Add ARIA, BLAKE2 and SHA support for ARMv7, ARMv8 and Intel

pull/461/head
Jeffrey Walton 2017-07-29 23:40:56 -04:00
parent 61691dd906
commit 8338d90c10
No known key found for this signature in database
GPG Key ID: B36AB348921B1838
15 changed files with 2596 additions and 2493 deletions

View File

@ -11,6 +11,7 @@ algparam.h
arc4.cpp
arc4.h
aria.cpp
aria-simd.cpp
aria.h
argnames.h
asn.cpp
@ -29,6 +30,7 @@ bench1.cpp
bench2.cpp
bfinit.cpp
blake2.cpp
blake2-simd.cpp
blake2.h
blowfish.cpp
blowfish.h
@ -178,6 +180,7 @@ mqv.cpp
mqv.h
nbtheory.cpp
nbtheory.h
neon.cpp
network.cpp
network.h
nr.h
@ -248,8 +251,8 @@ seed.h
serpent.cpp
serpent.h
serpentp.h
sha-simd.cpp
sha.cpp
sha-simd.cpp
sha.h
sha3.cpp
sha3.h

66
GNUmakefile Executable file → Normal file
View File

@ -2,7 +2,7 @@
##### System Attributes and Programs #####
###########################################################
TEMPDIR ?= /tmp
TMPDIR ?= /tmp
AR ?= ar
ARFLAGS ?= -cr # ar needs the dash on OpenBSD
@ -24,6 +24,7 @@ IS_PPC := $(shell uname -m | $(EGREP) -i -c "ppc|power")
IS_ARM32 := $(shell uname -m | $(EGREP) -v "arm64" | $(EGREP) -i -c "arm")
IS_ARM64 := $(shell uname -m | $(EGREP) -i -c "aarch64")
IS_ARMV8 ?= $(shell uname -m | $(EGREP) -i -c 'aarch32|aarch64')
IS_NEON ?= $(shell uname -m | $(EGREP) -i -c 'armv7|armv8|aarch32|aarch64')
IS_SPARC := $(shell uname -m | $(EGREP) -i -c "sparc")
IS_SPARC64 := $(shell uname -m | $(EGREP) -i -c "sparc64")
@ -194,14 +195,16 @@ endif # -DCRYPTOPP_DISABLE_SSSE3
endif # -DCRYPTOPP_DISABLE_ASM
endif # CXXFLAGS
HAS_CRC := $(shell $(CXX) $(CXXFLAGS) -msse4.2 -o $(TEMPDIR)/t.o -c crc-simd.cpp; echo $$?)
ifeq ($(HAS_CRC),0)
CRC_FLAG := -msse4.2
ifeq ($(findstring -DCRYPTOPP_DISABLE_SSE4,$(CXXFLAGS)),)
SSE42_FLAG = $(shell echo | $(CXX) $(CXXFLAGS) -msse4.2 -dM -E - | grep -i -c -q __SSE4_2__ && echo "-msse4.2")
ifeq ($(findstring -DCRYPTOPP_DISABLE_AESNI,$(CXXFLAGS)),)
AES_FLAG = $(shell echo | $(CXX) $(CXXFLAGS) -maes -dM -E - | grep -i -c -q __AES__ && echo "-maes")
ifeq ($(findstring -DCRYPTOPP_DISABLE_SHA,$(CXXFLAGS)),)
SHA_FLAG = $(shell echo | $(CXX) $(CXXFLAGS) -msse4.2 -msha -dM -E - | grep -i -c -q __SHA__ && echo "-msse4.2 -msha")
BLAKE2_FLAG = $(SSE42_FLAG)
CRC_FLAG = $(SSE42_FLAG)
endif
endif
HAS_SHA := $(shell $(CXX) $(CXXFLAGS) -msse4.2 -msha -o $(TEMPDIR)/t.o -c sha.cpp; echo $$?)
ifeq ($(HAS_SHA),0)
SHA_FLAG := -msse4.2 -msha
endif
# BEGIN_NATIVE_ARCH
@ -293,15 +296,23 @@ CXXFLAGS += -pipe
endif
endif
ifeq ($(IS_NEON),1)
NEON_FLAG = $(shell echo | $(CXX) $(CXXFLAGS) -mfpu=neon -dM -E - | grep -i -c -q __SSE4_2__ && echo "-mfpu=neon")
GCM_FLAG = $(NEON_FLAG)
ARIA_FLAG = $(NEON_FLAG)
BLAKE2_FLAG = $(NEON_FLAG)
NEON_FLAG = $(NEON_FLAG)
endif
ifeq ($(IS_ARMV8),1)
HAS_CRC := $(shell $(CXX) $(CXXFLAGS) -march=armv8-a+crc -o $(TEMPDIR)/t.o -c crc-simd.cpp; echo $$?)
ifeq ($(HAS_CRC),0)
CRC_FLAG := -march=armv8-a+crc
endif
HAS_SHA := $(shell $(CXX) $(CXXFLAGS) -march=armv8-a+crc -o $(TEMPDIR)/t.o -c sha.cpp; echo $$?)
ifeq ($(HAS_SHA),0)
SHA_FLAG := -march=armv8-a+crypto
endif
ARMV8A_FLAG = $(shell echo | $(CXX) $(CXXFLAGS) -march=armv8-a -dM -E - | grep -i -c -q __ARM_NEON && echo "-march=armv8-a")
CRC_FLAG = $(shell echo | $(CXX) $(CXXFLAGS) -march=armv8-a+crc -dM -E - | grep -i -c -q __ARM_FEATURE_CRC32 && echo "-march=armv8-a+crc")
AES_FLAG = $(shell echo | $(CXX) $(CXXFLAGS) -march=armv8-a+crypto -dM -E - | grep -i -c -q __ARM_FEATURE_CRYPTO && echo "-march=armv8-a+crypto")
SHA_FLAG = $(shell echo | $(CXX) $(CXXFLAGS) -march=armv8-a+crypto -dM -E - | grep -i -c -q __ARM_FEATURE_CRYPTO && echo "-march=armv8-a+crypto")
GCM_FLAG = $(ARMV8A_FLAG)
ARIA_FLAG = $(ARMV8A_FLAG)
BLAKE2_FLAG = $(ARMV8A_FLAG)
NEON_FLAG = $(ARMV8A_FLAG)
endif
endif # IS_X86
@ -545,12 +556,13 @@ endif
endif # Nasm
# List test.cpp first to tame C++ static initialization problems.
TESTSRCS := adhoc.cpp test.cpp bench1.cpp bench2.cpp validat0.cpp validat1.cpp validat2.cpp validat3.cpp datatest.cpp regtest1.cpp regtest2.cpp regtest3.cpp fipsalgt.cpp dlltest.cpp
TESTSRCS := adhoc.cpp test.cpp bench1.cpp bench2.cpp validat0.cpp validat1.cpp validat2.cpp validat3.cpp datatest.cpp regtest1.cpp regtest2.cpp regtest3.cpp dlltest.cpp fipsalgt.cpp
TESTOBJS := $(TESTSRCS:.cpp=.o)
LIBOBJS := $(filter-out $(TESTOBJS),$(OBJS))
# List cryptlib.cpp first, then cpu.cpp, then integer.cpp to tame C++ static initialization problems.
DLLSRCS := cryptlib.cpp cpu.cpp integer.cpp shacal2.cpp md5.cpp shark.cpp zinflate.cpp gf2n.cpp salsa.cpp xtr.cpp oaep.cpp poly1305.cpp polynomi.cpp rc2.cpp default.cpp wait.cpp wake.cpp twofish.cpp iterhash.cpp adler32.cpp elgamal.cpp marss.cpp blowfish.cpp ecp.cpp filters.cpp strciphr.cpp camellia.cpp ida.cpp zlib.cpp des.cpp crc.cpp crc-simd.cpp algparam.cpp dessp.cpp tea.cpp eax.cpp network.cpp emsa2.cpp pkcspad.cpp squaretb.cpp idea.cpp authenc.cpp kalyna.cpp threefish.cpp hmac.cpp zdeflate.cpp xtrcrypt.cpp queue.cpp mars.cpp rc5.cpp blake2.cpp hrtimer.cpp eprecomp.cpp hex.cpp dsa.cpp sha.cpp fips140.cpp gzip.cpp seal.cpp files.cpp base32.cpp vmac.cpp tigertab.cpp sharkbox.cpp safer.cpp randpool.cpp esign.cpp arc4.cpp osrng.cpp skipjack.cpp seed.cpp sha3.cpp sosemanuk.cpp bfinit.cpp rabin.cpp 3way.cpp rw.cpp rdrand.cpp rsa.cpp rdtables.cpp gost.cpp socketft.cpp tftables.cpp nbtheory.cpp panama.cpp modes.cpp rijndael.cpp casts.cpp chacha.cpp gfpcrypt.cpp poly1305.cpp dll.cpp ec2n.cpp blumshub.cpp algebra.cpp basecode.cpp base64.cpp cbcmac.cpp rc6.cpp dh2.cpp gf256.cpp mqueue.cpp misc.cpp pssr.cpp channels.cpp tiger.cpp cast.cpp rng.cpp square.cpp asn.cpp whrlpool.cpp md4.cpp dh.cpp ccm.cpp md2.cpp mqv.cpp gf2_32.cpp ttmac.cpp luc.cpp trdlocal.cpp pubkey.cpp gcm.cpp ripemd.cpp eccrypto.cpp serpent.cpp cmac.cpp
DLLSRCS := cryptlib.cpp cpu.cpp integer.cpp 3way.cpp adler32.cpp algebra.cpp algparam.cpp arc4.cpp aria.cpp asn.cpp authenc.cpp base32.cpp base64.cpp basecode.cpp bfinit.cpp blake2-simd.cpp blake2.cpp blowfish.cpp blumshub.cpp camellia.cpp cast.cpp casts.cpp cbcmac.cpp ccm.cpp chacha.cpp channels.cpp cmac.cpp crc-simd.cpp crc.cpp default.cpp des.cpp dessp.cpp dh.cpp dh2.cpp dll.cpp dsa.cpp eax.cpp ec2n.cpp eccrypto.cpp ecp.cpp elgamal.cpp emsa2.cpp eprecomp.cpp esign.cpp files.cpp filters.cpp fips140.cpp fipstest.cpp gcm.cpp gf256.cpp gf2_32.cpp gf2n.cpp gfpcrypt.cpp gost.cpp gzip.cpp hex.cpp hmac.cpp hrtimer.cpp ida.cpp idea.cpp iterhash.cpp kalyna.cpp kalynatab.cpp keccak.cpp luc.cpp mars.cpp marss.cpp md2.cpp md4.cpp md5.cpp misc.cpp modes.cpp mqueue.cpp mqv.cpp nbtheory.cpp neon.cpp network.cpp oaep.cpp ospstore.cpp osrng.cpp panama.cpp pkcspad.cpp poly1305.cpp polynomi.cpp pssr.cpp pubkey.cpp queue.cpp rabin.cpp randpool.cpp rc2.cpp rc5.cpp rc6.cpp rdrand.cpp rdtables.cpp rijndael.cpp ripemd.cpp rng.cpp rsa.cpp rw.cpp safer.cpp salsa.cpp seal.cpp seed.cpp serpent.cpp sha-simd.cpp sha.cpp sha3.cpp shacal2.cpp shark.cpp sharkbox.cpp skipjack.cpp socketft.cpp sosemanuk.cpp square.cpp squaretb.cpp strciphr.cpp tea.cpp tftables.cpp threefish.cpp tiger.cpp tigertab.cpp trdlocal.cpp ttmac.cpp twofish.cpp vmac.cpp wait.cpp wake.cpp whrlpool.cpp xtr.cpp xtrcrypt.cpp zdeflate.cpp zinflate.cpp zlib.cpp
DLLOBJS := $(DLLSRCS:.cpp=.export.o)
# Import lib testing
@ -573,8 +585,8 @@ static: libcryptopp.a
shared dynamic: libcryptopp.so$(SOLIB_VERSION_SUFFIX)
endif
.PHONY: deps
deps GNUmakefile.deps:
.PHONY: dep deps depend
dep deps depend GNUmakefile.deps:
$(CXX) $(strip $(CXXFLAGS)) -MM *.cpp > GNUmakefile.deps
# CXXFLAGS are tuned earlier.
@ -841,11 +853,23 @@ rdrand-%.o:
./rdrand-nasm.sh
endif
# SSE4.2 or NEON available
aria-simd.o : aria-simd.cpp
$(CXX) $(strip $(CXXFLAGS) $(ARIA_FLAG) -c) $<
# SSE4.2 or NEON available
neon.o : neon.cpp
$(CXX) $(strip $(CXXFLAGS) $(NEON_FLAG) -c) $<
# SSE4.2 or ARMv8a available
blake2-simd.o : blake2-simd.cpp
$(CXX) $(strip $(CXXFLAGS) $(BLAKE2_FLAG) -c) $<
# SSE4.2 or ARMv8a available
crc-simd.o : crc-simd.cpp
$(CXX) $(strip $(CXXFLAGS) $(CRC_FLAG) -c) $<
# SSE4.2/SHANI or ARMv8a available
# SSE4.2/SHA-NI or ARMv8a available
sha-simd.o : sha-simd.cpp
$(CXX) $(strip $(CXXFLAGS) $(SHA_FLAG) -c) $<

75
aria-simd.cpp Normal file
View File

@ -0,0 +1,75 @@
// crc-simd.cpp - written and placed in the public domain by
// Jeffrey Walton, Uri Blumenthal and Marcel Raad.
//
// This source file uses intrinsics to gain access to ARMv7a and
// ARMv8a NEON instructions. A separate source file is needed
// because additional CXXFLAGS are required to enable the
// appropriate instructions sets in some build configurations.
#include "pch.h"
#include "config.h"
#include "misc.h"
#if (CRYPTOPP_ARM_NEON_AVAILABLE) && defined(__GNUC__)
# include "arm_neon.h"
#endif
NAMESPACE_BEGIN(CryptoPP)
#if (CRYPTOPP_ARM_NEON_AVAILABLE)
template <unsigned int N>
inline void ARIA_GSRK_NEON(const uint32x4_t X, const uint32x4_t Y, byte RK[16])
{
static const unsigned int Q1 = (4-(N/32)) % 4;
static const unsigned int Q2 = (3-(N/32)) % 4;
static const unsigned int R = N % 32;
vst1q_u32(reinterpret_cast<uint32_t*>(RK),
veorq_u32(X, veorq_u32(
vshrq_n_u32(vextq_u32(Y, Y, Q1), R),
vshlq_n_u32(vextq_u32(Y, Y, Q2), 32-R))));
}
void ARIA_UncheckedSetKey_Schedule_NEON(byte* rk, word32* ws, unsigned int keylen)
{
const uint32x4_t w0 = vld1q_u32((const uint32_t*)(ws+ 0));
const uint32x4_t w1 = vld1q_u32((const uint32_t*)(ws+ 8));
const uint32x4_t w2 = vld1q_u32((const uint32_t*)(ws+12));
const uint32x4_t w3 = vld1q_u32((const uint32_t*)(ws+16));
ARIA_GSRK_NEON<19>(w0, w1, rk + 0);
ARIA_GSRK_NEON<19>(w1, w2, rk + 16);
ARIA_GSRK_NEON<19>(w2, w3, rk + 32);
ARIA_GSRK_NEON<19>(w3, w0, rk + 48);
ARIA_GSRK_NEON<31>(w0, w1, rk + 64);
ARIA_GSRK_NEON<31>(w1, w2, rk + 80);
ARIA_GSRK_NEON<31>(w2, w3, rk + 96);
ARIA_GSRK_NEON<31>(w3, w0, rk + 112);
ARIA_GSRK_NEON<67>(w0, w1, rk + 128);
ARIA_GSRK_NEON<67>(w1, w2, rk + 144);
ARIA_GSRK_NEON<67>(w2, w3, rk + 160);
ARIA_GSRK_NEON<67>(w3, w0, rk + 176);
ARIA_GSRK_NEON<97>(w0, w1, rk + 192);
if (keylen > 16)
{
ARIA_GSRK_NEON<97>(w1, w2, rk + 208);
ARIA_GSRK_NEON<97>(w2, w3, rk + 224);
if (keylen > 24)
{
ARIA_GSRK_NEON< 97>(w3, w0, rk + 240);
ARIA_GSRK_NEON<109>(w0, w1, rk + 256);
}
}
}
void ARIA_ProcessAndXorBlock_Xor_NEON(const byte* xorBlock, byte* outBlock)
{
vst1q_u32(reinterpret_cast<uint32_t*>(outBlock), veorq_u32(
vld1q_u32(reinterpret_cast<const uint32_t*>(outBlock)),
vld1q_u32(reinterpret_cast<const uint32_t*>(xorBlock))));
}
#endif
NAMESPACE_END

View File

@ -15,7 +15,7 @@
# define CRYPTOPP_ENABLE_ARIA_SSSE3_INTRINSICS 1
#endif
#if CRYPTOPP_BOOL_NEON_INTRINSICS_AVAILABLE
#if CRYPTOPP_ARM_NEON_AVAILABLE
# define CRYPTOPP_ENABLE_ARIA_NEON_INTRINSICS 1
#endif
@ -222,6 +222,11 @@ inline byte ARIA_BRF(const word32 x, const int y) {
#define ARIA_FO {SBL1_M(t[0],t[1],t[2],t[3]) ARIA_MM(t[0],t[1],t[2],t[3]) ARIA_P(t[0],t[1],t[2],t[3]) ARIA_MM(t[0],t[1],t[2],t[3])}
#define ARIA_FE {SBL2_M(t[0],t[1],t[2],t[3]) ARIA_MM(t[0],t[1],t[2],t[3]) ARIA_P(t[2],t[3],t[0],t[1]) ARIA_MM(t[0],t[1],t[2],t[3])}
#if (CRYPTOPP_ARM_NEON_AVAILABLE)
extern void ARIA_UncheckedSetKey_Schedule_NEON(byte* rk, word32* ws, unsigned int keylen);
extern void ARIA_ProcessAndXorBlock_Xor_NEON(const byte* xorBlock, byte* outblock);
#endif
// n-bit right shift of Y XORed to X
template <unsigned int N>
inline void ARIA_GSRK(const word32 X[4], const word32 Y[4], byte RK[16])
@ -235,21 +240,6 @@ inline void ARIA_GSRK(const word32 X[4], const word32 Y[4], byte RK[16])
reinterpret_cast<word32*>(RK)[3] = (X[3]) ^ ((Y[(Q+3)%4])>>R) ^ ((Y[(Q+2)%4])<<(32-R));
}
#if CRYPTOPP_ENABLE_ARIA_NEON_INTRINSICS
template <unsigned int N>
inline void ARIA_GSRK_NEON(const uint32x4_t X, const uint32x4_t Y, byte RK[16])
{
static const unsigned int Q1 = (4-(N/32)) % 4;
static const unsigned int Q2 = (3-(N/32)) % 4;
static const unsigned int R = N % 32;
vst1q_u32(reinterpret_cast<uint32_t*>(RK),
veorq_u32(X, veorq_u32(
vshrq_n_u32(vextq_u32(Y, Y, Q1), R),
vshlq_n_u32(vextq_u32(Y, Y, Q2), 32-R))));
}
#endif
void ARIA::Base::UncheckedSetKey(const byte *key, unsigned int keylen, const NameValuePairs &params)
{
CRYPTOPP_UNUSED(params);
@ -385,36 +375,7 @@ void ARIA::Base::UncheckedSetKey(const byte *key, unsigned int keylen, const Nam
#if CRYPTOPP_ENABLE_ARIA_NEON_INTRINSICS
if (HasNEON())
{
const uint32x4_t w0 = vld1q_u32((const uint32_t*)(m_w.data()+0));
const uint32x4_t w1 = vld1q_u32((const uint32_t*)(m_w.data()+8));
const uint32x4_t w2 = vld1q_u32((const uint32_t*)(m_w.data()+12));
const uint32x4_t w3 = vld1q_u32((const uint32_t*)(m_w.data()+16));
ARIA_GSRK_NEON<19>(w0, w1, rk + 0);
ARIA_GSRK_NEON<19>(w1, w2, rk + 16);
ARIA_GSRK_NEON<19>(w2, w3, rk + 32);
ARIA_GSRK_NEON<19>(w3, w0, rk + 48);
ARIA_GSRK_NEON<31>(w0, w1, rk + 64);
ARIA_GSRK_NEON<31>(w1, w2, rk + 80);
ARIA_GSRK_NEON<31>(w2, w3, rk + 96);
ARIA_GSRK_NEON<31>(w3, w0, rk + 112);
ARIA_GSRK_NEON<67>(w0, w1, rk + 128);
ARIA_GSRK_NEON<67>(w1, w2, rk + 144);
ARIA_GSRK_NEON<67>(w2, w3, rk + 160);
ARIA_GSRK_NEON<67>(w3, w0, rk + 176);
ARIA_GSRK_NEON<97>(w0, w1, rk + 192);
if (keylen > 16)
{
ARIA_GSRK_NEON<97>(w1, w2, rk + 208);
ARIA_GSRK_NEON<97>(w2, w3, rk + 224);
if (keylen > 24)
{
ARIA_GSRK_NEON< 97>(w3, w0, rk + 240);
ARIA_GSRK_NEON<109>(w0, w1, rk + 256);
}
}
ARIA_UncheckedSetKey_Schedule_NEON(rk, m_w, keylen);
}
else
#endif // CRYPTOPP_ENABLE_ARIA_NEON_INTRINSICS
@ -621,18 +582,13 @@ void ARIA::Base::ProcessAndXorBlock(const byte *inBlock, const byte *xorBlock, b
if (HasNEON())
{
if (xorBlock != NULLPTR)
{
vst1q_u32(reinterpret_cast<uint32_t*>(outBlock),
veorq_u32(
vld1q_u32((const uint32_t*)outBlock),
vld1q_u32((const uint32_t*)xorBlock)));
}
ARIA_ProcessAndXorBlock_Xor_NEON(xorBlock, outBlock);
}
else
#endif // CRYPTOPP_ENABLE_ARIA_NEON_INTRINSICS
{
if (xorBlock != NULLPTR)
for (unsigned int n=0; n<16; ++n)
for (unsigned int n=0; n<ARIA::BLOCKSIZE; ++n)
outBlock[n] ^= xorBlock[n];
}
}

2162
blake2-simd.cpp Normal file

File diff suppressed because it is too large Load Diff

2275
blake2.cpp

File diff suppressed because it is too large Load Diff

View File

@ -516,9 +516,9 @@ NAMESPACE_END
#if (CRYPTOPP_BOOL_ARM32 || CRYPTOPP_BOOL_ARM64)
// Requires ARMv7 and ACLE 1.0. Testing shows ARMv7 is really ARMv7a under most toolchains.
#if !defined(CRYPTOPP_BOOL_NEON_INTRINSICS_AVAILABLE) && !defined(CRYPTOPP_DISABLE_ASM)
# if defined(__ARM_NEON__) || defined(__ARM_NEON) || defined(_M_ARM)
# define CRYPTOPP_BOOL_NEON_INTRINSICS_AVAILABLE 1
#if !defined(CRYPTOPP_ARM_NEON_AVAILABLE) && !defined(CRYPTOPP_DISABLE_ASM)
# if defined(__ARM_NEON__) || defined(__ARM_NEON) || (CRYPTOPP_GCC_VERSION >= 40800) || (CRYPTOPP_LLVM_CLANG_VERSION >= 30500)
# define CRYPTOPP_ARM_NEON_AVAILABLE 1
# endif
#endif
@ -563,7 +563,7 @@ NAMESPACE_END
// ***************** Miscellaneous ********************
#if CRYPTOPP_BOOL_SSE2_INTRINSICS_AVAILABLE || CRYPTOPP_BOOL_SSE2_ASM_AVAILABLE || CRYPTOPP_BOOL_NEON_INTRINSICS_AVAILABLE || defined(CRYPTOPP_X64_MASM_AVAILABLE)
#if CRYPTOPP_BOOL_SSE2_INTRINSICS_AVAILABLE || CRYPTOPP_BOOL_SSE2_ASM_AVAILABLE || CRYPTOPP_ARM_NEON_AVAILABLE || defined(CRYPTOPP_X64_MASM_AVAILABLE)
#define CRYPTOPP_BOOL_ALIGN16 1
#else
#define CRYPTOPP_BOOL_ALIGN16 0

100
cpu.cpp
View File

@ -334,15 +334,16 @@ bool CRYPTOPP_SECTION_INIT g_hasNEON = false, CRYPTOPP_SECTION_INIT g_hasPMULL =
bool CRYPTOPP_SECTION_INIT g_hasAES = false, CRYPTOPP_SECTION_INIT g_hasSHA1 = false, CRYPTOPP_SECTION_INIT g_hasSHA2 = false;
word32 CRYPTOPP_SECTION_INIT g_cacheLineSize = CRYPTOPP_L1_CACHE_LINE_SIZE;
extern bool CPU_TryNEON_ARM();
extern bool CPU_TryAES_ARMV8();
extern bool CPU_TrySHA1_ARMV8();
extern bool CPU_TrySHA2_ARMV8();
extern bool CPU_TryCRC32_ARMV8();
extern bool CPU_TryPMULL_ARMV8();
#ifndef CRYPTOPP_MS_STYLE_INLINE_ASSEMBLY
extern "C"
{
static jmp_buf s_jmpNoNEON;
static void SigIllHandlerNEON(int)
{
longjmp(s_jmpNoNEON, 1);
}
static jmp_buf s_jmpNoPMULL;
static void SigIllHandlerPMULL(int)
{
@ -359,70 +360,20 @@ extern "C"
static bool TryNEON()
{
#if (CRYPTOPP_BOOL_NEON_INTRINSICS_AVAILABLE)
# if defined(CRYPTOPP_MS_STYLE_INLINE_ASSEMBLY)
volatile bool result = true;
__try
{
uint32_t v1[4] = {1,1,1,1};
uint32x4_t x1 = vld1q_u32(v1);
uint64_t v2[2] = {1,1};
uint64x2_t x2 = vld1q_u64(v2);
uint32x4_t x3 = vdupq_n_u32(2);
x3 = vsetq_lane_u32(vgetq_lane_u32(x1,0),x3,0);
x3 = vsetq_lane_u32(vgetq_lane_u32(x1,3),x3,3);
uint64x2_t x4 = vdupq_n_u64(2);
x4 = vsetq_lane_u64(vgetq_lane_u64(x2,0),x4,0);
x4 = vsetq_lane_u64(vgetq_lane_u64(x2,1),x4,1);
result = !!(vgetq_lane_u32(x3,0) | vgetq_lane_u64(x4,1));
}
__except (EXCEPTION_EXECUTE_HANDLER)
{
return false;
}
return result;
# else
// longjmp and clobber warnings. Volatile is required.
// http://github.com/weidai11/cryptopp/issues/24 and http://stackoverflow.com/q/7721854
volatile bool result = true;
volatile SigHandler oldHandler = signal(SIGILL, SigIllHandlerNEON);
if (oldHandler == SIG_ERR)
return false;
volatile sigset_t oldMask;
if (sigprocmask(0, NULLPTR, (sigset_t*)&oldMask))
return false;
if (setjmp(s_jmpNoNEON))
result = false;
else
{
uint32_t v1[4] = {1,1,1,1};
uint32x4_t x1 = vld1q_u32(v1);
uint64_t v2[2] = {1,1};
uint64x2_t x2 = vld1q_u64(v2);
uint32x4_t x3 = {0,0,0,0};
x3 = vsetq_lane_u32(vgetq_lane_u32(x1,0),x3,0);
x3 = vsetq_lane_u32(vgetq_lane_u32(x1,3),x3,3);
uint64x2_t x4 = {0,0};
x4 = vsetq_lane_u64(vgetq_lane_u64(x2,0),x4,0);
x4 = vsetq_lane_u64(vgetq_lane_u64(x2,1),x4,1);
// Hack... GCC optimizes away the code and returns true
result = !!(vgetq_lane_u32(x3,0) | vgetq_lane_u64(x4,1));
}
sigprocmask(SIG_SETMASK, (sigset_t*)&oldMask, NULLPTR);
signal(SIGILL, oldHandler);
return result;
# endif
#if (CRYPTOPP_ARMV8A_CRC32_AVAILABLE)
return CPU_TryCRC32_ARMV8();
#else
return false;
#endif // CRYPTOPP_BOOL_NEON_INTRINSICS_AVAILABLE
#endif
}
static bool TryCRC32()
{
#if (CRYPTOPP_ARMV8A_CRC32_AVAILABLE)
return CPU_TryCRC32_ARMV8();
#else
return false;
#endif
}
static bool TryPMULL()
@ -492,19 +443,6 @@ static bool TryPMULL()
#endif // CRYPTOPP_ARMV8A_CRYPTO_AVAILABLE
}
extern bool CPU_TryCRC32_ARMV8();
extern bool CPU_TrySHA1_ARMV8();
extern bool CPU_TrySHA2_ARMV8();
static bool TryCRC32()
{
#if (CRYPTOPP_ARMV8A_CRC32_AVAILABLE)
return CPU_TryCRC32_ARMV8();
#else
return false;
#endif
}
static bool TryAES()
{
#if (CRYPTOPP_ARMV8A_CRYPTO_AVAILABLE)

View File

@ -48,9 +48,9 @@
# If you use 'make sources' from Linux makefile, then add 'winpipes.cpp'. Platform specific
# classes, like 'rdrand.cpp', should not be included. Add them under the X86 and X64 rules.
LIB_SRCS = cryptlib.cpp cpu.cpp integer.cpp shacal2.cpp md5.cpp shark.cpp zinflate.cpp gf2n.cpp salsa.cpp xtr.cpp oaep.cpp rc2.cpp default.cpp wait.cpp wake.cpp twofish.cpp iterhash.cpp adler32.cpp algparam.cpp marss.cpp blowfish.cpp ecp.cpp strciphr.cpp aria.cpp camellia.cpp dh2.cpp ida.cpp zlib.cpp elgamal.cpp crc.cpp crc-simd.cpp dessp.cpp tea.cpp eax.cpp network.cpp sha.cpp emsa2.cpp pkcspad.cpp squaretb.cpp idea.cpp authenc.cpp hmac.cpp xtrcrypt.cpp queue.cpp mars.cpp rc5.cpp md2.cpp hrtimer.cpp vmac.cpp eprecomp.cpp hex.cpp dsa.cpp fips140.cpp gzip.cpp seal.cpp blake2.cpp files.cpp base32.cpp sharkbox.cpp safer.cpp randpool.cpp sosemanuk.cpp arc4.cpp osrng.cpp skipjack.cpp seed.cpp sha3.cpp filters.cpp bfinit.cpp rabin.cpp 3way.cpp rw.cpp rdtables.cpp rsa.cpp tftables.cpp gost.cpp socketft.cpp nbtheory.cpp panama.cpp modes.cpp rijndael.cpp casts.cpp algebra.cpp esign.cpp gfpcrypt.cpp dll.cpp ec2n.cpp poly1305.cpp polynomi.cpp blumshub.cpp des.cpp basecode.cpp zdeflate.cpp base64.cpp rc6.cpp gf256.cpp mqueue.cpp misc.cpp pssr.cpp channels.cpp rng.cpp threefish.cpp tiger.cpp cast.cpp square.cpp asn.cpp chacha.cpp whrlpool.cpp md4.cpp dh.cpp ccm.cpp mqv.cpp tigertab.cpp gf2_32.cpp cbcmac.cpp ttmac.cpp luc.cpp trdlocal.cpp pubkey.cpp gcm.cpp ripemd.cpp kalyna.cpp kalynatab.cpp keccak.cpp eccrypto.cpp serpent.cpp cmac.cpp winpipes.cpp
LIB_SRCS = cryptlib.cpp cpu.cpp integer.cpp shacal2.cpp md5.cpp shark.cpp zinflate.cpp gf2n.cpp salsa.cpp xtr.cpp oaep.cpp rc2.cpp default.cpp wait.cpp wake.cpp twofish.cpp iterhash.cpp adler32.cpp algparam.cpp marss.cpp blowfish.cpp ecp.cpp strciphr.cpp aria.cpp aria-simd.cpp camellia.cpp dh2.cpp ida.cpp zlib.cpp elgamal.cpp crc.cpp crc-simd.cpp dessp.cpp tea.cpp eax.cpp network.cpp sha.cpp sha-simd.cpp emsa2.cpp pkcspad.cpp squaretb.cpp idea.cpp authenc.cpp hmac.cpp xtrcrypt.cpp queue.cpp mars.cpp rc5.cpp md2.cpp hrtimer.cpp vmac.cpp eprecomp.cpp hex.cpp dsa.cpp fips140.cpp gzip.cpp seal.cpp blake2.cpp blake2-simd.cpp files.cpp base32.cpp sharkbox.cpp safer.cpp randpool.cpp sosemanuk.cpp arc4.cpp osrng.cpp skipjack.cpp seed.cpp sha3.cpp filters.cpp bfinit.cpp rabin.cpp 3way.cpp rw.cpp rdtables.cpp rsa.cpp tftables.cpp gost.cpp socketft.cpp nbtheory.cpp panama.cpp modes.cpp rijndael.cpp casts.cpp algebra.cpp esign.cpp gfpcrypt.cpp dll.cpp ec2n.cpp poly1305.cpp polynomi.cpp blumshub.cpp des.cpp basecode.cpp zdeflate.cpp base64.cpp rc6.cpp gf256.cpp mqueue.cpp misc.cpp pssr.cpp channels.cpp rng.cpp threefish.cpp tiger.cpp cast.cpp square.cpp asn.cpp chacha.cpp whrlpool.cpp md4.cpp dh.cpp ccm.cpp mqv.cpp tigertab.cpp gf2_32.cpp cbcmac.cpp ttmac.cpp luc.cpp trdlocal.cpp pubkey.cpp gcm.cpp ripemd.cpp kalyna.cpp kalynatab.cpp keccak.cpp eccrypto.cpp serpent.cpp cmac.cpp winpipes.cpp
LIB_OBJS = cryptlib.obj cpu.obj integer.obj shacal2.obj md5.obj shark.obj zinflate.obj gf2n.obj salsa.obj xtr.obj oaep.obj rc2.obj default.obj wait.obj wake.obj twofish.obj iterhash.obj adler32.obj algparam.obj marss.obj blowfish.obj ecp.obj strciphr.obj aria.obj camellia.obj dh2.obj ida.obj zlib.obj elgamal.obj crc.obj crc-simd.obj dessp.obj tea.obj eax.obj network.obj sha.obj emsa2.obj pkcspad.obj squaretb.obj idea.obj authenc.obj hmac.obj xtrcrypt.obj queue.obj mars.obj rc5.obj md2.obj hrtimer.obj vmac.obj eprecomp.obj hex.obj dsa.obj fips140.obj gzip.obj seal.obj blake2.obj files.obj base32.obj sharkbox.obj safer.obj randpool.obj sosemanuk.obj arc4.obj osrng.obj skipjack.obj seed.obj sha3.obj filters.obj bfinit.obj rabin.obj 3way.obj rw.obj rdtables.obj rsa.obj tftables.obj gost.obj socketft.obj nbtheory.obj panama.obj modes.obj rijndael.obj casts.obj algebra.obj esign.obj gfpcrypt.obj dll.obj ec2n.obj poly1305.obj polynomi.obj blumshub.obj des.obj basecode.obj zdeflate.obj base64.obj rc6.obj gf256.obj mqueue.obj misc.obj pssr.obj channels.obj rng.obj threefish.obj tiger.obj cast.obj square.obj asn.obj chacha.obj whrlpool.obj md4.obj dh.obj ccm.obj mqv.obj tigertab.obj gf2_32.obj cbcmac.obj ttmac.obj luc.obj trdlocal.obj pubkey.obj gcm.obj ripemd.obj kalyna.obj kalynatab.obj keccak.obj eccrypto.obj serpent.obj cmac.obj winpipes.obj
LIB_OBJS = cryptlib.obj cpu.obj integer.obj shacal2.obj md5.obj shark.obj zinflate.obj gf2n.obj salsa.obj xtr.obj oaep.obj rc2.obj default.obj wait.obj wake.obj twofish.obj iterhash.obj adler32.obj algparam.obj marss.obj blowfish.obj ecp.obj strciphr.obj aria.obj aria-simd.obj camellia.obj dh2.obj ida.obj zlib.obj elgamal.obj crc.obj crc-simd.obj dessp.obj tea.obj eax.obj network.obj sha.obj sha-simd.obj emsa2.obj pkcspad.obj squaretb.obj idea.obj authenc.obj hmac.obj xtrcrypt.obj queue.obj mars.obj rc5.obj md2.obj hrtimer.obj vmac.obj eprecomp.obj hex.obj dsa.obj fips140.obj gzip.obj seal.obj blake2.obj blake2-simd.obj files.obj base32.obj sharkbox.obj safer.obj randpool.obj sosemanuk.obj arc4.obj osrng.obj skipjack.obj seed.obj sha3.obj filters.obj bfinit.obj rabin.obj 3way.obj rw.obj rdtables.obj rsa.obj tftables.obj gost.obj socketft.obj nbtheory.obj panama.obj modes.obj rijndael.obj casts.obj algebra.obj esign.obj gfpcrypt.obj dll.obj ec2n.obj poly1305.obj polynomi.obj blumshub.obj des.obj basecode.obj zdeflate.obj base64.obj rc6.obj gf256.obj mqueue.obj misc.obj pssr.obj channels.obj rng.obj threefish.obj tiger.obj cast.obj square.obj asn.obj chacha.obj whrlpool.obj md4.obj dh.obj ccm.obj mqv.obj tigertab.obj gf2_32.obj cbcmac.obj ttmac.obj luc.obj trdlocal.obj pubkey.obj gcm.obj ripemd.obj kalyna.obj kalynatab.obj keccak.obj eccrypto.obj serpent.obj cmac.obj winpipes.obj
TEST_SRCS = bench1.cpp bench2.cpp test.cpp validat0.cpp validat1.cpp validat2.cpp validat3.cpp datatest.cpp regtest1.cpp regtest2.cpp regtest3.cpp fipsalgt.cpp dlltest.cpp fipstest.cpp

View File

@ -250,15 +250,15 @@ if [[ ("$SUNCC_510_OR_ABOVE" -ne "0") ]]; then
HAVE_OFAST=0
fi
if [[ (-z "$TMP") ]]; then
if [[ (-z "$TMPDIR") ]]; then
if [[ (-d "/tmp") ]]; then
TMP=/tmp
TMPDIR=/tmp
elif [[ (-d "/temp") ]]; then
TMP=/temp
TMPDIR=/temp
elif [[ (-d "$HOME/tmp") ]]; then
TMP="$HOME/tmp"
TMPDIR="$HOME/tmp"
else
echo "Please set TMP to a valid directory"
echo "Please set TMPDIR to a valid directory"
[[ "$0" = "$BASH_SOURCE" ]] && exit 1 || return 1
fi
fi
@ -267,74 +267,74 @@ fi
rm -f adhoc.cpp > /dev/null 2>&1
cp adhoc.cpp.proto adhoc.cpp
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_CXX17") ]]; then
HAVE_CXX17=0
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=c++17 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=c++17 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_CXX17=1
fi
fi
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_GNU17") ]]; then
HAVE_GNU17=0
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=gnu++17 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=gnu++17 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_GNU17=1
fi
fi
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_CXX14") ]]; then
HAVE_CXX14=0
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=c++14 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=c++14 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_CXX14=1
fi
fi
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_GNU14") ]]; then
HAVE_GNU14=0
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=gnu++14 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=gnu++14 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_GNU14=1
fi
fi
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_CXX11") ]]; then
HAVE_CXX11=0
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=c++11 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=c++11 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_CXX11=1
fi
fi
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_GNU11") ]]; then
HAVE_GNU11=0
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=gnu++11 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=gnu++11 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_GNU11=1
fi
fi
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_CXX03") ]]; then
HAVE_CXX03=0
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=c++03 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=c++03 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_CXX03=1
fi
fi
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_GNU03") ]]; then
HAVE_GNU03=0
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=gnu++03 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -std=gnu++03 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_GNU03=1
fi
@ -342,13 +342,13 @@ fi
# Use a fallback strategy so OPT_O0 can be used with DEBUG_CXXFLAGS
OPT_O0=
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -O0 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -O0 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
OPT_O0=-O0
else
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -xO0 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -xO0 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
OPT_O0=-xO0
fi
@ -356,13 +356,13 @@ fi
# Use a fallback strategy so OPT_O1 can be used with VALGRIND_CXXFLAGS
OPT_O1=
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -O1 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -O1 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
OPT_O1=-O1
else
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -xO1 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -xO1 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
OPT_O1=-xO1
fi
@ -370,13 +370,13 @@ fi
# Use a fallback strategy so OPT_O2 can be used with RELEASE_CXXFLAGS
OPT_O2=
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -O2 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -O2 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
OPT_O2=-O2
else
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -xO2 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -xO2 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
OPT_O2=-xO2
fi
@ -385,14 +385,14 @@ fi
if [[ (-z "$HAVE_O3") ]]; then
HAVE_O3=0
OPT_O3=
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -O3 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -O3 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
HAVE_O3=1
OPT_O3=-O3
else
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -xO3 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -xO3 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
HAVE_O3=1
OPT_O3=-xO3
@ -404,14 +404,14 @@ fi
if [[ ( (-z "$HAVE_O5") && ("$CLANG_COMPILER" -eq "0") ) ]]; then
HAVE_O5=0
OPT_O5=
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -O5 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -O5 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
HAVE_O5=1
OPT_O5=-O5
else
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -xO5 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -xO5 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
HAVE_O5=1
OPT_O5=-xO5
@ -423,8 +423,8 @@ fi
if [[ (-z "$HAVE_OS") ]]; then
HAVE_OS=0
OPT_OS=
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -Os adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -Os adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
HAVE_OS=1
OPT_OS=-Os
@ -435,8 +435,8 @@ fi
if [[ (-z "$HAVE_OFAST") ]]; then
HAVE_OFAST=0
OPT_OFAST=
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -Ofast adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -Ofast adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
HAVE_OFAST=1
OPT_OFAST=-Ofast
@ -445,13 +445,13 @@ fi
# Use a fallback strategy so OPT_G2 can be used with RELEASE_CXXFLAGS
OPT_G2=
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -g2 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -g2 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
OPT_G2=-g2
else
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -g adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -g adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
OPT_G2=-g
fi
@ -459,13 +459,13 @@ fi
# Use a fallback strategy so OPT_G3 can be used with DEBUG_CXXFLAGS
OPT_G3=
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -g3 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -g3 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
OPT_G3=-g3
else
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -g adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -g adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
OPT_G3=-g
fi
@ -473,10 +473,10 @@ fi
# Cygwin and noisy compiles
OPT_PIC=
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_PIC") ]]; then
HAVE_PIC=0
PIC_PROBLEMS=$("$CXX" -DCRYPTOPP_ADHOC_MAIN -fPIC adhoc.cpp -o "$TMP/adhoc.exe" 2>&1 | "$EGREP" -ic '(warning|error)')
PIC_PROBLEMS=$("$CXX" -DCRYPTOPP_ADHOC_MAIN -fPIC adhoc.cpp -o "$TMPDIR/adhoc.exe" 2>&1 | "$EGREP" -ic '(warning|error)')
if [[ "$PIC_PROBLEMS" -eq "0" ]]; then
HAVE_PIC=1
OPT_PIC=-fPIC
@ -484,12 +484,12 @@ if [[ (-z "$HAVE_PIC") ]]; then
fi
# GCC 4.8; Clang 3.4
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_UBSAN") ]]; then
HAVE_UBSAN=0
"$CXX" -DCRYPTOPP_ADHOC_MAIN -fsanitize=undefined adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -fsanitize=undefined adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
"$TMP/adhoc.exe" > /dev/null 2>&1
"$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
HAVE_UBSAN=1
fi
@ -497,12 +497,12 @@ if [[ (-z "$HAVE_UBSAN") ]]; then
fi
# GCC 4.8; Clang 3.4
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_ASAN") ]]; then
HAVE_ASAN=0
"$CXX" -DCRYPTOPP_ADHOC_MAIN -fsanitize=address adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -fsanitize=address adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
"$TMP/adhoc.exe" > /dev/null 2>&1
"$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
HAVE_ASAN=1
fi
@ -510,41 +510,41 @@ if [[ (-z "$HAVE_ASAN") ]]; then
fi
# GCC 6.0; maybe Clang
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_BSAN") ]]; then
HAVE_BSAN=0
"$CXX" -DCRYPTOPP_ADHOC_MAIN -fsanitize=bounds-strict adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -fsanitize=bounds-strict adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
"$TMP/adhoc.exe" > /dev/null 2>&1
"$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
HAVE_BSAN=1
fi
fi
fi
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_OMP") ]]; then
HAVE_OMP=0
if [[ "$GCC_COMPILER" -ne "0" ]]; then
"$CXX" -DCRYPTOPP_ADHOC_MAIN -fopenmp -O3 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -fopenmp -O3 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_OMP=1
OMP_FLAGS=(-fopenmp -O3)
fi
elif [[ "$INTEL_COMPILER" -ne "0" ]]; then
"$CXX" -DCRYPTOPP_ADHOC_MAIN -openmp -O3 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -openmp -O3 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_OMP=1
OMP_FLAGS=(-openmp -O3)
fi
elif [[ "$CLANG_COMPILER" -ne "0" ]]; then
"$CXX" -DCRYPTOPP_ADHOC_MAIN -fopenmp=libomp -O3 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -fopenmp=libomp -O3 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_OMP=1
OMP_FLAGS=(-fopenmp=libomp -O3)
fi
elif [[ "$SUN_COMPILER" -ne "0" ]]; then
"$CXX" -DCRYPTOPP_ADHOC_MAIN -xopenmp=parallel -xO3 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -xopenmp=parallel -xO3 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_OMP=1
OMP_FLAGS=(-xopenmp=parallel -xO3)
@ -552,33 +552,33 @@ if [[ (-z "$HAVE_OMP") ]]; then
fi
fi
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_INTEL_MULTIARCH") ]]; then
HAVE_INTEL_MULTIARCH=0
if [[ ("$IS_DARWIN" -ne "0") && ("$IS_X86" -ne "0" || "$IS_X64" -ne "0") ]]; then
"$CXX" -DCRYPTOPP_ADHOC_MAIN -arch i386 -arch x86_64 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -arch i386 -arch x86_64 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_INTEL_MULTIARCH=1
fi
fi
fi
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_PPC_MULTIARCH") ]]; then
HAVE_PPC_MULTIARCH=0
if [[ ("$IS_DARWIN" -ne "0") && ("$IS_PPC" -ne "0") ]]; then
"$CXX" -DCRYPTOPP_ADHOC_MAIN -arch ppc -arch ppc64 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -arch ppc -arch ppc64 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_PPC_MULTIARCH=1
fi
fi
fi
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ (-z "$HAVE_X32") ]]; then
HAVE_X32=0
if [[ "$IS_X32" -ne "0" ]]; then
"$CXX" -DCRYPTOPP_ADHOC_MAIN -mx32 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -mx32 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_X32=1
fi
@ -588,8 +588,8 @@ fi
# Hit or miss, mostly hit
if [[ (-z "$HAVE_NATIVE_ARCH") ]]; then
HAVE_NATIVE_ARCH=0
rm -f "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -march=native adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
rm -f "$TMPDIR/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -march=native adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ ("$?" -eq "0") ]]; then
HAVE_NATIVE_ARCH=1
fi
@ -603,7 +603,7 @@ if [[ (-z "$HAVE_LDGOLD") ]]; then
if [[ (! -z "$LD_GOLD") && (! -z "$ELF_FILE") ]]; then
LD_GOLD=$(file "$LD_GOLD" | cut -d":" -f 2 | "$EGREP" -i -c "elf")
if [[ ("$LD_GOLD" -ne "0") ]]; then
"$CXX" -DCRYPTOPP_ADHOC_MAIN -fuse-ld=gold adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -fuse-ld=gold adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_LDGOLD=1
fi
@ -688,10 +688,10 @@ fi
# Used to disassemble object modules so we can verify some aspects of code generation
if [[ (-z "$HAVE_DISASS") ]]; then
echo "int main(int argc, char* argv[]) {return 0;}" > "$TMP/test.cc"
"$CXX" "$TMP/test.cc" -o "$TMP/test.exe" > /dev/null 2>&1
echo "int main(int argc, char* argv[]) {return 0;}" > "$TMPDIR/test.cc"
"$CXX" "$TMPDIR/test.cc" -o "$TMPDIR/test.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
"$DISASS" "${DISASSARGS[@]}" "$TMP/test.exe" > /dev/null 2>&1
"$DISASS" "${DISASSARGS[@]}" "$TMPDIR/test.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
HAVE_DISASS=1
else
@ -1167,7 +1167,7 @@ if [[ ("$HAVE_DISASS" -ne "0" && ("$IS_X86" -ne "0" || "$IS_X64" -ne "0")) ]]; t
echo
OBJFILE=sha.o; rm -f "$OBJFILE" 2>/dev/null
CXX="$CXX" CXXFLAGS="$RELEASE_CXXFLAGS -DDISABLE_NATIVE_ARCH=1 -msse -msse2" "$MAKE" "${MAKEARGS[@]}" $OBJFILE 2>&1 | tee -a "$TEST_RESULTS"
CXX="$CXX" CXXFLAGS="$RELEASE_CXXFLAGS -DDISABLE_NATIVE_ARCH=1" "$MAKE" "${MAKEARGS[@]}" $OBJFILE 2>&1 | tee -a "$TEST_RESULTS"
DISASS_TEXT=$("$DISASS" "${DISASSARGS[@]}" "$OBJFILE" 2>/dev/null)
@ -1201,7 +1201,7 @@ if [[ ("$HAVE_DISASS" -ne "0" && ("$IS_X86" -ne "0" || "$IS_X64" -ne "0")) ]]; t
############################################
# Test CRC-32C code generation
"$CXX" -DCRYPTOPP_ADHOC_MAIN -msse4.2 adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -msse4.2 adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
X86_CRC32=1
fi
@ -1239,7 +1239,7 @@ if [[ ("$HAVE_DISASS" -ne "0" && ("$IS_X86" -ne "0" || "$IS_X64" -ne "0")) ]]; t
############################################
# Test AES-NI code generation
"$CXX" -DCRYPTOPP_ADHOC_MAIN -maes adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -maes adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
X86_AESNI=1
fi
@ -1251,7 +1251,7 @@ if [[ ("$HAVE_DISASS" -ne "0" && ("$IS_X86" -ne "0" || "$IS_X64" -ne "0")) ]]; t
echo
OBJFILE=rijndael.o; rm -f "$OBJFILE" 2>/dev/null
CXX="$CXX" CXXFLAGS="$RELEASE_CXXFLAGS -DDISABLE_NATIVE_ARCH=1 -msse -msse2" "$MAKE" "${MAKEARGS[@]}" $OBJFILE 2>&1 | tee -a "$TEST_RESULTS"
CXX="$CXX" CXXFLAGS="$RELEASE_CXXFLAGS -DDISABLE_NATIVE_ARCH=1" "$MAKE" "${MAKEARGS[@]}" $OBJFILE 2>&1 | tee -a "$TEST_RESULTS"
COUNT=0
FAILED=0
@ -1301,7 +1301,7 @@ if [[ ("$HAVE_DISASS" -ne "0" && ("$IS_X86" -ne "0" || "$IS_X64" -ne "0")) ]]; t
############################################
# X86 carryless multiply code generation
"$CXX" -DCRYPTOPP_ADHOC_MAIN -mpclmul adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -mpclmul adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
X86_PCLMUL=1
fi
@ -1313,7 +1313,7 @@ if [[ ("$HAVE_DISASS" -ne "0" && ("$IS_X86" -ne "0" || "$IS_X64" -ne "0")) ]]; t
echo
OBJFILE=gcm.o; rm -f "$OBJFILE" 2>/dev/null
CXX="$CXX" CXXFLAGS="$RELEASE_CXXFLAGS -DDISABLE_NATIVE_ARCH=1 -msse -msse2" "$MAKE" "${MAKEARGS[@]}" $OBJFILE 2>&1 | tee -a "$TEST_RESULTS"
CXX="$CXX" CXXFLAGS="$RELEASE_CXXFLAGS -DDISABLE_NATIVE_ARCH=1" "$MAKE" "${MAKEARGS[@]}" $OBJFILE 2>&1 | tee -a "$TEST_RESULTS"
COUNT=0
FAILED=0
@ -1339,11 +1339,11 @@ if [[ ("$HAVE_DISASS" -ne "0" && ("$IS_X86" -ne "0" || "$IS_X64" -ne "0")) ]]; t
############################################
# Test RDRAND and RDSEED code generation
"$CXX" -DCRYPTOPP_ADHOC_MAIN -mrdrnd adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -mrdrnd adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
X86_RDRAND=1
fi
"$CXX" -DCRYPTOPP_ADHOC_MAIN -mrdseed adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -mrdseed adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
X86_RDSEED=1
fi
@ -1355,7 +1355,7 @@ if [[ ("$HAVE_DISASS" -ne "0" && ("$IS_X86" -ne "0" || "$IS_X64" -ne "0")) ]]; t
echo
OBJFILE=rdrand.o; rm -f "$OBJFILE" 2>/dev/null
CXX="$CXX" CXXFLAGS="$RELEASE_CXXFLAGS -DDISABLE_NATIVE_ARCH=1 -msse -msse2" "$MAKE" "${MAKEARGS[@]}" $OBJFILE 2>&1 | tee -a "$TEST_RESULTS"
CXX="$CXX" CXXFLAGS="$RELEASE_CXXFLAGS -DDISABLE_NATIVE_ARCH=1" "$MAKE" "${MAKEARGS[@]}" $OBJFILE 2>&1 | tee -a "$TEST_RESULTS"
COUNT=0
FAILED=0
@ -1385,7 +1385,7 @@ if [[ ("$HAVE_DISASS" -ne "0" && ("$IS_X86" -ne "0" || "$IS_X64" -ne "0")) ]]; t
############################################
# X86 SHA code generation
"$CXX" -DCRYPTOPP_ADHOC_MAIN -msha adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -msha adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
X86_SHA=1
fi
@ -1397,7 +1397,7 @@ if [[ ("$HAVE_DISASS" -ne "0" && ("$IS_X86" -ne "0" || "$IS_X64" -ne "0")) ]]; t
echo
OBJFILE=sha-simd.o; rm -f "$OBJFILE" 2>/dev/null
CXX="$CXX" CXXFLAGS="$RELEASE_CXXFLAGS -DDISABLE_NATIVE_ARCH=1 -msse -msse2" "$MAKE" "${MAKEARGS[@]}" $OBJFILE 2>&1 | tee -a "$TEST_RESULTS"
CXX="$CXX" CXXFLAGS="$RELEASE_CXXFLAGS -DDISABLE_NATIVE_ARCH=1" "$MAKE" "${MAKEARGS[@]}" $OBJFILE 2>&1 | tee -a "$TEST_RESULTS"
COUNT=0
FAILED=0
@ -1465,7 +1465,7 @@ if [[ ("$HAVE_DISASS" -ne "0" && ("$IS_ARM32" -ne "0" || "$IS_ARM64" -ne "0")) ]
echo "Testing: ARM NEON code generation" | tee -a "$TEST_RESULTS"
echo
OBJFILE=aria.o; rm -f "$OBJFILE" 2>/dev/null
OBJFILE=aria-simd.o; rm -f "$OBJFILE" 2>/dev/null
CXX="$CXX" CXXFLAGS="$RELEASE_CXXFLAGS -DDISABLE_NATIVE_ARCH=1" "$MAKE" "${MAKEARGS[@]}" $OBJFILE 2>&1 | tee -a "$TEST_RESULTS"
COUNT=0
@ -1515,7 +1515,7 @@ if [[ ("$HAVE_DISASS" -ne "0" && ("$IS_ARM32" -ne "0" || "$IS_ARM64" -ne "0")) ]
############################################
# ARM CRC32 code generation
"$CXX" -DCRYPTOPP_ADHOC_MAIN -march=armv8-a+crc adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -march=armv8-a+crc adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
ARM_CRC32=1
fi
@ -1565,7 +1565,7 @@ if [[ ("$HAVE_DISASS" -ne "0" && ("$IS_ARM32" -ne "0" || "$IS_ARM64" -ne "0")) ]
############################################
# ARM carryless multiply code generation
"$CXX" -DCRYPTOPP_ADHOC_MAIN -march=armv8-a+crypto adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -march=armv8-a+crypto adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
ARM_PMULL=1
fi
@ -1603,7 +1603,7 @@ if [[ ("$HAVE_DISASS" -ne "0" && ("$IS_ARM32" -ne "0" || "$IS_ARM64" -ne "0")) ]
############################################
# ARM SHA code generation
"$CXX" -DCRYPTOPP_ADHOC_MAIN -march=armv8-a+crypto adhoc.cpp -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CXX" -DCRYPTOPP_ADHOC_MAIN -march=armv8-a+crypto adhoc.cpp -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
ARM_SHA=1
fi
@ -5098,7 +5098,7 @@ fi
if [[ ("$CLANG_COMPILER" -eq "0") ]]; then
CLANG_CXX=$(which clang++ 2>&1 | "$GREP" -v "no clang++" | head -1)
"$CLANG_CXX" -x c++ -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$CLANG_CXX" -x c++ -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
############################################
@ -5133,7 +5133,7 @@ fi
if [[ ("$GCC_COMPILER" -eq "0") ]]; then
GCC_CXX=$(which g++ 2>&1 | "$GREP" -v "no g++" | head -1)
"$GCC_CXX" -x c++ -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$GCC_CXX" -x c++ -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
############################################
@ -5171,7 +5171,7 @@ if [[ ("$INTEL_COMPILER" -eq "0") ]]; then
if [[ (-z "$INTEL_CXX") ]]; then
INTEL_CXX=$(find /opt/intel -name icpc 2>/dev/null | "$GREP" -iv composer | head -1)
fi
"$INTEL_CXX" -x c++ -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$INTEL_CXX" -x c++ -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
############################################
@ -5207,7 +5207,7 @@ if [[ ("$IS_DARWIN" -ne "0" && "$MACPORTS_COMPILER" -eq "0") ]]; then
MACPORTS_CXX=$(find /opt/local/bin -name 'g++-mp-4*' 2>/dev/null | head -1)
if [[ (! -z "$MACPORTS_CXX") ]]; then
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
############################################
@ -5240,7 +5240,7 @@ if [[ ("$IS_DARWIN" -ne "0" && "$MACPORTS_COMPILER" -eq "0") ]]; then
MACPORTS_CXX=$(find /opt/local/bin -name 'g++-mp-5*' 2>/dev/null | head -1)
if [[ (! -z "$MACPORTS_CXX") ]]; then
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
############################################
@ -5273,7 +5273,7 @@ if [[ ("$IS_DARWIN" -ne "0" && "$MACPORTS_COMPILER" -eq "0") ]]; then
MACPORTS_CXX=$(find /opt/local/bin -name 'g++-mp-6*' 2>/dev/null | head -1)
if [[ (! -z "$MACPORTS_CXX") ]]; then
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
############################################
@ -5306,7 +5306,7 @@ if [[ ("$IS_DARWIN" -ne "0" && "$MACPORTS_COMPILER" -eq "0") ]]; then
MACPORTS_CXX=$(find /opt/local/bin -name 'g++-mp-7*' 2>/dev/null | head -1)
if [[ (! -z "$MACPORTS_CXX") ]]; then
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
############################################
@ -5339,7 +5339,7 @@ if [[ ("$IS_DARWIN" -ne "0" && "$MACPORTS_COMPILER" -eq "0") ]]; then
MACPORTS_CXX=$(find /opt/local/bin -name 'clang++-mp-3.7*' 2>/dev/null | head -1)
if [[ (! -z "$MACPORTS_CXX") ]]; then
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
############################################
@ -5371,7 +5371,7 @@ if [[ ("$IS_DARWIN" -ne "0" && "$MACPORTS_COMPILER" -eq "0") ]]; then
MACPORTS_CXX=$(find /opt/local/bin -name 'clang++-mp-3.8*' 2>/dev/null | head -1)
if [[ (! -z "$MACPORTS_CXX") ]]; then
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
############################################
@ -5403,7 +5403,7 @@ if [[ ("$IS_DARWIN" -ne "0" && "$MACPORTS_COMPILER" -eq "0") ]]; then
MACPORTS_CXX=$(find /opt/local/bin -name 'clang++-mp-3.9*' 2>/dev/null | head -1)
if [[ (! -z "$MACPORTS_CXX") ]]; then
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
############################################
@ -5435,7 +5435,7 @@ if [[ ("$IS_DARWIN" -ne "0" && "$MACPORTS_COMPILER" -eq "0") ]]; then
MACPORTS_CXX=$(find /opt/local/bin -name 'clang++-mp-4*' 2>/dev/null | head -1)
if [[ (! -z "$MACPORTS_CXX") ]]; then
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMP/adhoc.exe" > /dev/null 2>&1
"$MACPORTS_CXX" -x c++ -std=c++11 -DCRYPTOPP_ADHOC_MAIN adhoc.cpp.proto -o "$TMPDIR/adhoc.exe" > /dev/null 2>&1
if [[ "$?" -eq "0" ]]; then
############################################

View File

@ -158,6 +158,7 @@
<ClCompile Include="algparam.cpp" />
<ClCompile Include="arc4.cpp" />
<ClCompile Include="aria.cpp" />
<ClCompile Include="aria-simd.cpp" />
<ClCompile Include="asn.cpp" />
<ClCompile Include="authenc.cpp" />
<ClCompile Include="base32.cpp" />
@ -165,6 +166,7 @@
<ClCompile Include="basecode.cpp" />
<ClCompile Include="bfinit.cpp" />
<ClCompile Include="blake2.cpp" />
<ClCompile Include="blake2-simd.cpp" />
<ClCompile Include="blowfish.cpp" />
<ClCompile Include="blumshub.cpp" />
<ClCompile Include="camellia.cpp" />

10
gcm.cpp
View File

@ -211,7 +211,7 @@ inline static void SSE2_Xor16(byte *a, const byte *b, const byte *c)
}
#endif
#if CRYPTOPP_BOOL_NEON_INTRINSICS_AVAILABLE
#if CRYPTOPP_ARM_NEON_AVAILABLE
inline static void NEON_Xor16(byte *a, const byte *b, const byte *c)
{
CRYPTOPP_ASSERT(IsAlignedOn(a,GetAlignmentOf<uint64x2_t>()));
@ -437,7 +437,7 @@ void GCM_Base::SetKeyWithoutResync(const byte *userKey, size_t keylength, const
for (k=1; k<j; k++)
SSE2_Xor16(table+i*256*16+(j+k)*16, table+i*256*16+j*16, table+i*256*16+k*16);
else
#elif CRYPTOPP_BOOL_NEON_INTRINSICS_AVAILABLE
#elif CRYPTOPP_ARM_NEON_AVAILABLE
if (HasNEON())
for (j=2; j<=0x80; j*=2)
for (k=1; k<j; k++)
@ -492,7 +492,7 @@ void GCM_Base::SetKeyWithoutResync(const byte *userKey, size_t keylength, const
SSE2_Xor16(table+1024+i*256+(j+k)*16, table+1024+i*256+j*16, table+1024+i*256+k*16);
}
else
#elif CRYPTOPP_BOOL_NEON_INTRINSICS_AVAILABLE
#elif CRYPTOPP_ARM_NEON_AVAILABLE
if (HasNEON())
for (j=2; j<=8; j*=2)
for (k=1; k<j; k++)
@ -582,7 +582,7 @@ unsigned int GCM_Base::OptimalDataAlignment() const
return
#if CRYPTOPP_BOOL_SSE2_ASM_AVAILABLE || defined(CRYPTOPP_X64_MASM_AVAILABLE)
HasSSE2() ? 16 :
#elif CRYPTOPP_BOOL_NEON_INTRINSICS_AVAILABLE
#elif CRYPTOPP_ARM_NEON_AVAILABLE
HasNEON() ? 16 :
#endif
GetBlockCipher().OptimalDataAlignment();
@ -749,7 +749,7 @@ size_t GCM_Base::AuthenticateBlocks(const byte *data, size_t len)
switch (2*(m_buffer.size()>=64*1024)
#if CRYPTOPP_BOOL_SSE2_ASM_AVAILABLE || defined(CRYPTOPP_X64_MASM_AVAILABLE)
+ HasSSE2()
//#elif CRYPTOPP_BOOL_NEON_INTRINSICS_AVAILABLE
//#elif CRYPTOPP_ARM_NEON_AVAILABLE
// + HasNEON()
#endif
)

102
neon.cpp Normal file
View File

@ -0,0 +1,102 @@
// crc-simd.cpp - written and placed in the public domain by
// Jeffrey Walton, Uri Blumenthal and Marcel Raad.
//
// This source file uses intrinsics to gain access to ARMv7a and
// ARMv8a NEON instructions. A separate source file is needed
// because additional CXXFLAGS are required to enable the
// appropriate instructions sets in some build configurations.
#include "pch.h"
#include "config.h"
#include "misc.h"
#if (CRYPTOPP_ARM_NEON_AVAILABLE) && defined(__GNUC__)
# include "arm_neon.h"
#endif
#ifdef CRYPTOPP_GNU_STYLE_INLINE_ASSEMBLY
# include <signal.h>
# include <setjmp.h>
#endif
NAMESPACE_BEGIN(CryptoPP)
#ifdef CRYPTOPP_GNU_STYLE_INLINE_ASSEMBLY
extern "C" {
typedef void (*SigHandler)(int);
static jmp_buf s_jmpSIGILL;
static void SigIllHandler(int)
{
longjmp(s_jmpSIGILL, 1);
}
};
#endif // Not CRYPTOPP_MS_STYLE_INLINE_ASSEMBLY
#if (CRYPTOPP_ARM_NEON_AVAILABLE)
bool CPU_TryNEON_ARM()
{
# if defined(CRYPTOPP_MS_STYLE_INLINE_ASSEMBLY)
volatile bool result = true;
__try
{
uint32_t v1[4] = {1,1,1,1};
uint32x4_t x1 = vld1q_u32(v1);
uint64_t v2[2] = {1,1};
uint64x2_t x2 = vld1q_u64(v2);
uint32x4_t x3 = vdupq_n_u32(2);
x3 = vsetq_lane_u32(vgetq_lane_u32(x1,0),x3,0);
x3 = vsetq_lane_u32(vgetq_lane_u32(x1,3),x3,3);
uint64x2_t x4 = vdupq_n_u64(2);
x4 = vsetq_lane_u64(vgetq_lane_u64(x2,0),x4,0);
x4 = vsetq_lane_u64(vgetq_lane_u64(x2,1),x4,1);
result = !!(vgetq_lane_u32(x3,0) | vgetq_lane_u64(x4,1));
}
__except (EXCEPTION_EXECUTE_HANDLER)
{
return false;
}
return result;
# else
// longjmp and clobber warnings. Volatile is required.
// http://github.com/weidai11/cryptopp/issues/24 and http://stackoverflow.com/q/7721854
volatile bool result = true;
volatile SigHandler oldHandler = signal(SIGILL, SigIllHandler);
if (oldHandler == SIG_ERR)
return false;
volatile sigset_t oldMask;
if (sigprocmask(0, NULLPTR, (sigset_t*)&oldMask))
return false;
if (setjmp(s_jmpSIGILL))
result = false;
else
{
uint32_t v1[4] = {1,1,1,1};
uint32x4_t x1 = vld1q_u32(v1);
uint64_t v2[2] = {1,1};
uint64x2_t x2 = vld1q_u64(v2);
uint32x4_t x3 = {0,0,0,0};
x3 = vsetq_lane_u32(vgetq_lane_u32(x1,0),x3,0);
x3 = vsetq_lane_u32(vgetq_lane_u32(x1,3),x3,3);
uint64x2_t x4 = {0,0};
x4 = vsetq_lane_u64(vgetq_lane_u64(x2,0),x4,0);
x4 = vsetq_lane_u64(vgetq_lane_u64(x2,1),x4,1);
// Hack... GCC optimizes away the code and returns true
result = !!(vgetq_lane_u32(x3,0) | vgetq_lane_u64(x4,1));
}
sigprocmask(SIG_SETMASK, (sigset_t*)&oldMask, NULLPTR);
signal(SIGILL, oldHandler);
return result;
# endif
}
#endif // CRYPTOPP_ARM_NEON_AVAILABLE
NAMESPACE_END

0
rdrand-masm.cmd Executable file → Normal file
View File

View File

@ -730,7 +730,7 @@ void CRYPTOPP_FASTCALL SHA256_HashBlocks_ARMV8A(word32 *state, const word32 *dat
STATE0 = vld1q_u32(&state[0]);
STATE1 = vld1q_u32(&state[4]);
const size_t BLOCKSIZE = 6;
const size_t BLOCKSIZE = 64;
while (length >= BLOCKSIZE)
{
// Save current hash