mmi;	$LDW	E=[r8]
-	$LDW	F=[r9]		}
-{ .mmi;	$LDW	G=[r10]
-	$LDW	H=[r11]
-	cmp.ne	p15,p14=0,r35	};;	// used in sha256_block
 
+// load A-H
+.Lpic_point:
+{ .mmi;	$LDW	A_=[r8],4*$SZ
+	$LDW	B_=[r9],4*$SZ
+	mov	Ktbl=ip		}
+{ .mmi;	$LDW	C_=[r10],4*$SZ
+	$LDW	D_=[r11],4*$SZ
+	mov	sgm0=$sigma0[2]	};;
+{ .mmi;	$LDW	E_=[r8]
+	$LDW	F_=[r9]
+	add	Ktbl=($TABLE#-.Lpic_point),Ktbl		}
+{ .mmi;	$LDW	G_=[r10]
+	$LDW	H_=[r11]
+	cmp.ne	p0,p16=0,r0	};;	// used in sha256_block
+___
+$code.=<<___ if ($BITS==64);
+{ .mii;	and	r8=7,input
+	and	input=~7,input;;
+	cmp.eq	p9,p0=1,r8	}
+{ .mmi;	cmp.eq	p10,p0=2,r8
+	cmp.eq	p11,p0=3,r8
+	cmp.eq	p12,p0=4,r8	}
+{ .mmi;	cmp.eq	p13,p0=5,r8
+	cmp.eq	p14,p0=6,r8
+	cmp.eq	p15,p0=7,r8	};;
+___
+$code.=<<___;
 .L_outer:
-{ .mii;	mov	ar.lc=15
-	mov	ar.ec=1		};;
-.align	32
-.L_first16:
 .rotr	X[16]
+{ .mmi;	mov	A=A_
+	mov	B=B_
+	mov	ar.lc=14	}
+{ .mmi;	mov	C=C_
+	mov	D=D_
+	mov	E=E_		}
+{ .mmi;	mov	F=F_
+	mov	G=G_
+	mov	ar.ec=2		}
+{ .mmi;	ld1	X[15]=[input],$SZ		// eliminated in 64-bit
+	mov	H=H_
+	mov	sgm1=$sigma1[2]	};;
+
 ___
 $t0="t0", $t1="t1", $code.=<<___ if ($BITS==32);
-{ .mib;	(p14)	add	r9=1,input
+.align	32
-	(p14)	add	r10=2,input	}
+.L_first16:
-{ .mib;	(p14)	add	r11=3,input
+{ .mmi;		add	r9=1-$SZ,input
-	(p15)	br.dptk.few	.L_host	};;
+		add	r10=2-$SZ,input
-{ .mmi;	(p14)	ld1	r8=[input],$SZ
+		add	r11=3-$SZ,input	};;
-	(p14)	ld1	r9=[r9]		}
+{ .mmi;		ld1	r9=[r9]
-{ .mmi;	(p14)	ld1	r10=[r10]
+		ld1	r10=[r10]
-	(p14)	ld1	r11=[r11]	};;
-{ .mii;	(p14)	dep	r9=r8,r9,8,8
-	(p14)	dep	r11=r10,r11,8,8	};;
-{ .mib;	(p14)	dep	X[15]=r9,r11,16,16 };;
-.L_host:
-{ .mib;	(p15)	$LDW	X[15]=[input],$SZ	// X[i]=*input++
 		dep.z	$t1=E,32,32	}
-{ .mib;		$LDW	K=[Ktbl],$SZ
+{ .mmi;		$LDW	K=[Ktbl],$SZ
+		ld1	r11=[r11]
 		zxt4	E=E		};;
-{ .mmi;		or	$t1=$t1,E
+{ .mii;		or	$t1=$t1,E
-		and	T1=F,E
+		dep	X[15]=X[15],r9,8,8
-		and	T2=A,B		}
+		dep	r11=r10,r11,8,8	};;
+{ .mmi;		and	T1=F,E
+		and	T2=A,B
+		dep	X[15]=X[15],r11,16,16	}
 { .mmi;		andcm	r8=G,E
 		and	r9=A,C
 		mux2	$t0=A,0x44	};;	// copy lower half to upper
-{ .mib;		xor	T1=T1,r8		// T1=((e & f) ^ (~e & g))
+{ .mmi;	(p16)	ld1	X[15-1]=[input],$SZ	// prefetch
+		xor	T1=T1,r8		// T1=((e & f) ^ (~e & g))
 		_rotr	r11=$t1,$Sigma1[0] }	// ROTR(e,14)
 { .mib;		and	r10=B,C
 		xor	T2=T2,r9	};;
 ___
 $t0="A", $t1="E", $code.=<<___ if ($BITS==64);
-{ .mmi;		$LDW	X[15]=[input],$SZ	// X[i]=*input++
+// in 64-bit mode I load whole X[16] at once and take care of alignment...
+{ .mmi;	add	r8=1*$SZ,input
+	add	r9=2*$SZ,input
+	add	r10=3*$SZ,input		};;
+{ .mmb;	$LDW	X[15]=[input],4*$SZ
+	$LDW	X[14]=[r8],4*$SZ
+(p9)	br.cond.dpnt.many	.L1byte	};;
+{ .mmb;	$LDW	X[13]=[r9],4*$SZ
+	$LDW	X[12]=[r10],4*$SZ
+(p10)	br.cond.dpnt.many	.L2byte	};;
+{ .mmb;	$LDW	X[11]=[input],4*$SZ
+	$LDW	X[10]=[r8],4*$SZ
+(p11)	br.cond.dpnt.many	.L3byte	};;
+{ .mmb;	$LDW	X[ 9]=[r9],4*$SZ
+	$LDW	X[ 8]=[r10],4*$SZ
+(p12)	br.cond.dpnt.many	.L4byte	};;
+{ .mmb;	$LDW	X[ 7]=[input],4*$SZ
+	$LDW	X[ 6]=[r8],4*$SZ
+(p13)	br.cond.dpnt.many	.L5byte	};;
+{ .mmb;	$LDW	X[ 5]=[r9],4*$SZ
+	$LDW	X[ 4]=[r10],4*$SZ
+(p14)	br.cond.dpnt.many	.L6byte	};;
+{ .mmb;	$LDW	X[ 3]=[input],4*$SZ
+	$LDW	X[ 2]=[r8],4*$SZ
+(p15)	br.cond.dpnt.many	.L7byte	};;
+{ .mmb;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	br.many	.L_first16		};;
+.L1byte:
+{ .mmi;	$LDW	X[13]=[r9],4*$SZ
+	$LDW	X[12]=[r10],4*$SZ
+	shrp	X[15]=X[15],X[14],56	};;
+{ .mmi;	$LDW	X[11]=[input],4*$SZ
+	$LDW	X[10]=[r8],4*$SZ
+	shrp	X[14]=X[14],X[13],56	}
+{ .mmi;	$LDW	X[ 9]=[r9],4*$SZ
+	$LDW	X[ 8]=[r10],4*$SZ
+	shrp	X[13]=X[13],X[12],56	};;
+{ .mmi;	$LDW	X[ 7]=[input],4*$SZ
+	$LDW	X[ 6]=[r8],4*$SZ
+	shrp	X[12]=X[12],X[11],56	}
+{ .mmi;	$LDW	X[ 5]=[r9],4*$SZ
+	$LDW	X[ 4]=[r10],4*$SZ
+	shrp	X[11]=X[11],X[10],56	};;
+{ .mmi;	$LDW	X[ 3]=[input],4*$SZ
+	$LDW	X[ 2]=[r8],4*$SZ
+	shrp	X[10]=X[10],X[ 9],56	}
+{ .mmi;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	shrp	X[ 9]=X[ 9],X[ 8],56	};;
+{ .mii;	$LDW	T1=[input]
+	shrp	X[ 8]=X[ 8],X[ 7],56
+	shrp	X[ 7]=X[ 7],X[ 6],56	}
+{ .mii;	shrp	X[ 6]=X[ 6],X[ 5],56
+	shrp	X[ 5]=X[ 5],X[ 4],56	};;
+{ .mii;	shrp	X[ 4]=X[ 4],X[ 3],56
+	shrp	X[ 3]=X[ 3],X[ 2],56	}
+{ .mii;	shrp	X[ 2]=X[ 2],X[ 1],56
+	shrp	X[ 1]=X[ 1],X[ 0],56	}
+{ .mib;	shrp	X[ 0]=X[ 0],T1,56
+	br.many	.L_first16		};;
+.L2byte:
+{ .mmi;	$LDW	X[11]=[input],4*$SZ
+	$LDW	X[10]=[r8],4*$SZ
+	shrp	X[15]=X[15],X[14],48	}
+{ .mmi;	$LDW	X[ 9]=[r9],4*$SZ
+	$LDW	X[ 8]=[r10],4*$SZ
+	shrp	X[14]=X[14],X[13],48	};;
+{ .mmi;	$LDW	X[ 7]=[input],4*$SZ
+	$LDW	X[ 6]=[r8],4*$SZ
+	shrp	X[13]=X[13],X[12],48	}
+{ .mmi;	$LDW	X[ 5]=[r9],4*$SZ
+	$LDW	X[ 4]=[r10],4*$SZ
+	shrp	X[12]=X[12],X[11],48	};;
+{ .mmi;	$LDW	X[ 3]=[input],4*$SZ
+	$LDW	X[ 2]=[r8],4*$SZ
+	shrp	X[11]=X[11],X[10],48	}
+{ .mmi;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	shrp	X[10]=X[10],X[ 9],48	};;
+{ .mii;	$LDW	T1=[input]
+	shrp	X[ 9]=X[ 9],X[ 8],48
+	shrp	X[ 8]=X[ 8],X[ 7],48	}
+{ .mii;	shrp	X[ 7]=X[ 7],X[ 6],48
+	shrp	X[ 6]=X[ 6],X[ 5],48	};;
+{ .mii;	shrp	X[ 5]=X[ 5],X[ 4],48
+	shrp	X[ 4]=X[ 4],X[ 3],48	}
+{ .mii;	shrp	X[ 3]=X[ 3],X[ 2],48
+	shrp	X[ 2]=X[ 2],X[ 1],48	}
+{ .mii;	shrp	X[ 1]=X[ 1],X[ 0],48
+	shrp	X[ 0]=X[ 0],T1,48	}
+{ .mfb;	br.many	.L_first16		};;
+.L3byte:
+{ .mmi;	$LDW	X[ 9]=[r9],4*$SZ
+	$LDW	X[ 8]=[r10],4*$SZ
+	shrp	X[15]=X[15],X[14],40	};;
+{ .mmi;	$LDW	X[ 7]=[input],4*$SZ
+	$LDW	X[ 6]=[r8],4*$SZ
+	shrp	X[14]=X[14],X[13],40	}
+{ .mmi;	$LDW	X[ 5]=[r9],4*$SZ
+	$LDW	X[ 4]=[r10],4*$SZ
+	shrp	X[13]=X[13],X[12],40	};;
+{ .mmi;	$LDW	X[ 3]=[input],4*$SZ
+	$LDW	X[ 2]=[r8],4*$SZ
+	shrp	X[12]=X[12],X[11],40	}
+{ .mmi;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	shrp	X[11]=X[11],X[10],40	};;
+{ .mii;	$LDW	T1=[input]
+	shrp	X[10]=X[10],X[ 9],40
+	shrp	X[ 9]=X[ 9],X[ 8],40	}
+{ .mii;	shrp	X[ 8]=X[ 8],X[ 7],40
+	shrp	X[ 7]=X[ 7],X[ 6],40	};;
+{ .mii;	shrp	X[ 6]=X[ 6],X[ 5],40
+	shrp	X[ 5]=X[ 5],X[ 4],40	}
+{ .mii;	shrp	X[ 4]=X[ 4],X[ 3],40
+	shrp	X[ 3]=X[ 3],X[ 2],40	}
+{ .mii;	shrp	X[ 2]=X[ 2],X[ 1],40
+	shrp	X[ 1]=X[ 1],X[ 0],40	}
+{ .mib;	shrp	X[ 0]=X[ 0],T1,40
+	br.many	.L_first16		};;
+.L4byte:
+{ .mmi;	$LDW	X[ 7]=[input],4*$SZ
+	$LDW	X[ 6]=[r8],4*$SZ
+	shrp	X[15]=X[15],X[14],32	}
+{ .mmi;	$LDW	X[ 5]=[r9],4*$SZ
+	$LDW	X[ 4]=[r10],4*$SZ
+	shrp	X[14]=X[14],X[13],32	};;
+{ .mmi;	$LDW	X[ 3]=[input],4*$SZ
+	$LDW	X[ 2]=[r8],4*$SZ
+	shrp	X[13]=X[13],X[12],32	}
+{ .mmi;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	shrp	X[12]=X[12],X[11],32	};;
+{ .mii;	$LDW	T1=[input]
+	shrp	X[11]=X[11],X[10],32
+	shrp	X[10]=X[10],X[ 9],32	}
+{ .mii;	shrp	X[ 9]=X[ 9],X[ 8],32
+	shrp	X[ 8]=X[ 8],X[ 7],32	};;
+{ .mii;	shrp	X[ 7]=X[ 7],X[ 6],32
+	shrp	X[ 6]=X[ 6],X[ 5],32	}
+{ .mii;	shrp	X[ 5]=X[ 5],X[ 4],32
+	shrp	X[ 4]=X[ 4],X[ 3],32	}
+{ .mii;	shrp	X[ 3]=X[ 3],X[ 2],32
+	shrp	X[ 2]=X[ 2],X[ 1],32	}
+{ .mii;	shrp	X[ 1]=X[ 1],X[ 0],32
+	shrp	X[ 0]=X[ 0],T1,32	}
+{ .mfb;	br.many	.L_first16		};;
+.L5byte:
+{ .mmi;	$LDW	X[ 5]=[r9],4*$SZ
+	$LDW	X[ 4]=[r10],4*$SZ
+	shrp	X[15]=X[15],X[14],24	};;
+{ .mmi;	$LDW	X[ 3]=[input],4*$SZ
+	$LDW	X[ 2]=[r8],4*$SZ
+	shrp	X[14]=X[14],X[13],24	}
+{ .mmi;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	shrp	X[13]=X[13],X[12],24	};;
+{ .mii;	$LDW	T1=[input]
+	shrp	X[12]=X[12],X[11],24
+	shrp	X[11]=X[11],X[10],24	}
+{ .mii;	shrp	X[10]=X[10],X[ 9],24
+	shrp	X[ 9]=X[ 9],X[ 8],24	};;
+{ .mii;	shrp	X[ 8]=X[ 8],X[ 7],24
+	shrp	X[ 7]=X[ 7],X[ 6],24	}
+{ .mii;	shrp	X[ 6]=X[ 6],X[ 5],24
+	shrp	X[ 5]=X[ 5],X[ 4],24	}
+{ .mii;	shrp	X[ 4]=X[ 4],X[ 3],24
+	shrp	X[ 3]=X[ 3],X[ 2],24	}
+{ .mii;	shrp	X[ 2]=X[ 2],X[ 1],24
+	shrp	X[ 1]=X[ 1],X[ 0],24	}
+{ .mib;	shrp	X[ 0]=X[ 0],T1,24
+	br.many	.L_first16		};;
+.L6byte:
+{ .mmi;	$LDW	X[ 3]=[input],4*$SZ
+	$LDW	X[ 2]=[r8],4*$SZ
+	shrp	X[15]=X[15],X[14],16	}
+{ .mmi;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	shrp	X[14]=X[14],X[13],16	};;
+{ .mii;	$LDW	T1=[input]
+	shrp	X[13]=X[13],X[12],16
+	shrp	X[12]=X[12],X[11],16	}
+{ .mii;	shrp	X[11]=X[11],X[10],16
+	shrp	X[10]=X[10],X[ 9],16	};;
+{ .mii;	shrp	X[ 9]=X[ 9],X[ 8],16
+	shrp	X[ 8]=X[ 8],X[ 7],16	}
+{ .mii;	shrp	X[ 7]=X[ 7],X[ 6],16
+	shrp	X[ 6]=X[ 6],X[ 5],16	}
+{ .mii;	shrp	X[ 5]=X[ 5],X[ 4],16
+	shrp	X[ 4]=X[ 4],X[ 3],16	}
+{ .mii;	shrp	X[ 3]=X[ 3],X[ 2],16
+	shrp	X[ 2]=X[ 2],X[ 1],16	}
+{ .mii;	shrp	X[ 1]=X[ 1],X[ 0],16
+	shrp	X[ 0]=X[ 0],T1,16	}
+{ .mfb;	br.many	.L_first16		};;
+.L7byte:
+{ .mmi;	$LDW	X[ 1]=[r9],4*$SZ
+	$LDW	X[ 0]=[r10],4*$SZ
+	shrp	X[15]=X[15],X[14],8	};;
+{ .mii;	$LDW	T1=[input]
+	shrp	X[14]=X[14],X[13],8
+	shrp	X[13]=X[13],X[12],8	}
+{ .mii;	shrp	X[12]=X[12],X[11],8
+	shrp	X[11]=X[11],X[10],8	};;
+{ .mii;	shrp	X[10]=X[10],X[ 9],8
+	shrp	X[ 9]=X[ 9],X[ 8],8	}
+{ .mii;	shrp	X[ 8]=X[ 8],X[ 7],8
+	shrp	X[ 7]=X[ 7],X[ 6],8	}
+{ .mii;	shrp	X[ 6]=X[ 6],X[ 5],8
+	shrp	X[ 5]=X[ 5],X[ 4],8	}
+{ .mii;	shrp	X[ 4]=X[ 4],X[ 3],8
+	shrp	X[ 3]=X[ 3],X[ 2],8	}
+{ .mii;	shrp	X[ 2]=X[ 2],X[ 1],8
+	shrp	X[ 1]=X[ 1],X[ 0],8	}
+{ .mib;	shrp	X[ 0]=X[ 0],T1,8
+	br.many	.L_first16		};;
+
+.align	32
+.L_first16:
+{ .mmi;		$LDW	K=[Ktbl],$SZ
 		and	T1=F,E
 		and	T2=A,B		}
-{ .mmi;		$LDW	K=[Ktbl],$SZ
+{ .mmi;		//$LDW	X[15]=[input],$SZ	// X[i]=*input++
 		andcm	r8=G,E
 		and	r9=A,C		};;
 { .mmi;		xor	T1=T1,r8		//T1=((e & f) ^ (~e & g))
@@ -236,13 +478,14 @@ $code.=<<___;
 { .mmi;		xor	r10=r8,r10		// r10=Sigma0(a)
 		mov	B=A
 		add	A=T1,T2		};;
-.L_first16_ctop:
 { .mib;		add	E=E,T1
 		add	A=A,r10			// T2=Maj(a,b,c)+Sigma0(a)
 	br.ctop.sptk	.L_first16	};;
+.L_first16_end:
+
+{ .mii;	mov	ar.lc=$rounds-17
+	mov	ar.ec=1			};;
 
-{ .mib;	mov	ar.lc=$rounds-17	}
-{ .mib;	mov	ar.ec=1			};;
 .align	32
 .L_rest:
 .rotr	X[16]
@@ -311,46 +554,38 @@ $code.=<<___;
 { .mmi;		xor	r10=r8,r10		// r10=Sigma0(a)
 		mov	B=A
 		add	A=T1,T2		};;
-.L_rest_ctop:
 { .mib;		add	E=E,T1
 		add	A=A,r10			// T2=Maj(a,b,c)+Sigma0(a)
 	br.ctop.sptk	.L_rest	};;
+.L_rest_end:
+
+{ .mmi;	add	A_=A_,A
+	add	B_=B_,B
+	add	C_=C_,C			}
+{ .mmi;	add	D_=D_,D
+	add	E_=E_,E
+	cmp.ltu	p16,p0=1,num		};;
+{ .mmi;	add	F_=F_,F
+	add	G_=G_,G
+	add	H_=H_,H			}
+{ .mmb;	add	Ktbl=-$SZ*$rounds,Ktbl
+(p16)	add	num=-1,num
+(p16)	br.dptk.many	.L_outer	};;
 
 { .mib;	add	r8=0*$SZ,ctx
 	add	r9=1*$SZ,ctx		}
 { .mib;	add	r10=2*$SZ,ctx
 	add	r11=3*$SZ,ctx		};;
-{ .mmi;	$LDW	r32=[r8],4*$SZ
+{ .mmi;	$STW	[r8]=A_,4*$SZ
-	$LDW	r33=[r9],4*$SZ		}
+	$STW	[r9]=B_,4*$SZ
-{ .mmi;	$LDW	r34=[r10],4*$SZ
+	mov	ar.lc=lcsave		}
-	$LDW	r35=[r11],4*$SZ
+{ .mmi;	$STW	[r10]=C_,4*$SZ
-	cmp.ltu	p6,p7=1,num		};;
+	$STW	[r11]=D_,4*$SZ
-{ .mmi;	$LDW	r36=[r8],-4*$SZ
+	mov	pr=prsave,0x1ffff	};;
-	$LDW	r37=[r9],-4*$SZ
+{ .mmb;	$STW	[r8]=E_
-(p6)	add	Ktbl=-$SZ*$rounds,Ktbl	}
+	$STW	[r9]=F_			}
-{ .mmi;	$LDW	r38=[r10],-4*$SZ
+{ .mmb;	$STW	[r10]=G_
-	$LDW	r39=[r11],-4*$SZ
+	$STW	[r11]=H_
-(p7)	mov	ar.lc=r3		};;
-{ .mmi;	add	A=A,r32
-	add	B=B,r33
-	add	C=C,r34			}
-{ .mmi;	add	D=D,r35
-	add	E=E,r36
-	add	F=F,r37			};;
-{ .mmi;	$STW	[r8]=A,4*$SZ
-	$STW	[r9]=B,4*$SZ
-	add	G=G,r38			}
-{ .mmi;	$STW	[r10]=C,4*$SZ
-	$STW	[r11]=D,4*$SZ
-	add	H=H,r39			};;
-{ .mmi;	$STW	[r8]=E
-	$STW	[r9]=F
-(p6)	add	num=-1,num		}
-{ .mmb;	$STW	[r10]=G
-	$STW	[r11]=H
-(p6)	br.dptk.many	.L_outer	};;
-
-{ .mib;	mov	pr=prsave,0x1ffff
 	br.ret.sptk.many	b0	};;
 .endp	$func#
 ___
@@ -360,6 +595,9 @@ $code =~ s/_rotr(\s+)([^=]+)=([^,]+),([0-9]+)/shrp$1$2=$3,$3,$4/gm;
 if ($BITS==64) {
     $code =~ s/mux2(\s+)\S+/nop.i$1 0x0/gm;
     $code =~ s/mux1(\s+)\S+/nop.i$1 0x0/gm	if ($big_endian);
+    $code =~ s/(shrp\s+X\[[^=]+)=([^,]+),([^,]+),([1-9]+)/$1=$3,$2,64-$4/gm
+    						if (!$big_endian);
+    $code =~ s/ld1(\s+)X\[\S+/nop.m$1 0x0/gm;
 }
 
 print $code;
@@ -384,6 +622,7 @@ K256:	data4	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
 	data4	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
 	data4	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
 .size	K256#,$SZ*$rounds
+stringz	"SHA256 block transform for IA64, CRYPTOGAMS by <appro\@openssl.org>"
 ___
 print<<___ if ($BITS==64);
 .align	64
@@ -429,4 +668,5 @@ K512:	data8	0x428a2f98d728ae22,0x7137449123ef65cd
 	data8	0x4cc5d4becb3e42b6,0x597f299cfc657e2a
 	data8	0x5fcb6fab3ad6faec,0x6c44198c4a475817
 .size	K512#,$SZ*$rounds
+stringz	"SHA512 block transform for IA64, CRYPTOGAMS by <appro\@openssl.org>"
 ___

crypto/sha/asm/sha512-s390x.pl

@@ -0,0 +1,290 @@
+#!/usr/bin/env perl
+
+# ====================================================================
+# Written by Andy Polyakov <appro@fy.chalmers.se> for the OpenSSL
+# project. The module is, however, dual licensed under OpenSSL and
+# CRYPTOGAMS licenses depending on where you obtain it. For further
+# details see http://www.openssl.org/~appro/cryptogams/.
+# ====================================================================
+
+# SHA256/512 block procedures for s390x.
+
+# April 2007.
+#
+# sha256_block_data_order is reportedly >3 times faster than gcc 3.3
+# generated code (must be a bug in compiler, as improvement is
+# "pathologically" high, in particular in comparison to other SHA
+# modules). But the real twist is that it detects if hardware support
+# for SHA256 is available and in such case utilizes it. Then the
+# performance can reach >6.5x of assembler one for larger chunks.
+#
+# sha512_block_data_order is ~70% faster than gcc 3.3 generated code.
+
+$t0="%r0";
+$t1="%r1";
+$ctx="%r2";
+$inp="%r3";
+$len="%r4";	# used as index in inner loop
+
+$A="%r5";
+$B="%r6";
+$C="%r7";
+$D="%r8";
+$E="%r9";
+$F="%r10";
+$G="%r11";
+$H="%r12";	@V=($A,$B,$C,$D,$E,$F,$G,$H);
+$tbl="%r13";
+$T1="%r14";
+$sp="%r15";
+
+$output=shift;
+open STDOUT,">$output";
+
+if ($output =~ /512/) {
+	$label="512";
+	$SZ=8;
+	$LD="lg";	# load from memory
+	$ST="stg";	# store to memory
+	$ADD="alg";	# add with memory operand
+	$ROT="rllg";	# rotate left
+	$SHR="srlg";	# logical right shift [see even at the end]
+	@Sigma0=(25,30,36);
+	@Sigma1=(23,46,50);
+	@sigma0=(56,63, 7);
+	@sigma1=( 3,45, 6);
+	$rounds=80;
+	$kimdfunc=0;	# 0 means unknown/unsupported/unimplemented
+} else {
+	$label="256";
+	$SZ=4;
+	$LD="llgf";	# load from memory
+	$ST="st";	# store to memory
+	$ADD="al";	# add with memory operand
+	$ROT="rll";	# rotate left
+	$SHR="srl";	# logical right shift
+	@Sigma0=(10,19,30);
+	@Sigma1=( 7,21,26);
+	@sigma0=(14,25, 3);
+	@sigma1=(13,15,10);
+	$rounds=64;
+	$kimdfunc=0;	# magic function code for kimd instruction,
+			# 0 disables hardware support
+}
+$Func="sha${label}_block_data_order";
+$Table="K${label}";
+$frame=160+16*$SZ;
+
+sub BODY_00_15 {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___ if ($i<16);
+	$LD	$T1,`$i*$SZ`($inp)	### $i
+___
+$code.=<<___;
+	$ROT	$t0,$e,$Sigma1[0]
+	$ROT	$t1,$e,$Sigma1[1]
+	xgr	$t0,$t1
+	$ROT	$t1,$t1,`$Sigma1[2]-$Sigma1[1]`
+	xgr	$t0,$t1			# Sigma1(e)
+	$ST	$T1,`160+$SZ*($i%16)`($sp)
+	algr	$T1,$t0			# T1+=Sigma1(e)
+	algr	$T1,$h			# T1+=h
+	$ADD	$T1,`$i*$SZ`($len,$tbl)	# T1+=K[i]
+	lgr	$t0,$f
+	xgr	$t0,$g
+	ngr	$t0,$e
+	xgr	$t0,$g			# Ch(e,f,g)
+	algr	$T1,$t0			# T1+=Ch(e,f,g)
+	$ROT	$h,$a,$Sigma0[0]
+	$ROT	$t0,$a,$Sigma0[1]
+	xgr	$h,$t0
+	$ROT	$t0,$t0,`$Sigma0[2]-$Sigma0[1]`
+	xgr	$h,$t0			# h=Sigma0(a)
+	lgr	$t0,$a
+	ogr	$t0,$b
+	ngr	$t0,$c
+	lgr	$t1,$a
+	ngr	$t1,$b
+	ogr	$t0,$t1			# Maj(a,b,c)
+	algr	$h,$t0			# h+=Maj(a,b,c)
+	algr	$d,$T1			# d+=T1
+	algr	$h,$T1			# h+=T1
+___
+}
+
+sub BODY_16_XX {
+my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___;
+	$LD	$T1,`160+$SZ*(($i+1)%16)`($sp)	### $i
+	$LD	$t1,`160+$SZ*(($i+14)%16)`($sp)
+	$ROT	$t0,$T1,$sigma0[0]
+	$SHR	$T1,$sigma0[2]
+	xgr	$T1,$t0
+	$ROT	$t0,$t0,`$sigma0[1]-$sigma0[0]`
+	xgr	$T1,$t0				# sigma0(X[i+1])
+	$ROT	$t0,$t1,$sigma1[0]
+	$ADD	$T1,`160+$SZ*($i%16)`($sp)	# +=X[i]
+	$SHR	$t1,$sigma1[2]
+	xgr	$t1,$t0
+	$ADD	$T1,`160+$SZ*(($i+9)%16)`($sp)	# +=X[i+9]
+	$ROT	$t0,$t0,`$sigma1[1]-$sigma1[0]`
+	xgr	$t1,$t0				# sigma1(X[i+14])
+	algr	$T1,$t1				# +=sigma1(X[i+14])
+___
+	&BODY_00_15(@_);
+}
+
+$code.=<<___;
+.text
+.align	64
+.type	$Table,\@object
+$Table:
+___
+$code.=<<___ if ($SZ==4);
+	.long	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
+	.long	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
+	.long	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
+	.long	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
+	.long	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
+	.long	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
+	.long	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
+	.long	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
+	.long	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
+	.long	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
+	.long	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
+	.long	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
+	.long	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
+	.long	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
+	.long	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
+	.long	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
+___
+$code.=<<___ if ($SZ==8);
+	.quad	0x428a2f98d728ae22,0x7137449123ef65cd
+	.quad	0xb5c0fbcfec4d3b2f,0xe9b5dba58189dbbc
+	.quad	0x3956c25bf348b538,0x59f111f1b605d019
+	.quad	0x923f82a4af194f9b,0xab1c5ed5da6d8118
+	.quad	0xd807aa98a3030242,0x12835b0145706fbe
+	.quad	0x243185be4ee4b28c,0x550c7dc3d5ffb4e2
+	.quad	0x72be5d74f27b896f,0x80deb1fe3b1696b1
+	.quad	0x9bdc06a725c71235,0xc19bf174cf692694
+	.quad	0xe49b69c19ef14ad2,0xefbe4786384f25e3
+	.quad	0x0fc19dc68b8cd5b5,0x240ca1cc77ac9c65
+	.quad	0x2de92c6f592b0275,0x4a7484aa6ea6e483
+	.quad	0x5cb0a9dcbd41fbd4,0x76f988da831153b5
+	.quad	0x983e5152ee66dfab,0xa831c66d2db43210
+	.quad	0xb00327c898fb213f,0xbf597fc7beef0ee4
+	.quad	0xc6e00bf33da88fc2,0xd5a79147930aa725
+	.quad	0x06ca6351e003826f,0x142929670a0e6e70
+	.quad	0x27b70a8546d22ffc,0x2e1b21385c26c926
+	.quad	0x4d2c6dfc5ac42aed,0x53380d139d95b3df
+	.quad	0x650a73548baf63de,0x766a0abb3c77b2a8
+	.quad	0x81c2c92e47edaee6,0x92722c851482353b
+	.quad	0xa2bfe8a14cf10364,0xa81a664bbc423001
+	.quad	0xc24b8b70d0f89791,0xc76c51a30654be30
+	.quad	0xd192e819d6ef5218,0xd69906245565a910
+	.quad	0xf40e35855771202a,0x106aa07032bbd1b8
+	.quad	0x19a4c116b8d2d0c8,0x1e376c085141ab53
+	.quad	0x2748774cdf8eeb99,0x34b0bcb5e19b48a8
+	.quad	0x391c0cb3c5c95a63,0x4ed8aa4ae3418acb
+	.quad	0x5b9cca4f7763e373,0x682e6ff3d6b2b8a3
+	.quad	0x748f82ee5defb2fc,0x78a5636f43172f60
+	.quad	0x84c87814a1f0ab72,0x8cc702081a6439ec
+	.quad	0x90befffa23631e28,0xa4506cebde82bde9
+	.quad	0xbef9a3f7b2c67915,0xc67178f2e372532b
+	.quad	0xca273eceea26619c,0xd186b8c721c0c207
+	.quad	0xeada7dd6cde0eb1e,0xf57d4f7fee6ed178
+	.quad	0x06f067aa72176fba,0x0a637dc5a2c898a6
+	.quad	0x113f9804bef90dae,0x1b710b35131c471b
+	.quad	0x28db77f523047d84,0x32caab7b40c72493
+	.quad	0x3c9ebe0a15c9bebc,0x431d67c49c100d4c
+	.quad	0x4cc5d4becb3e42b6,0x597f299cfc657e2a
+	.quad	0x5fcb6fab3ad6faec,0x6c44198c4a475817
+___
+$code.=<<___;
+.size	$Table,.-$Table
+.globl	$Func
+.type	$Func,\@function
+$Func:
+___
+$code.=<<___ if ($kimdfunc);
+	lghi	%r0,0
+	la	%r1,16($sp)
+	.long	0xb93e0002	# kimd %r0,%r2
+	lg	%r0,16($sp)
+	tmhh	%r0,`0x8000>>$kimdfunc`
+	jz	.Lsoftware
+	lghi	%r0,$kimdfunc
+	lgr	%r1,$ctx
+	lgr	%r2,$inp
+	sllg	%r3,$len,`log(16*$SZ)/log(2)`
+	.long	0xb93e0002	# kimd %r0,%r2
+	brc	1,.-4		# pay attention to "partial completion"
+	br	%r14
+.Lsoftware:
+___
+$code.=<<___;
+	sllg	$len,$len,`log(16*$SZ)/log(2)`
+	la	$len,0($inp,$len)
+	stmg	$len,%r15,32($sp)
+	lgr	%r0,$sp
+	aghi	$sp,-$frame
+	stg	%r0,0($sp)
+
+	bras	$tbl,.Lpic
+.Lpic:	aghi	$tbl,$Table-.Lpic
+
+	$LD	$A,`0*$SZ`($ctx)
+	$LD	$B,`1*$SZ`($ctx)
+	$LD	$C,`2*$SZ`($ctx)
+	$LD	$D,`3*$SZ`($ctx)
+	$LD	$E,`4*$SZ`($ctx)
+	$LD	$F,`5*$SZ`($ctx)
+	$LD	$G,`6*$SZ`($ctx)
+	$LD	$H,`7*$SZ`($ctx)
+
+.Lloop:
+	lghi	$len,0
+___
+for ($i=0;$i<16;$i++)	{ &BODY_00_15($i,@V); unshift(@V,pop(@V)); }
+$code.=".Lrounds_16_xx:\n";
+for (;$i<32;$i++)	{ &BODY_16_XX($i,@V); unshift(@V,pop(@V)); }
+$code.=<<___;
+	aghi	$len,`16*$SZ`
+	lghi	$t0,`($rounds-16)*$SZ`
+	clgr	$len,$t0
+	jne	.Lrounds_16_xx
+
+	$ADD	$A,`0*$SZ`($ctx)
+	$ADD	$B,`1*$SZ`($ctx)
+	$ADD	$C,`2*$SZ`($ctx)
+	$ADD	$D,`3*$SZ`($ctx)
+	$ADD	$E,`4*$SZ`($ctx)
+	$ADD	$F,`5*$SZ`($ctx)
+	$ADD	$G,`6*$SZ`($ctx)
+	$ADD	$H,`7*$SZ`($ctx)
+	$ST	$A,`0*$SZ`($ctx)
+	$ST	$B,`1*$SZ`($ctx)
+	$ST	$C,`2*$SZ`($ctx)
+	$ST	$D,`3*$SZ`($ctx)
+	$ST	$E,`4*$SZ`($ctx)
+	$ST	$F,`5*$SZ`($ctx)
+	$ST	$G,`6*$SZ`($ctx)
+	$ST	$H,`7*$SZ`($ctx)
+	la	$inp,`16*$SZ`($inp)
+	clg	$inp,`$frame+32`($sp)
+	jne	.Lloop
+
+	lmg	%r6,%r15,`$frame+48`($sp)	
+	br	%r14
+.size	$Func,.-$Func
+.string	"SHA${label} block transform for s390x, CRYPTOGAMS by <appro\@openssl.org>"
+___
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+# unlike 32-bit shift 64-bit one takes three arguments
+$code =~ s/(srlg\s+)(%r[0-9]+),/$1$2,$2,/gm;
+
+print $code;
+close STDOUT;

crypto/sha/asm/sha512-x86_64.pl

@@ -0,0 +1,344 @@
+#!/usr/bin/env perl
+#
+# ====================================================================
+# Written 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.
+# ====================================================================
+#
+# sha256/512_block procedure for x86_64.
+#
+# 40% improvement over compiler-generated code on Opteron. On EM64T
+# sha256 was observed to run >80% faster and sha512 - >40%. No magical
+# tricks, just straight implementation... I really wonder why gcc
+# [being armed with inline assembler] fails to generate as fast code.
+# The only thing which is cool about this module is that it's very
+# same instruction sequence used for both SHA-256 and SHA-512. In
+# former case the instructions operate on 32-bit operands, while in
+# latter - on 64-bit ones. All I had to do is to get one flavor right,
+# the other one passed the test right away:-)
+#
+# sha256_block runs in ~1005 cycles on Opteron, which gives you
+# asymptotic performance of 64*1000/1005=63.7MBps times CPU clock
+# frequency in GHz. sha512_block runs in ~1275 cycles, which results
+# in 128*1000/1275=100MBps per GHz. Is there room for improvement?
+# Well, if you compare it to IA-64 implementation, which maintains
+# X[16] in register bank[!], tends to 4 instructions per CPU clock
+# cycle and runs in 1003 cycles, 1275 is very good result for 3-way
+# issue Opteron pipeline and X[16] maintained in memory. So that *if*
+# there is a way to improve it, *then* the only way would be to try to
+# offload X[16] updates to SSE unit, but that would require "deeper"
+# loop unroll, which in turn would naturally cause size blow-up, not
+# to mention increased complexity! And once again, only *if* it's
+# actually possible to noticeably improve overall ILP, instruction
+# level parallelism, on a given CPU implementation in this case.
+#
+# Special note on Intel EM64T. While Opteron CPU exhibits perfect
+# perfromance ratio of 1.5 between 64- and 32-bit flavors [see above],
+# [currently available] EM64T CPUs apparently are far from it. On the
+# contrary, 64-bit version, sha512_block, is ~30% *slower* than 32-bit
+# sha256_block:-( This is presumably because 64-bit shifts/rotates
+# apparently are not atomic instructions, but implemented in microcode.
+
+$output=shift;
+
+$0 =~ m/(.*[\/\\])[^\/\\]+$/; $dir=$1;
+( $xlate="${dir}x86_64-xlate.pl" and -f $xlate ) or
+( $xlate="${dir}../../perlasm/x86_64-xlate.pl" and -f $xlate) or
+die "can't locate x86_64-xlate.pl";
+
+open STDOUT,"| $^X $xlate $output";
+
+if ($output =~ /512/) {
+	$func="sha512_block_data_order";
+	$TABLE="K512";
+	$SZ=8;
+	@ROT=($A,$B,$C,$D,$E,$F,$G,$H)=("%rax","%rbx","%rcx","%rdx",
+					"%r8", "%r9", "%r10","%r11");
+	($T1,$a0,$a1,$a2)=("%r12","%r13","%r14","%r15");
+	@Sigma0=(28,34,39);
+	@Sigma1=(14,18,41);
+	@sigma0=(1,  8, 7);
+	@sigma1=(19,61, 6);
+	$rounds=80;
+} else {
+	$func="sha256_block_data_order";
+	$TABLE="K256";
+	$SZ=4;
+	@ROT=($A,$B,$C,$D,$E,$F,$G,$H)=("%eax","%ebx","%ecx","%edx",
+					"%r8d","%r9d","%r10d","%r11d");
+	($T1,$a0,$a1,$a2)=("%r12d","%r13d","%r14d","%r15d");
+	@Sigma0=( 2,13,22);
+	@Sigma1=( 6,11,25);
+	@sigma0=( 7,18, 3);
+	@sigma1=(17,19,10);
+	$rounds=64;
+}
+
+$ctx="%rdi";	# 1st arg
+$round="%rdi";	# zaps $ctx
+$inp="%rsi";	# 2nd arg
+$Tbl="%rbp";
+
+$_ctx="16*$SZ+0*8(%rsp)";
+$_inp="16*$SZ+1*8(%rsp)";
+$_end="16*$SZ+2*8(%rsp)";
+$_rsp="16*$SZ+3*8(%rsp)";
+$framesz="16*$SZ+4*8";
+
+
+sub ROUND_00_15()
+{ my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___;
+	mov	$e,$a0
+	mov	$e,$a1
+	mov	$f,$a2
+
+	ror	\$$Sigma1[0],$a0
+	ror	\$$Sigma1[1],$a1
+	xor	$g,$a2			# f^g
+
+	xor	$a1,$a0
+	ror	\$`$Sigma1[2]-$Sigma1[1]`,$a1
+	and	$e,$a2			# (f^g)&e
+	mov	$T1,`$SZ*($i&0xf)`(%rsp)
+
+	xor	$a1,$a0			# Sigma1(e)
+	xor	$g,$a2			# Ch(e,f,g)=((f^g)&e)^g
+	add	$h,$T1			# T1+=h
+
+	mov	$a,$h
+	add	$a0,$T1			# T1+=Sigma1(e)
+
+	add	$a2,$T1			# T1+=Ch(e,f,g)
+	mov	$a,$a0
+	mov	$a,$a1
+
+	ror	\$$Sigma0[0],$h
+	ror	\$$Sigma0[1],$a0
+	mov	$a,$a2
+	add	($Tbl,$round,$SZ),$T1	# T1+=K[round]
+
+	xor	$a0,$h
+	ror	\$`$Sigma0[2]-$Sigma0[1]`,$a0
+	or	$c,$a1			# a|c
+
+	xor	$a0,$h			# h=Sigma0(a)
+	and	$c,$a2			# a&c
+	add	$T1,$d			# d+=T1
+
+	and	$b,$a1			# (a|c)&b
+	add	$T1,$h			# h+=T1
+
+	or	$a2,$a1			# Maj(a,b,c)=((a|c)&b)|(a&c)
+	lea	1($round),$round	# round++
+
+	add	$a1,$h			# h+=Maj(a,b,c)
+___
+}
+
+sub ROUND_16_XX()
+{ my ($i,$a,$b,$c,$d,$e,$f,$g,$h) = @_;
+
+$code.=<<___;
+	mov	`$SZ*(($i+1)&0xf)`(%rsp),$a0
+	mov	`$SZ*(($i+14)&0xf)`(%rsp),$T1
+
+	mov	$a0,$a2
+
+	shr	\$$sigma0[2],$a0
+	ror	\$$sigma0[0],$a2
+
+	xor	$a2,$a0
+	ror	\$`$sigma0[1]-$sigma0[0]`,$a2
+
+	xor	$a2,$a0			# sigma0(X[(i+1)&0xf])
+	mov	$T1,$a1
+
+	shr	\$$sigma1[2],$T1
+	ror	\$$sigma1[0],$a1
+
+	xor	$a1,$T1
+	ror	\$`$sigma1[1]-$sigma1[0]`,$a1
+
+	xor	$a1,$T1			# sigma1(X[(i+14)&0xf])
+
+	add	$a0,$T1
+
+	add	`$SZ*(($i+9)&0xf)`(%rsp),$T1
+
+	add	`$SZ*($i&0xf)`(%rsp),$T1
+___
+	&ROUND_00_15(@_);
+}
+
+$code=<<___;
+.text
+
+.globl	$func
+.type	$func,\@function,4
+.align	16
+$func:
+	push	%rbx
+	push	%rbp
+	push	%r12
+	push	%r13
+	push	%r14
+	push	%r15
+	mov	%rsp,%rbp		# copy %rsp
+	shl	\$4,%rdx		# num*16
+	sub	\$$framesz,%rsp
+	lea	($inp,%rdx,$SZ),%rdx	# inp+num*16*$SZ
+	and	\$-64,%rsp		# align stack frame
+	mov	$ctx,$_ctx		# save ctx, 1st arg
+	mov	$inp,$_inp		# save inp, 2nd arh
+	mov	%rdx,$_end		# save end pointer, "3rd" arg
+	mov	%rbp,$_rsp		# save copy of %rsp
+
+	.picmeup $Tbl
+	lea	$TABLE-.($Tbl),$Tbl
+
+	mov	$SZ*0($ctx),$A
+	mov	$SZ*1($ctx),$B
+	mov	$SZ*2($ctx),$C
+	mov	$SZ*3($ctx),$D
+	mov	$SZ*4($ctx),$E
+	mov	$SZ*5($ctx),$F
+	mov	$SZ*6($ctx),$G
+	mov	$SZ*7($ctx),$H
+	jmp	.Lloop
+
+.align	16
+.Lloop:
+	xor	$round,$round
+___
+	for($i=0;$i<16;$i++) {
+		$code.="	mov	$SZ*$i($inp),$T1\n";
+		$code.="	bswap	$T1\n";
+		&ROUND_00_15($i,@ROT);
+		unshift(@ROT,pop(@ROT));
+	}
+$code.=<<___;
+	jmp	.Lrounds_16_xx
+.align	16
+.Lrounds_16_xx:
+___
+	for(;$i<32;$i++) {
+		&ROUND_16_XX($i,@ROT);
+		unshift(@ROT,pop(@ROT));
+	}
+
+$code.=<<___;
+	cmp	\$$rounds,$round
+	jb	.Lrounds_16_xx
+
+	mov	$_ctx,$ctx
+	lea	16*$SZ($inp),$inp
+
+	add	$SZ*0($ctx),$A
+	add	$SZ*1($ctx),$B
+	add	$SZ*2($ctx),$C
+	add	$SZ*3($ctx),$D
+	add	$SZ*4($ctx),$E
+	add	$SZ*5($ctx),$F
+	add	$SZ*6($ctx),$G
+	add	$SZ*7($ctx),$H
+
+	cmp	$_end,$inp
+
+	mov	$A,$SZ*0($ctx)
+	mov	$B,$SZ*1($ctx)
+	mov	$C,$SZ*2($ctx)
+	mov	$D,$SZ*3($ctx)
+	mov	$E,$SZ*4($ctx)
+	mov	$F,$SZ*5($ctx)
+	mov	$G,$SZ*6($ctx)
+	mov	$H,$SZ*7($ctx)
+	jb	.Lloop
+
+	mov	$_rsp,%rsp
+	pop	%r15
+	pop	%r14
+	pop	%r13
+	pop	%r12
+	pop	%rbp
+	pop	%rbx
+
+	ret
+.size	$func,.-$func
+___
+
+if ($SZ==4) {
+$code.=<<___;
+.align	64
+.type	$TABLE,\@object
+$TABLE:
+	.long	0x428a2f98,0x71374491,0xb5c0fbcf,0xe9b5dba5
+	.long	0x3956c25b,0x59f111f1,0x923f82a4,0xab1c5ed5
+	.long	0xd807aa98,0x12835b01,0x243185be,0x550c7dc3
+	.long	0x72be5d74,0x80deb1fe,0x9bdc06a7,0xc19bf174
+	.long	0xe49b69c1,0xefbe4786,0x0fc19dc6,0x240ca1cc
+	.long	0x2de92c6f,0x4a7484aa,0x5cb0a9dc,0x76f988da
+	.long	0x983e5152,0xa831c66d,0xb00327c8,0xbf597fc7
+	.long	0xc6e00bf3,0xd5a79147,0x06ca6351,0x14292967
+	.long	0x27b70a85,0x2e1b2138,0x4d2c6dfc,0x53380d13
+	.long	0x650a7354,0x766a0abb,0x81c2c92e,0x92722c85
+	.long	0xa2bfe8a1,0xa81a664b,0xc24b8b70,0xc76c51a3
+	.long	0xd192e819,0xd6990624,0xf40e3585,0x106aa070
+	.long	0x19a4c116,0x1e376c08,0x2748774c,0x34b0bcb5
+	.long	0x391c0cb3,0x4ed8aa4a,0x5b9cca4f,0x682e6ff3
+	.long	0x748f82ee,0x78a5636f,0x84c87814,0x8cc70208
+	.long	0x90befffa,0xa4506ceb,0xbef9a3f7,0xc67178f2
+___
+} else {
+$code.=<<___;
+.align	64
+.type	$TABLE,\@object
+$TABLE:
+	.quad	0x428a2f98d728ae22,0x7137449123ef65cd
+	.quad	0xb5c0fbcfec4d3b2f,0xe9b5dba58189dbbc
+	.quad	0x3956c25bf348b538,0x59f111f1b605d019
+	.quad	0x923f82a4af194f9b,0xab1c5ed5da6d8118
+	.quad	0xd807aa98a3030242,0x12835b0145706fbe
+	.quad	0x243185be4ee4b28c,0x550c7dc3d5ffb4e2
+	.quad	0x72be5d74f27b896f,0x80deb1fe3b1696b1
+	.quad	0x9bdc06a725c71235,0xc19bf174cf692694
+	.quad	0xe49b69c19ef14ad2,0xefbe4786384f25e3
+	.quad	0x0fc19dc68b8cd5b5,0x240ca1cc77ac9c65
+	.quad	0x2de92c6f592b0275,0x4a7484aa6ea6e483
+	.quad	0x5cb0a9dcbd41fbd4,0x76f988da831153b5
+	.quad	0x983e5152ee66dfab,0xa831c66d2db43210
+	.quad	0xb00327c898fb213f,0xbf597fc7beef0ee4
+	.quad	0xc6e00bf33da88fc2,0xd5a79147930aa725
+	.quad	0x06ca6351e003826f,0x142929670a0e6e70
+	.quad	0x27b70a8546d22ffc,0x2e1b21385c26c926
+	.quad	0x4d2c6dfc5ac42aed,0x53380d139d95b3df
+	.quad	0x650a73548baf63de,0x766a0abb3c77b2a8
+	.quad	0x81c2c92e47edaee6,0x92722c851482353b
+	.quad	0xa2bfe8a14cf10364,0xa81a664bbc423001
+	.quad	0xc24b8b70d0f89791,0xc76c51a30654be30
+	.quad	0xd192e819d6ef5218,0xd69906245565a910
+	.quad	0xf40e35855771202a,0x106aa07032bbd1b8
+	.quad	0x19a4c116b8d2d0c8,0x1e376c085141ab53
+	.quad	0x2748774cdf8eeb99,0x34b0bcb5e19b48a8
+	.quad	0x391c0cb3c5c95a63,0x4ed8aa4ae3418acb
+	.quad	0x5b9cca4f7763e373,0x682e6ff3d6b2b8a3
+	.quad	0x748f82ee5defb2fc,0x78a5636f43172f60
+	.quad	0x84c87814a1f0ab72,0x8cc702081a6439ec
+	.quad	0x90befffa23631e28,0xa4506cebde82bde9
+	.quad	0xbef9a3f7b2c67915,0xc67178f2e372532b
+	.quad	0xca273eceea26619c,0xd186b8c721c0c207
+	.quad	0xeada7dd6cde0eb1e,0xf57d4f7fee6ed178
+	.quad	0x06f067aa72176fba,0x0a637dc5a2c898a6
+	.quad	0x113f9804bef90dae,0x1b710b35131c471b
+	.quad	0x28db77f523047d84,0x32caab7b40c72493
+	.quad	0x3c9ebe0a15c9bebc,0x431d67c49c100d4c
+	.quad	0x4cc5d4becb3e42b6,0x597f299cfc657e2a
+	.quad	0x5fcb6fab3ad6faec,0x6c44198c4a475817
+___
+}
+
+$code =~ s/\`([^\`]*)\`/eval $1/gem;
+print $code;
+close STDOUT;

crypto/sha/sha1_one.c

@@ -61,7 +61,7 @@
 #include <openssl/sha.h>
 #include <openssl/crypto.h>
 
-#if !defined(OPENSSL_NO_SHA1) && !defined(OPENSSL_FIPS)
+#if !defined(OPENSSL_NO_SHA1)
 unsigned char *SHA1(const unsigned char *d, size_t n, unsigned char *md)
 	{
 	SHA_CTX c;

crypto/sha/sha1dgst.c

@@ -63,19 +63,13 @@
 #define SHA_1
 
 #include <openssl/opensslv.h>
+#include <openssl/fips.h>
 
-#ifndef OPENSSL_FIPS
 const char SHA1_version[]="SHA1" OPENSSL_VERSION_PTEXT;
 
 /* The implementation is in ../md32_common.h */
 
 #include "sha_locl.h"
 
-#else
-
-static void *dummy=&dummy;
-
-#endif
-
 #endif
 

crypto/sha/sha256.c

@@ -15,12 +15,13 @@
 #include <openssl/fips.h>
 #include <openssl/opensslv.h>
 
-#ifndef OPENSSL_FIPS
-
 const char SHA256_version[]="SHA-256" OPENSSL_VERSION_PTEXT;
 
 int SHA224_Init (SHA256_CTX *c)
 	{
+#ifdef OPENSSL_FIPS
+	FIPS_selftest_check();
+#endif
 	c->h[0]=0xc1059ed8UL;	c->h[1]=0x367cd507UL;
 	c->h[2]=0x3070dd17UL;	c->h[3]=0xf70e5939UL;
 	c->h[4]=0xffc00b31UL;	c->h[5]=0x68581511UL;
@@ -32,6 +33,9 @@ int SHA224_Init (SHA256_CTX *c)
 
 int SHA256_Init (SHA256_CTX *c)
 	{
+#ifdef OPENSSL_FIPS
+	FIPS_selftest_check();
+#endif
 	c->h[0]=0x6a09e667UL;	c->h[1]=0xbb67ae85UL;
 	c->h[2]=0x3c6ef372UL;	c->h[3]=0xa54ff53aUL;
 	c->h[4]=0x510e527fUL;	c->h[5]=0x9b05688cUL;
@@ -72,17 +76,11 @@ int SHA224_Update(SHA256_CTX *c, const void *data, size_t len)
 int SHA224_Final (unsigned char *md, SHA256_CTX *c)
 {   return SHA256_Final (md,c);   }
 
-#ifndef	SHA_LONG_LOG2
-#define	SHA_LONG_LOG2	2	/* default to 32 bits */
-#endif
-
 #define	DATA_ORDER_IS_BIG_ENDIAN
 
 #define	HASH_LONG		SHA_LONG
-#define	HASH_LONG_LOG2		SHA_LONG_LOG2
 #define	HASH_CTX		SHA256_CTX
 #define	HASH_CBLOCK		SHA_CBLOCK
-#define	HASH_LBLOCK		SHA_LBLOCK
 /*
  * Note that FIPS180-2 discusses "Truncation of the Hash Function Output."
  * default: case below covers for it. It's not clear however if it's
@@ -115,16 +113,15 @@ int SHA224_Final (unsigned char *md, SHA256_CTX *c)
 #define	HASH_UPDATE		SHA256_Update
 #define	HASH_TRANSFORM		SHA256_Transform
 #define	HASH_FINAL		SHA256_Final
-#define	HASH_BLOCK_HOST_ORDER	sha256_block_host_order
 #define	HASH_BLOCK_DATA_ORDER	sha256_block_data_order
-void sha256_block_host_order (SHA256_CTX *ctx, const void *in, size_t num);
+#ifndef SHA256_ASM
+static
+#endif
 void sha256_block_data_order (SHA256_CTX *ctx, const void *in, size_t num);
 
 #include "md32_common.h"
 
-#ifdef SHA256_ASM
+#ifndef SHA256_ASM
-void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host);
-#else
 static const SHA_LONG K256[64] = {
 	0x428a2f98UL,0x71374491UL,0xb5c0fbcfUL,0xe9b5dba5UL,
 	0x3956c25bUL,0x59f111f1UL,0x923f82a4UL,0xab1c5ed5UL,
@@ -158,10 +155,10 @@ static const SHA_LONG K256[64] = {
 
 #ifdef OPENSSL_SMALL_FOOTPRINT
 
-static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host)
+static void sha256_block_data_order (SHA256_CTX *ctx, const void *in, size_t num)
 	{
 	unsigned MD32_REG_T a,b,c,d,e,f,g,h,s0,s1,T1,T2;
-	SHA_LONG	X[16];
+	SHA_LONG	X[16],l;
 	int i;
 	const unsigned char *data=in;
 
@@ -170,25 +167,6 @@ static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host)
 	a = ctx->h[0];	b = ctx->h[1];	c = ctx->h[2];	d = ctx->h[3];
 	e = ctx->h[4];	f = ctx->h[5];	g = ctx->h[6];	h = ctx->h[7];
 
-	if (host)
-		{
-		const SHA_LONG *W=(const SHA_LONG *)data;
-
-		for (i=0;i<16;i++)
-			{
-			T1 = X[i] = W[i];
-			T1 += h + Sigma1(e) + Ch(e,f,g) + K256[i];
-			T2 = Sigma0(a) + Maj(a,b,c);
-			h = g;	g = f;	f = e;	e = d + T1;
-			d = c;	c = b;	b = a;	a = T1 + T2;
-			}
-
-		data += SHA256_CBLOCK;
-		}
-	else
-		{
-		SHA_LONG l;
-
 	for (i=0;i<16;i++)
 		{
 		HOST_c2l(data,l); T1 = X[i] = l;
@@ -197,7 +175,6 @@ static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host)
 		h = g;	g = f;	f = e;	e = d + T1;
 		d = c;	c = b;	b = a;	a = T1 + T2;
 		}
-		}
 
 	for (;i<64;i++)
 		{
@@ -230,19 +207,20 @@ static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host)
 	T1 = X[(i)&0x0f] += s0 + s1 + X[(i+9)&0x0f];	\
 	ROUND_00_15(i,a,b,c,d,e,f,g,h);		} while (0)
 
-static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host)
+static void sha256_block_data_order (SHA256_CTX *ctx, const void *in, size_t num)
 	{
 	unsigned MD32_REG_T a,b,c,d,e,f,g,h,s0,s1,T1;
 	SHA_LONG	X[16];
 	int i;
 	const unsigned char *data=in;
+	const union { long one; char little; } is_endian = {1};
 
 			while (num--) {
 
 	a = ctx->h[0];	b = ctx->h[1];	c = ctx->h[2];	d = ctx->h[3];
 	e = ctx->h[4];	f = ctx->h[5];	g = ctx->h[6];	h = ctx->h[7];
 
-	if (host)
+	if (!is_endian.little && sizeof(SHA_LONG)==4 && ((size_t)in%4)==0)
 		{
 		const SHA_LONG *W=(const SHA_LONG *)data;
 
@@ -308,17 +286,4 @@ static void sha256_block (SHA256_CTX *ctx, const void *in, size_t num, int host)
 #endif
 #endif /* SHA256_ASM */
 
-/*
- * Idea is to trade couple of cycles for some space. On IA-32 we save
- * about 4K in "big footprint" case. In "small footprint" case any gain
- * is appreciated:-)
- */
-void HASH_BLOCK_HOST_ORDER (SHA256_CTX *ctx, const void *in, size_t num)
-{   sha256_block (ctx,in,num,1);   }
-
-void HASH_BLOCK_DATA_ORDER (SHA256_CTX *ctx, const void *in, size_t num)
-{   sha256_block (ctx,in,num,0);   }
-
 #endif /* OPENSSL_NO_SHA256 */
-
-#endif

crypto/sha/sha512.c

@@ -6,7 +6,7 @@
  */
 #include <openssl/opensslconf.h>
 #include <openssl/fips.h>
-#if !defined(OPENSSL_NO_SHA) && !defined(OPENSSL_NO_SHA512) && !defined(OPENSSL_FIPS)
+#if !defined(OPENSSL_NO_SHA) && !defined(OPENSSL_NO_SHA512)
 /*
  * IMPLEMENTATION NOTES.
  *
@@ -53,12 +53,18 @@
 
 const char SHA512_version[]="SHA-512" OPENSSL_VERSION_PTEXT;
 
-#if defined(_M_IX86) || defined(_M_AMD64) || defined(__i386) || defined(__x86_64)
+#if defined(__i386) || defined(__i386__) || defined(_M_IX86) || \
+    defined(__x86_64) || defined(_M_AMD64) || defined(_M_X64) || \
+    defined(__s390__) || defined(__s390x__) || \
+    defined(SHA512_ASM)
 #define SHA512_BLOCK_CAN_MANAGE_UNALIGNED_DATA
 #endif
 
 int SHA384_Init (SHA512_CTX *c)
 	{
+#ifdef OPENSSL_FIPS
+	FIPS_selftest_check();
+#endif
 	c->h[0]=U64(0xcbbb9d5dc1059ed8);
 	c->h[1]=U64(0x629a292a367cd507);
 	c->h[2]=U64(0x9159015a3070dd17);
@@ -74,6 +80,9 @@ int SHA384_Init (SHA512_CTX *c)
 
 int SHA512_Init (SHA512_CTX *c)
 	{
+#ifdef OPENSSL_FIPS
+	FIPS_selftest_check();
+#endif
 	c->h[0]=U64(0x6a09e667f3bcc908);
 	c->h[1]=U64(0xbb67ae8584caa73b);
 	c->h[2]=U64(0x3c6ef372fe94f82b);
@@ -90,7 +99,7 @@ int SHA512_Init (SHA512_CTX *c)
 #ifndef SHA512_ASM
 static
 #endif
-void sha512_block (SHA512_CTX *ctx, const void *in, size_t num);
+void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num);
 
 int SHA512_Final (unsigned char *md, SHA512_CTX *c)
 	{
@@ -101,7 +110,7 @@ int SHA512_Final (unsigned char *md, SHA512_CTX *c)
 	n++;
 	if (n > (sizeof(c->u)-16))
 		memset (p+n,0,sizeof(c->u)-n), n=0,
-		sha512_block (c,p,1);
+		sha512_block_data_order (c,p,1);
 
 	memset (p+n,0,sizeof(c->u)-16-n);
 #ifdef	B_ENDIAN
@@ -126,7 +135,7 @@ int SHA512_Final (unsigned char *md, SHA512_CTX *c)
 	p[sizeof(c->u)-16] = (unsigned char)(c->Nh>>56);
 #endif
 
-	sha512_block (c,p,1);
+	sha512_block_data_order (c,p,1);
 
 	if (md==0) return 0;
 
@@ -198,7 +207,7 @@ int SHA512_Update (SHA512_CTX *c, const void *_data, size_t len)
 		else	{
 			memcpy (p+c->num,data,n), c->num = 0;
 			len-=n, data+=n;
-			sha512_block (c,p,1);
+			sha512_block_data_order (c,p,1);
 			}
 		}
 
@@ -208,12 +217,12 @@ int SHA512_Update (SHA512_CTX *c, const void *_data, size_t len)
 		if ((size_t)data%sizeof(c->u.d[0]) != 0)
 			while (len >= sizeof(c->u))
 				memcpy (p,data,sizeof(c->u)),
-				sha512_block (c,p,1),
+				sha512_block_data_order (c,p,1),
 				len  -= sizeof(c->u),
 				data += sizeof(c->u);
 		else
 #endif
-			sha512_block (c,data,len/sizeof(c->u)),
+			sha512_block_data_order (c,data,len/sizeof(c->u)),
 			data += len,
 			len  %= sizeof(c->u),
 			data -= len;
@@ -228,7 +237,7 @@ int SHA384_Update (SHA512_CTX *c, const void *data, size_t len)
 {   return SHA512_Update (c,data,len);   }
 
 void SHA512_Transform (SHA512_CTX *c, const unsigned char *data)
-{   sha512_block (c,data,1);  }
+{   sha512_block_data_order (c,data,1);  }
 
 unsigned char *SHA384(const unsigned char *d, size_t n, unsigned char *md)
 	{
@@ -302,11 +311,70 @@ static const SHA_LONG64 K512[80] = {
 #ifndef PEDANTIC
 # if defined(__GNUC__) && __GNUC__>=2 && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM)
 #  if defined(__x86_64) || defined(__x86_64__)
+#   define ROTR(a,n)	({ unsigned long ret;		\
+				asm ("rorq %1,%0"	\
+				: "=r"(ret)		\
+				: "J"(n),"0"(a)		\
+				: "cc"); ret;		})
+#   if !defined(B_ENDIAN)
 #    define PULL64(x) ({ SHA_LONG64 ret=*((const SHA_LONG64 *)(&(x)));	\
 				asm ("bswapq	%0"		\
 				: "=r"(ret)			\
 				: "0"(ret)); ret;		})
 #   endif
+#  elif (defined(__i386) || defined(__i386__)) && !defined(B_ENDIAN)
+#   if defined(I386_ONLY)
+#    define PULL64(x) ({ const unsigned int *p=(const unsigned int *)(&(x));\
+			unsigned int hi,lo;			\
+				asm("xchgb %%ah,%%al;xchgb %%dh,%%dl;"\
+				    "roll $16,%%eax; roll $16,%%edx; "\
+				    "xchgb %%ah,%%al;xchgb %%dh,%%dl;" \
+				: "=a"(lo),"=d"(hi)		\
+				: "0"(p[1]),"1"(p[0]) : "cc");	\
+				((SHA_LONG64)hi)<<32|lo;	})
+#   else
+#    define PULL64(x) ({ const unsigned int *p=(const unsigned int *)(&(x));\
+			unsigned int hi,lo;			\
+				asm ("bswapl %0; bswapl %1;"	\
+				: "=r"(lo),"=r"(hi)		\
+				: "0"(p[1]),"1"(p[0]));		\
+				((SHA_LONG64)hi)<<32|lo;	})
+#   endif
+#  elif (defined(_ARCH_PPC) && defined(__64BIT__)) || defined(_ARCH_PPC64)
+#   define ROTR(a,n)	({ unsigned long ret;		\
+				asm ("rotrdi %0,%1,%2"	\
+				: "=r"(ret)		\
+				: "r"(a),"K"(n)); ret;	})
+#  endif
+# elif defined(_MSC_VER)
+#  if defined(_WIN64)	/* applies to both IA-64 and AMD64 */
+#   define ROTR(a,n)	_rotr64((a),n)
+#  endif
+#  if defined(_M_IX86) && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM)
+#   if defined(I386_ONLY)
+    static SHA_LONG64 __fastcall __pull64be(const void *x)
+    {	_asm	mov	edx, [ecx + 0]
+	_asm	mov	eax, [ecx + 4]
+	_asm	xchg	dh,dl
+	_asm	xchg	ah,al
+	_asm	rol	edx,16
+	_asm	rol	eax,16
+	_asm	xchg	dh,dl
+	_asm	xchg	ah,al
+    }
+#   else
+    static SHA_LONG64 __fastcall __pull64be(const void *x)
+    {	_asm	mov	edx, [ecx + 0]
+	_asm	mov	eax, [ecx + 4]
+	_asm	bswap	edx
+	_asm	bswap	eax
+    }
+#   endif
+#   define PULL64(x) __pull64be(&(x))
+#   if _MSC_VER<=1200
+#    pragma inline_depth(0)
+#   endif
+#  endif
 # endif
 #endif
 
@@ -315,27 +383,6 @@ static const SHA_LONG64 K512[80] = {
 #define PULL64(x) (B(x,0)|B(x,1)|B(x,2)|B(x,3)|B(x,4)|B(x,5)|B(x,6)|B(x,7))
 #endif
 
-#ifndef PEDANTIC
-# if defined(_MSC_VER)
-#  if defined(_WIN64)	/* applies to both IA-64 and AMD64 */
-#   define ROTR(a,n)	_rotr64((a),n)
-#  endif
-# elif defined(__GNUC__) && __GNUC__>=2 && !defined(OPENSSL_NO_ASM) && !defined(OPENSSL_NO_INLINE_ASM)
-#  if defined(__x86_64) || defined(__x86_64__)
-#   define ROTR(a,n)	({ unsigned long ret;		\
-				asm ("rorq %1,%0"	\
-				: "=r"(ret)		\
-				: "J"(n),"0"(a)		\
-				: "cc"); ret;		})
-#  elif defined(_ARCH_PPC) && defined(__64BIT__)
-#   define ROTR(a,n)	({ unsigned long ret;		\
-				asm ("rotrdi %0,%1,%2"	\
-				: "=r"(ret)		\
-				: "r"(a),"K"(n)); ret;	})
-#  endif
-# endif
-#endif
-
 #ifndef ROTR
 #define ROTR(x,s)	(((x)>>s) | (x)<<(64-s))
 #endif
@@ -358,7 +405,7 @@ static const SHA_LONG64 K512[80] = {
 
 #ifdef OPENSSL_SMALL_FOOTPRINT
 
-static void sha512_block (SHA512_CTX *ctx, const void *in, size_t num)
+static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num)
 	{
 	const SHA_LONG64 *W=in;
 	SHA_LONG64	a,b,c,d,e,f,g,h,s0,s1,T1,T2;
@@ -419,7 +466,7 @@ static void sha512_block (SHA512_CTX *ctx, const void *in, size_t num)
 	T1 = X[(i)&0x0f] += s0 + s1 + X[(i+9)&0x0f];	\
 	ROUND_00_15(i,a,b,c,d,e,f,g,h);		} while (0)
 
-static void sha512_block (SHA512_CTX *ctx, const void *in, size_t num)
+static void sha512_block_data_order (SHA512_CTX *ctx, const void *in, size_t num)
 	{
 	const SHA_LONG64 *W=in;
 	SHA_LONG64	a,b,c,d,e,f,g,h,s0,s1,T1;

crypto/sha/sha_locl.h

@@ -62,17 +62,11 @@
 #include <openssl/opensslconf.h>
 #include <openssl/sha.h>
 
-#ifndef SHA_LONG_LOG2
-#define SHA_LONG_LOG2	2	/* default to 32 bits */
-#endif
-
 #define DATA_ORDER_IS_BIG_ENDIAN
 
 #define HASH_LONG               SHA_LONG
-#define HASH_LONG_LOG2          SHA_LONG_LOG2
 #define HASH_CTX                SHA_CTX
 #define HASH_CBLOCK             SHA_CBLOCK
-#define HASH_LBLOCK             SHA_LBLOCK
 #define HASH_MAKE_STRING(c,s)   do {	\
 	unsigned long ll;		\
 	ll=(c)->h0; HOST_l2c(ll,(s));	\
@@ -88,12 +82,10 @@
 # define HASH_TRANSFORM          	SHA_Transform
 # define HASH_FINAL              	SHA_Final
 # define HASH_INIT			SHA_Init
-# define HASH_BLOCK_HOST_ORDER   	sha_block_host_order
 # define HASH_BLOCK_DATA_ORDER   	sha_block_data_order
 # define Xupdate(a,ix,ia,ib,ic,id)	(ix=(a)=(ia^ib^ic^id))
 
-  void sha_block_host_order (SHA_CTX *c, const void *p,size_t num);
+static void sha_block_data_order (SHA_CTX *c, const void *p,size_t num);
-  void sha_block_data_order (SHA_CTX *c, const void *p,size_t num);
 
 #elif defined(SHA_1)
 
@@ -101,7 +93,6 @@
 # define HASH_TRANSFORM          	SHA1_Transform
 # define HASH_FINAL              	SHA1_Final
 # define HASH_INIT			SHA1_Init
-# define HASH_BLOCK_HOST_ORDER   	sha1_block_host_order
 # define HASH_BLOCK_DATA_ORDER   	sha1_block_data_order
 # if defined(__MWERKS__) && defined(__MC68K__)
    /* Metrowerks for Motorola fails otherwise:-( <appro@fy.chalmers.se> */
@@ -114,24 +105,10 @@
 					)
 # endif
 
-# ifdef SHA1_ASM
+#ifndef SHA1_ASM
-#  if defined(__i386) || defined(__i386__) || defined(_M_IX86) || defined(__INTEL__)
+static
-#   if !defined(B_ENDIAN)
+#endif
-#    define sha1_block_host_order		sha1_block_asm_host_order
+void sha1_block_data_order (SHA_CTX *c, const void *p,size_t num);
-#    define DONT_IMPLEMENT_BLOCK_HOST_ORDER
-#    define sha1_block_data_order		sha1_block_asm_data_order
-#    define DONT_IMPLEMENT_BLOCK_DATA_ORDER
-#    define HASH_BLOCK_DATA_ORDER_ALIGNED	sha1_block_asm_data_order
-#   endif
-#  elif defined(__ia64) || defined(__ia64__) || defined(_M_IA64)
-#   define sha1_block_host_order		sha1_block_asm_host_order
-#   define DONT_IMPLEMENT_BLOCK_HOST_ORDER
-#   define sha1_block_data_order		sha1_block_asm_data_order
-#   define DONT_IMPLEMENT_BLOCK_DATA_ORDER
-#  endif
-# endif
-  void sha1_block_host_order (SHA_CTX *c, const void *p,size_t num);
-  void sha1_block_data_order (SHA_CTX *c, const void *p,size_t num);
 
 #else
 # error "Either SHA_0 or SHA_1 must be defined."
@@ -151,6 +128,9 @@ FIPS_NON_FIPS_MD_Init(SHA)
 int HASH_INIT (SHA_CTX *c)
 #endif
 	{
+#if defined(SHA_1) && defined(OPENSSL_FIPS)
+	FIPS_selftest_check();
+#endif
 	c->h0=INIT_DATA_h0;
 	c->h1=INIT_DATA_h1;
 	c->h2=INIT_DATA_h2;
@@ -233,133 +213,8 @@ int HASH_INIT (SHA_CTX *c)
 # define X(i)	XX[i]
 #endif
 
-#ifndef DONT_IMPLEMENT_BLOCK_HOST_ORDER
+#if !defined(SHA_1) || !defined(SHA1_ASM)
-void HASH_BLOCK_HOST_ORDER (SHA_CTX *c, const void *d, size_t num)
+static void HASH_BLOCK_DATA_ORDER (SHA_CTX *c, const void *p, size_t num)
-	{
-	const SHA_LONG *W=d;
-	register unsigned MD32_REG_T A,B,C,D,E,T;
-#ifndef MD32_XARRAY
-	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
-
-	A=c->h0;
-	B=c->h1;
-	C=c->h2;
-	D=c->h3;
-	E=c->h4;
-
-	for (;;)
-		{
-	BODY_00_15( 0,A,B,C,D,E,T,W[ 0]);
-	BODY_00_15( 1,T,A,B,C,D,E,W[ 1]);
-	BODY_00_15( 2,E,T,A,B,C,D,W[ 2]);
-	BODY_00_15( 3,D,E,T,A,B,C,W[ 3]);
-	BODY_00_15( 4,C,D,E,T,A,B,W[ 4]);
-	BODY_00_15( 5,B,C,D,E,T,A,W[ 5]);
-	BODY_00_15( 6,A,B,C,D,E,T,W[ 6]);
-	BODY_00_15( 7,T,A,B,C,D,E,W[ 7]);
-	BODY_00_15( 8,E,T,A,B,C,D,W[ 8]);
-	BODY_00_15( 9,D,E,T,A,B,C,W[ 9]);
-	BODY_00_15(10,C,D,E,T,A,B,W[10]);
-	BODY_00_15(11,B,C,D,E,T,A,W[11]);
-	BODY_00_15(12,A,B,C,D,E,T,W[12]);
-	BODY_00_15(13,T,A,B,C,D,E,W[13]);
-	BODY_00_15(14,E,T,A,B,C,D,W[14]);
-	BODY_00_15(15,D,E,T,A,B,C,W[15]);
-
-	BODY_16_19(16,C,D,E,T,A,B,X( 0),W[ 0],W[ 2],W[ 8],W[13]);
-	BODY_16_19(17,B,C,D,E,T,A,X( 1),W[ 1],W[ 3],W[ 9],W[14]);
-	BODY_16_19(18,A,B,C,D,E,T,X( 2),W[ 2],W[ 4],W[10],W[15]);
-	BODY_16_19(19,T,A,B,C,D,E,X( 3),W[ 3],W[ 5],W[11],X( 0));
-
-	BODY_20_31(20,E,T,A,B,C,D,X( 4),W[ 4],W[ 6],W[12],X( 1));
-	BODY_20_31(21,D,E,T,A,B,C,X( 5),W[ 5],W[ 7],W[13],X( 2));
-	BODY_20_31(22,C,D,E,T,A,B,X( 6),W[ 6],W[ 8],W[14],X( 3));
-	BODY_20_31(23,B,C,D,E,T,A,X( 7),W[ 7],W[ 9],W[15],X( 4));
-	BODY_20_31(24,A,B,C,D,E,T,X( 8),W[ 8],W[10],X( 0),X( 5));
-	BODY_20_31(25,T,A,B,C,D,E,X( 9),W[ 9],W[11],X( 1),X( 6));
-	BODY_20_31(26,E,T,A,B,C,D,X(10),W[10],W[12],X( 2),X( 7));
-	BODY_20_31(27,D,E,T,A,B,C,X(11),W[11],W[13],X( 3),X( 8));
-	BODY_20_31(28,C,D,E,T,A,B,X(12),W[12],W[14],X( 4),X( 9));
-	BODY_20_31(29,B,C,D,E,T,A,X(13),W[13],W[15],X( 5),X(10));
-	BODY_20_31(30,A,B,C,D,E,T,X(14),W[14],X( 0),X( 6),X(11));
-	BODY_20_31(31,T,A,B,C,D,E,X(15),W[15],X( 1),X( 7),X(12));
-
-	BODY_32_39(32,E,T,A,B,C,D,X( 0),X( 2),X( 8),X(13));
-	BODY_32_39(33,D,E,T,A,B,C,X( 1),X( 3),X( 9),X(14));
-	BODY_32_39(34,C,D,E,T,A,B,X( 2),X( 4),X(10),X(15));
-	BODY_32_39(35,B,C,D,E,T,A,X( 3),X( 5),X(11),X( 0));
-	BODY_32_39(36,A,B,C,D,E,T,X( 4),X( 6),X(12),X( 1));
-	BODY_32_39(37,T,A,B,C,D,E,X( 5),X( 7),X(13),X( 2));
-	BODY_32_39(38,E,T,A,B,C,D,X( 6),X( 8),X(14),X( 3));
-	BODY_32_39(39,D,E,T,A,B,C,X( 7),X( 9),X(15),X( 4));
-
-	BODY_40_59(40,C,D,E,T,A,B,X( 8),X(10),X( 0),X( 5));
-	BODY_40_59(41,B,C,D,E,T,A,X( 9),X(11),X( 1),X( 6));
-	BODY_40_59(42,A,B,C,D,E,T,X(10),X(12),X( 2),X( 7));
-	BODY_40_59(43,T,A,B,C,D,E,X(11),X(13),X( 3),X( 8));
-	BODY_40_59(44,E,T,A,B,C,D,X(12),X(14),X( 4),X( 9));
-	BODY_40_59(45,D,E,T,A,B,C,X(13),X(15),X( 5),X(10));
-	BODY_40_59(46,C,D,E,T,A,B,X(14),X( 0),X( 6),X(11));
-	BODY_40_59(47,B,C,D,E,T,A,X(15),X( 1),X( 7),X(12));
-	BODY_40_59(48,A,B,C,D,E,T,X( 0),X( 2),X( 8),X(13));
-	BODY_40_59(49,T,A,B,C,D,E,X( 1),X( 3),X( 9),X(14));
-	BODY_40_59(50,E,T,A,B,C,D,X( 2),X( 4),X(10),X(15));
-	BODY_40_59(51,D,E,T,A,B,C,X( 3),X( 5),X(11),X( 0));
-	BODY_40_59(52,C,D,E,T,A,B,X( 4),X( 6),X(12),X( 1));
-	BODY_40_59(53,B,C,D,E,T,A,X( 5),X( 7),X(13),X( 2));
-	BODY_40_59(54,A,B,C,D,E,T,X( 6),X( 8),X(14),X( 3));
-	BODY_40_59(55,T,A,B,C,D,E,X( 7),X( 9),X(15),X( 4));
-	BODY_40_59(56,E,T,A,B,C,D,X( 8),X(10),X( 0),X( 5));
-	BODY_40_59(57,D,E,T,A,B,C,X( 9),X(11),X( 1),X( 6));
-	BODY_40_59(58,C,D,E,T,A,B,X(10),X(12),X( 2),X( 7));
-	BODY_40_59(59,B,C,D,E,T,A,X(11),X(13),X( 3),X( 8));
-
-	BODY_60_79(60,A,B,C,D,E,T,X(12),X(14),X( 4),X( 9));
-	BODY_60_79(61,T,A,B,C,D,E,X(13),X(15),X( 5),X(10));
-	BODY_60_79(62,E,T,A,B,C,D,X(14),X( 0),X( 6),X(11));
-	BODY_60_79(63,D,E,T,A,B,C,X(15),X( 1),X( 7),X(12));
-	BODY_60_79(64,C,D,E,T,A,B,X( 0),X( 2),X( 8),X(13));
-	BODY_60_79(65,B,C,D,E,T,A,X( 1),X( 3),X( 9),X(14));
-	BODY_60_79(66,A,B,C,D,E,T,X( 2),X( 4),X(10),X(15));
-	BODY_60_79(67,T,A,B,C,D,E,X( 3),X( 5),X(11),X( 0));
-	BODY_60_79(68,E,T,A,B,C,D,X( 4),X( 6),X(12),X( 1));
-	BODY_60_79(69,D,E,T,A,B,C,X( 5),X( 7),X(13),X( 2));
-	BODY_60_79(70,C,D,E,T,A,B,X( 6),X( 8),X(14),X( 3));
-	BODY_60_79(71,B,C,D,E,T,A,X( 7),X( 9),X(15),X( 4));
-	BODY_60_79(72,A,B,C,D,E,T,X( 8),X(10),X( 0),X( 5));
-	BODY_60_79(73,T,A,B,C,D,E,X( 9),X(11),X( 1),X( 6));
-	BODY_60_79(74,E,T,A,B,C,D,X(10),X(12),X( 2),X( 7));
-	BODY_60_79(75,D,E,T,A,B,C,X(11),X(13),X( 3),X( 8));
-	BODY_60_79(76,C,D,E,T,A,B,X(12),X(14),X( 4),X( 9));
-	BODY_60_79(77,B,C,D,E,T,A,X(13),X(15),X( 5),X(10));
-	BODY_60_79(78,A,B,C,D,E,T,X(14),X( 0),X( 6),X(11));
-	BODY_60_79(79,T,A,B,C,D,E,X(15),X( 1),X( 7),X(12));
-	
-	c->h0=(c->h0+E)&0xffffffffL; 
-	c->h1=(c->h1+T)&0xffffffffL;
-	c->h2=(c->h2+A)&0xffffffffL;
-	c->h3=(c->h3+B)&0xffffffffL;
-	c->h4=(c->h4+C)&0xffffffffL;
-
-	if (--num == 0) break;
-
-	A=c->h0;
-	B=c->h1;
-	C=c->h2;
-	D=c->h3;
-	E=c->h4;
-
-	W+=SHA_LBLOCK;
-		}
-	}
-#endif
-
-#ifndef DONT_IMPLEMENT_BLOCK_DATA_ORDER
-void HASH_BLOCK_DATA_ORDER (SHA_CTX *c, const void *p, size_t num)
 	{
 	const unsigned char *data=p;
 	register unsigned MD32_REG_T A,B,C,D,E,T,l;
@@ -378,7 +233,34 @@ void HASH_BLOCK_DATA_ORDER (SHA_CTX *c, const void *p, size_t num)
 
 	for (;;)
 			{
+	const union { long one; char little; } is_endian = {1};
 
+	if (!is_endian.little && sizeof(SHA_LONG)==4 && ((size_t)p%4)==0)
+		{
+		const SHA_LONG *W=(const SHA_LONG *)data;
+
+		X( 0) = W[0];				X( 1) = W[ 1];
+		BODY_00_15( 0,A,B,C,D,E,T,X( 0));	X( 2) = W[ 2];
+		BODY_00_15( 1,T,A,B,C,D,E,X( 1));	X( 3) = W[ 3];
+		BODY_00_15( 2,E,T,A,B,C,D,X( 2));	X( 4) = W[ 4];
+		BODY_00_15( 3,D,E,T,A,B,C,X( 3));	X( 5) = W[ 5];
+		BODY_00_15( 4,C,D,E,T,A,B,X( 4));	X( 6) = W[ 6];
+		BODY_00_15( 5,B,C,D,E,T,A,X( 5));	X( 7) = W[ 7];
+		BODY_00_15( 6,A,B,C,D,E,T,X( 6));	X( 8) = W[ 8];
+		BODY_00_15( 7,T,A,B,C,D,E,X( 7));	X( 9) = W[ 9];
+		BODY_00_15( 8,E,T,A,B,C,D,X( 8));	X(10) = W[10];
+		BODY_00_15( 9,D,E,T,A,B,C,X( 9));	X(11) = W[11];
+		BODY_00_15(10,C,D,E,T,A,B,X(10));	X(12) = W[12];
+		BODY_00_15(11,B,C,D,E,T,A,X(11));	X(13) = W[13];
+		BODY_00_15(12,A,B,C,D,E,T,X(12));	X(14) = W[14];
+		BODY_00_15(13,T,A,B,C,D,E,X(13));	X(15) = W[15];
+		BODY_00_15(14,E,T,A,B,C,D,X(14));
+		BODY_00_15(15,D,E,T,A,B,C,X(15));
+
+		data += SHA_CBLOCK;
+		}
+	else
+		{
 		HOST_c2l(data,l); X( 0)=l;		HOST_c2l(data,l); X( 1)=l;
 		BODY_00_15( 0,A,B,C,D,E,T,X( 0));	HOST_c2l(data,l); X( 2)=l;
 		BODY_00_15( 1,T,A,B,C,D,E,X( 1));	HOST_c2l(data,l); X( 3)=l;
@@ -396,6 +278,7 @@ void HASH_BLOCK_DATA_ORDER (SHA_CTX *c, const void *p, size_t num)
 		BODY_00_15(13,T,A,B,C,D,E,X(13));	HOST_c2l(data,l); X(15)=l;
 		BODY_00_15(14,E,T,A,B,C,D,X(14));
 		BODY_00_15(15,D,E,T,A,B,C,X(15));
+		}
 
 	BODY_16_19(16,C,D,E,T,A,B,X( 0),X( 0),X( 2),X( 8),X(13));
 	BODY_16_19(17,B,C,D,E,T,A,X( 1),X( 1),X( 3),X( 9),X(14));
@@ -515,54 +398,8 @@ void HASH_BLOCK_DATA_ORDER (SHA_CTX *c, const void *p, size_t num)
 	E=D, D=C, C=ROTATE(B,30), B=A;	\
 	A=ROTATE(A,5)+T+xa;	    } while(0)
 
-#ifndef DONT_IMPLEMENT_BLOCK_HOST_ORDER
+#if !defined(SHA_1) || !defined(SHA1_ASM)
-void HASH_BLOCK_HOST_ORDER (SHA_CTX *c, const void *d, size_t num)
+static void HASH_BLOCK_DATA_ORDER (SHA_CTX *c, const void *p, size_t num)
-	{
-	const SHA_LONG *W=d;
-	register unsigned MD32_REG_T A,B,C,D,E,T;
-	int i;
-	SHA_LONG	X[16];
-
-	A=c->h0;
-	B=c->h1;
-	C=c->h2;
-	D=c->h3;
-	E=c->h4;
-
-	for (;;)
-		{
-	for (i=0;i<16;i++)
-	{ X[i]=W[i]; BODY_00_15(X[i]); }
-	for (i=0;i<4;i++)
-	{ BODY_16_19(X[i],       X[i+2],      X[i+8],     X[(i+13)&15]); }
-	for (;i<24;i++)
-	{ BODY_20_39(X[i&15],    X[(i+2)&15], X[(i+8)&15],X[(i+13)&15]); }
-	for (i=0;i<20;i++)
-	{ BODY_40_59(X[(i+8)&15],X[(i+10)&15],X[i&15],    X[(i+5)&15]);  }
-	for (i=4;i<24;i++)
-	{ BODY_60_79(X[(i+8)&15],X[(i+10)&15],X[i&15],    X[(i+5)&15]);  }
-	
-	c->h0=(c->h0+A)&0xffffffffL; 
-	c->h1=(c->h1+B)&0xffffffffL;
-	c->h2=(c->h2+C)&0xffffffffL;
-	c->h3=(c->h3+D)&0xffffffffL;
-	c->h4=(c->h4+E)&0xffffffffL;
-
-	if (--num == 0) break;
-
-	A=c->h0;
-	B=c->h1;
-	C=c->h2;
-	D=c->h3;
-	E=c->h4;
-
-	W+=SHA_LBLOCK;
-		}
-	}
-#endif
-
-#ifndef DONT_IMPLEMENT_BLOCK_DATA_ORDER
-void HASH_BLOCK_DATA_ORDER (SHA_CTX *c, const void *p, size_t num)
 	{
 	const unsigned char *data=p;
 	register unsigned MD32_REG_T A,B,C,D,E,T,l;

crypto/stack/safestack.h

@@ -520,28 +520,6 @@ STACK_OF(type) \
 #define sk_CRYPTO_dynlock_sort(st) SKM_sk_sort(CRYPTO_dynlock, (st))
 #define sk_CRYPTO_dynlock_is_sorted(st) SKM_sk_is_sorted(CRYPTO_dynlock, (st))
 
-#define sk_CRYPTO_dynlock_new(st) SKM_sk_new(CRYPTO_dynlock, (st))
-#define sk_CRYPTO_dynlock_new_null() SKM_sk_new_null(CRYPTO_dynlock)
-#define sk_CRYPTO_dynlock_free(st) SKM_sk_free(CRYPTO_dynlock, (st))
-#define sk_CRYPTO_dynlock_num(st) SKM_sk_num(CRYPTO_dynlock, (st))
-#define sk_CRYPTO_dynlock_value(st, i) SKM_sk_value(CRYPTO_dynlock, (st), (i))
-#define sk_CRYPTO_dynlock_set(st, i, val) SKM_sk_set(CRYPTO_dynlock, (st), (i), (val))
-#define sk_CRYPTO_dynlock_zero(st) SKM_sk_zero(CRYPTO_dynlock, (st))
-#define sk_CRYPTO_dynlock_push(st, val) SKM_sk_push(CRYPTO_dynlock, (st), (val))
-#define sk_CRYPTO_dynlock_unshift(st, val) SKM_sk_unshift(CRYPTO_dynlock, (st), (val))
-#define sk_CRYPTO_dynlock_find(st, val) SKM_sk_find(CRYPTO_dynlock, (st), (val))
-#define sk_CRYPTO_dynlock_find_ex(st, val) SKM_sk_find_ex(CRYPTO_dynlock, (st), (val))
-#define sk_CRYPTO_dynlock_delete(st, i) SKM_sk_delete(CRYPTO_dynlock, (st), (i))
-#define sk_CRYPTO_dynlock_delete_ptr(st, ptr) SKM_sk_delete_ptr(CRYPTO_dynlock, (st), (ptr))
-#define sk_CRYPTO_dynlock_insert(st, val, i) SKM_sk_insert(CRYPTO_dynlock, (st), (val), (i))
-#define sk_CRYPTO_dynlock_set_cmp_func(st, cmp) SKM_sk_set_cmp_func(CRYPTO_dynlock, (st), (cmp))
-#define sk_CRYPTO_dynlock_dup(st) SKM_sk_dup(CRYPTO_dynlock, st)
-#define sk_CRYPTO_dynlock_pop_free(st, free_func) SKM_sk_pop_free(CRYPTO_dynlock, (st), (free_func))
-#define sk_CRYPTO_dynlock_shift(st) SKM_sk_shift(CRYPTO_dynlock, (st))
-#define sk_CRYPTO_dynlock_pop(st) SKM_sk_pop(CRYPTO_dynlock, (st))
-#define sk_CRYPTO_dynlock_sort(st) SKM_sk_sort(CRYPTO_dynlock, (st))
-#define sk_CRYPTO_dynlock_is_sorted(st) SKM_sk_is_sorted(CRYPTO_dynlock, (st))
-
 #define sk_DIST_POINT_new(st) SKM_sk_new(DIST_POINT, (st))
 #define sk_DIST_POINT_new_null() SKM_sk_new_null(DIST_POINT)
 #define sk_DIST_POINT_free(st) SKM_sk_free(DIST_POINT, (st))

crypto/store/str_lib.c

@@ -236,7 +236,7 @@ const STORE_METHOD *STORE_set_method(STORE *store, const STORE_METHOD *meth)
 #define check_store(s,fncode,fnname,fnerrcode) \
 	do \
 		{ \
-		if ((s) == NULL || (s)->meth) \
+		if ((s) == NULL || (s)->meth == NULL) \
 			{ \
 			STOREerr((fncode), ERR_R_PASSED_NULL_PARAMETER); \
 			return 0; \

crypto/x86_64cpuid.pl

@@ -1,19 +1,12 @@
 #!/usr/bin/env perl
 
 $output=shift;
-$win64a=1 if ($output =~ /win64a\.[s|asm]/);
+$masm=1 if ($output =~ /\.asm/);
 open STDOUT,">$output" || die "can't open $output: $!";
 
-print<<___ if(defined($win64a));
+print<<___ if(defined($masm));
 _TEXT	SEGMENT
 PUBLIC	OPENSSL_rdtsc
-ALIGN	16
-OPENSSL_rdtsc	PROC
-	rdtsc
-	shl	rdx,32
-	or	rax,rdx
-	ret
-OPENSSL_rdtsc	ENDP
 
 PUBLIC	OPENSSL_atomic_add
 ALIGN	16
@@ -45,35 +38,16 @@ OPENSSL_wipe_cpu	PROC
 	lea	rax,QWORD PTR[rsp+8]
 	ret
 OPENSSL_wipe_cpu	ENDP
-
-OPENSSL_ia32_cpuid	PROC
-	mov	r8,rbx
-	mov	eax,1
-	cpuid
-	shl	rcx,32
-	mov	eax,edx
-	mov	rbx,r8
-	or	rax,rcx
-	ret
-OPENSSL_ia32_cpuid	ENDP
 _TEXT	ENDS
 
 CRT\$XIU	SEGMENT
 EXTRN	OPENSSL_cpuid_setup:PROC
 DQ	OPENSSL_cpuid_setup
 CRT\$XIU	ENDS
-END
+
 ___
-print<<___ if(!defined($win64a));
+print<<___ if(!defined($masm));
 .text
-.globl	OPENSSL_rdtsc
-.align	16
-OPENSSL_rdtsc:
-	rdtsc
-	shlq	\$32,%rdx
-	orq	%rdx,%rax
-	ret
-.size	OPENSSL_rdtsc,.-OPENSSL_rdtsc
 
 .globl	OPENSSL_atomic_add
 .type	OPENSSL_atomic_add,\@function
@@ -120,19 +94,66 @@ OPENSSL_wipe_cpu:
 	ret
 .size	OPENSSL_wipe_cpu,.-OPENSSL_wipe_cpu
 
-.globl	OPENSSL_ia32_cpuid
-.align	16
-OPENSSL_ia32_cpuid:
-	movq	%rbx,%r8
-	movl	\$1,%eax
-	cpuid
-	shlq	\$32,%rcx
-	movl	%edx,%eax
-	movq	%r8,%rbx
-	orq	%rcx,%rax
-	ret
-.size	OPENSSL_ia32_cpuid,.-OPENSSL_ia32_cpuid
-
 .section	.init
 	call	OPENSSL_cpuid_setup
+
 ___
+
+open STDOUT,"| $^X perlasm/x86_64-xlate.pl $output";
+print<<___;
+.text
+
+.globl	OPENSSL_rdtsc
+.type	OPENSSL_rdtsc,\@abi-omnipotent
+.align	16
+OPENSSL_rdtsc:
+	rdtsc
+	shl	\$32,%rdx
+	or	%rdx,%rax
+	ret
+.size	OPENSSL_rdtsc,.-OPENSSL_rdtsc
+
+.globl	OPENSSL_ia32_cpuid
+.type	OPENSSL_ia32_cpuid,\@abi-omnipotent
+.align	16
+OPENSSL_ia32_cpuid:
+	mov	%rbx,%r8
+
+	xor	%eax,%eax
+	cpuid
+	xor	%eax,%eax
+	cmp	\$0x756e6547,%ebx	# "Genu"
+	setne	%al
+	mov	%eax,%r9d
+	cmp	\$0x49656e69,%edx	# "ineI"
+	setne	%al
+	or	%eax,%r9d
+	cmp	\$0x6c65746e,%ecx	# "ntel"
+	setne	%al
+	or	%eax,%r9d
+
+	mov	\$1,%eax
+	cpuid
+	cmp	\$0,%r9d
+	jne	.Lnotintel
+	or	\$0x00100000,%edx	# use reserved 20th bit to engage RC4_CHAR
+	and	\$15,%ah
+	cmp	\$15,%ah		# examine Family ID
+	je	.Lnotintel
+	or	\$0x40000000,%edx	# use reserved 30th bit to skip unrolled loop
+.Lnotintel:
+	bt	\$28,%edx		# test hyper-threading bit
+	jnc	.Ldone
+	shr	\$16,%ebx
+	cmp	\$1,%bl			# see if cache is shared
+	ja	.Ldone
+	and	\$0xefffffff,%edx	# ~(1<<28)
+.Ldone:
+	shl	\$32,%rcx
+	mov	%edx,%eax
+	mov	%r8,%rbx
+	or	%rcx,%rax
+	ret
+.size	OPENSSL_ia32_cpuid,.-OPENSSL_ia32_cpuid
+___
+close STDOUT;	# flush

crypto/x86cpuid.pl

@@ -19,13 +19,40 @@ for (@ARGV) { $sse2=1 if (/-DOPENSSL_IA32_SSE2/); }
 	&pop	("eax");
 	&xor	("ecx","eax");
 	&bt	("ecx",21);
-	&jnc	(&label("nocpuid"));
+	&jnc	(&label("done"));
+	&xor	("eax","eax");
+	&cpuid	();
+	&xor	("eax","eax");
+	&cmp	("ebx",0x756e6547);	# "Genu"
+	&setne	(&LB("eax"));
+	&mov	("ebp","eax");
+	&cmp	("edx",0x49656e69);	# "ineI"
+	&setne	(&LB("eax"));
+	&or	("ebp","eax");
+	&cmp	("ecx",0x6c65746e);	# "ntel"
+	&setne	(&LB("eax"));
+	&or	("ebp","eax");
 	&mov	("eax",1);
 	&cpuid	();
-&set_label("nocpuid");
+	&cmp	("ebp",0);
+	&jne	(&label("notP4"));
+	&and	(&HB("eax"),15);	# familiy ID
+	&cmp	(&HB("eax"),15);	# P4?
+	&jne	(&label("notP4"));
+	&or	("edx",1<<20);		# use reserved bit to engage RC4_CHAR
+&set_label("notP4");
+	&bt	("edx",28);		# test hyper-threading bit
+	&jnc	(&label("done"));
+	&shr	("ebx",16);
+	&cmp	(&LB("ebx"),1);		# see if cache is shared(*)
+	&ja	(&label("done"));
+	&and	("edx",0xefffffff);	# clear hyper-threading bit if not
+&set_label("done");
 	&mov	("eax","edx");
 	&mov	("edx","ecx");
 &function_end("OPENSSL_ia32_cpuid");
+# (*)	on Core2 this value is set to 2 denoting the fact that L2
+#	cache is shared between cores.
 
 &external_label("OPENSSL_ia32cap_P");
 

doc/apps/enc.pod

@@ -227,6 +227,14 @@ Blowfish and RC5 algorithms use a 128 bit key.
  rc5-ecb            RC5 cipher in ECB mode
  rc5-ofb            RC5 cipher in OFB mode
 
+ aes-[128|192|256]-cbc	128/192/256 bit AES in CBC mode
+ aes-[128|192|256]	Alias for aes-[128|192|256]-cbc
+ aes-[128|192|256]-cfb	128/192/256 bit AES in 128 bit CFB mode
+ aes-[128|192|256]-cfb1	128/192/256 bit AES in 1 bit CFB mode
+ aes-[128|192|256]-cfb8	128/192/256 bit AES in 8 bit CFB mode
+ aes-[128|192|256]-ecb	128/192/256 bit AES in ECB mode
+ aes-[128|192|256]-ofb	128/192/256 bit AES in OFB mode
+
 =head1 EXAMPLES
 
 Just base64 encode a binary file:

doc/apps/openssl.pod

@@ -227,6 +227,22 @@ SHA Digest
 
 SHA-1 Digest
 
+=item B<sha224>
+
+SHA-224 Digest
+
+=item B<sha256>
+
+SHA-256 Digest
+
+=item B<sha384>
+
+SHA-384 Digest
+
+=item B<sha512>
+
+SHA-512 Digest
+
 =back
 
 =head2 ENCODING AND CIPHER COMMANDS

doc/apps/rsautl.pod

@@ -152,7 +152,7 @@ The final BIT STRING contains the actual signature. It can be extracted with:
 
 The certificate public key can be extracted with:
  
- openssl x509 -in test/testx509.pem -pubout -noout >pubkey.pem
+ openssl x509 -in test/testx509.pem -pubkey -noout >pubkey.pem
 
 The signature can be analysed with: