diff --git a/cipher/Makefile.am b/cipher/Makefile.am
index 1e67771e..3f00ed4a 100644
--- a/cipher/Makefile.am
+++ b/cipher/Makefile.am
@@ -1,148 +1,148 @@
# Makefile for cipher modules
# Copyright (C) 1998, 1999, 2000, 2001, 2002,
# 2003, 2009 Free Software Foundation, Inc.
#
# This file is part of Libgcrypt.
#
# Libgcrypt is free software; you can redistribute it and/or modify
# it under the terms of the GNU Lesser General Public License as
# published by the Free Software Foundation; either version 2.1 of
# the License, or (at your option) any later version.
#
# Libgcrypt is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU Lesser General Public License for more details.
#
# You should have received a copy of the GNU Lesser General Public
# License along with this program; if not, see .
# Process this file with automake to produce Makefile.in
# Need to include ../src in addition to top_srcdir because gcrypt.h is
# a built header.
AM_CPPFLAGS = -I../src -I$(top_srcdir)/src -I../mpi -I$(top_srcdir)/mpi
AM_CFLAGS = $(GPG_ERROR_CFLAGS)
AM_CCASFLAGS = $(NOEXECSTACK_FLAGS)
EXTRA_DIST = gost-s-box.c
CLEANFILES = gost-s-box
DISTCLEANFILES = gost-sb.h
noinst_LTLIBRARIES = libcipher.la
GCRYPT_MODULES = @GCRYPT_CIPHERS@ @GCRYPT_PUBKEY_CIPHERS@ \
@GCRYPT_DIGESTS@ @GCRYPT_KDFS@
libcipher_la_DEPENDENCIES = $(GCRYPT_MODULES)
libcipher_la_LIBADD = $(GCRYPT_MODULES)
libcipher_la_SOURCES = \
cipher.c cipher-internal.h \
cipher-cbc.c \
cipher-cfb.c \
cipher-ofb.c \
cipher-ctr.c \
cipher-aeswrap.c \
cipher-ccm.c \
cipher-cmac.c \
cipher-gcm.c cipher-gcm-intel-pclmul.c cipher-gcm-armv7-neon.S \
cipher-gcm-armv8-aarch32-ce.S cipher-gcm-armv8-aarch64-ce.S \
cipher-poly1305.c \
cipher-ocb.c \
cipher-xts.c \
cipher-eax.c \
cipher-selftest.c cipher-selftest.h \
pubkey.c pubkey-internal.h pubkey-util.c \
md.c \
mac.c mac-internal.h \
mac-hmac.c mac-cmac.c mac-gmac.c mac-poly1305.c \
poly1305.c poly1305-internal.h \
kdf.c kdf-internal.h \
hmac-tests.c \
bithelp.h \
bufhelp.h \
primegen.c \
hash-common.c hash-common.h \
dsa-common.c rsa-common.c \
sha1.h
EXTRA_libcipher_la_SOURCES = \
asm-common-amd64.h \
asm-common-aarch64.h \
asm-poly1305-amd64.h \
arcfour.c arcfour-amd64.S \
blowfish.c blowfish-amd64.S blowfish-arm.S \
cast5.c cast5-amd64.S cast5-arm.S \
chacha20.c chacha20-amd64-ssse3.S chacha20-amd64-avx2.S \
chacha20-armv7-neon.S chacha20-aarch64.S \
crc.c crc-intel-pclmul.c \
des.c des-amd64.S \
dsa.c \
elgamal.c \
ecc.c ecc-curves.c ecc-misc.c ecc-common.h \
ecc-ecdsa.c ecc-eddsa.c ecc-gost.c \
idea.c \
gost28147.c gost.h \
gostr3411-94.c \
md4.c \
md5.c \
rijndael.c rijndael-internal.h rijndael-tables.h \
rijndael-aesni.c rijndael-padlock.c \
rijndael-amd64.S rijndael-arm.S \
rijndael-ssse3-amd64.c rijndael-ssse3-amd64-asm.S \
rijndael-armv8-ce.c rijndael-armv8-aarch32-ce.S \
rijndael-armv8-aarch64-ce.S rijndael-aarch64.S \
rmd160.c \
rsa.c \
salsa20.c salsa20-amd64.S salsa20-armv7-neon.S \
scrypt.c \
seed.c \
serpent.c serpent-sse2-amd64.S \
serpent-avx2-amd64.S serpent-armv7-neon.S \
sha1.c sha1-ssse3-amd64.S sha1-avx-amd64.S sha1-avx-bmi2-amd64.S \
- sha1-armv7-neon.S sha1-armv8-aarch32-ce.S sha1-armv8-aarch64-ce.S \
- sha1-intel-shaext.c \
+ sha1-avx2-bmi2-amd64.S sha1-armv7-neon.S sha1-armv8-aarch32-ce.S \
+ sha1-armv8-aarch64-ce.S sha1-intel-shaext.c \
sha256.c sha256-ssse3-amd64.S sha256-avx-amd64.S \
sha256-avx2-bmi2-amd64.S \
sha256-armv8-aarch32-ce.S sha256-armv8-aarch64-ce.S \
sha256-intel-shaext.c \
sha512.c sha512-ssse3-amd64.S sha512-avx-amd64.S \
sha512-avx2-bmi2-amd64.S \
sha512-armv7-neon.S sha512-arm.S \
sm3.c \
keccak.c keccak_permute_32.h keccak_permute_64.h keccak-armv7-neon.S \
stribog.c \
tiger.c \
whirlpool.c whirlpool-sse2-amd64.S \
twofish.c twofish-amd64.S twofish-arm.S twofish-aarch64.S \
twofish-avx2-amd64.S \
rfc2268.c \
camellia.c camellia.h camellia-glue.c camellia-aesni-avx-amd64.S \
camellia-aesni-avx2-amd64.S camellia-arm.S camellia-aarch64.S \
blake2.c \
blake2b-amd64-avx2.S blake2s-amd64-avx.S
gost28147.lo: gost-sb.h
gost-sb.h: gost-s-box
./gost-s-box $@
gost-s-box: gost-s-box.c
$(CC_FOR_BUILD) -o $@ $(srcdir)/gost-s-box.c
if ENABLE_O_FLAG_MUNGING
o_flag_munging = sed -e 's/-O\([2-9s][2-9s]*\)/-O1/' -e 's/-Ofast/-O1/g'
else
o_flag_munging = cat
endif
# We need to lower the optimization for this module.
tiger.o: $(srcdir)/tiger.c
`echo $(COMPILE) -c $(srcdir)/tiger.c | $(o_flag_munging) `
tiger.lo: $(srcdir)/tiger.c
`echo $(LTCOMPILE) -c $(srcdir)/tiger.c | $(o_flag_munging) `
diff --git a/cipher/hash-common.h b/cipher/hash-common.h
index 23f81ed7..0b3ade11 100644
--- a/cipher/hash-common.h
+++ b/cipher/hash-common.h
@@ -1,59 +1,62 @@
/* hash-common.h - Declarations of common code for hash algorithms.
* Copyright (C) 2008 Free Software Foundation, Inc.
*
* This file is part of Libgcrypt.
*
* Libgcrypt is free software; you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as
* published by the Free Software Foundation; either version 2.1 of
* the License, or (at your option) any later version.
*
* Libgcrypt is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this program; if not, see .
*/
#ifndef GCRY_HASH_COMMON_H
#define GCRY_HASH_COMMON_H
#include "types.h"
const char * _gcry_hash_selftest_check_one
/**/ (int algo,
int datamode, const void *data, size_t datalen,
const void *expect, size_t expectlen);
/* Type for the md_write helper function. */
typedef unsigned int (*_gcry_md_block_write_t) (void *c,
const unsigned char *blks,
size_t nblks);
#if (defined(USE_SHA512) || defined(USE_WHIRLPOOL))
/* SHA-512 and Whirlpool needs u64. SHA-512 needs larger buffer. */
# define MD_BLOCK_MAX_BLOCKSIZE 128
# define MD_NBLOCKS_TYPE u64
#else
# define MD_BLOCK_MAX_BLOCKSIZE 64
# define MD_NBLOCKS_TYPE u32
#endif
+/* SHA1 needs 2x64 bytes and SHA-512 needs 128 bytes. */
+#define MD_BLOCK_CTX_BUFFER_SIZE 128
+
typedef struct gcry_md_block_ctx
{
- byte buf[MD_BLOCK_MAX_BLOCKSIZE];
+ byte buf[MD_BLOCK_CTX_BUFFER_SIZE];
MD_NBLOCKS_TYPE nblocks;
MD_NBLOCKS_TYPE nblocks_high;
int count;
size_t blocksize;
_gcry_md_block_write_t bwrite;
} gcry_md_block_ctx_t;
void
_gcry_md_block_write( void *context, const void *inbuf_arg, size_t inlen);
#endif /*GCRY_HASH_COMMON_H*/
diff --git a/cipher/sha1-avx-amd64.S b/cipher/sha1-avx-amd64.S
index 5f5b9c0e..143e4066 100644
--- a/cipher/sha1-avx-amd64.S
+++ b/cipher/sha1-avx-amd64.S
@@ -1,426 +1,428 @@
/* sha1-avx-amd64.S - Intel AVX accelerated SHA-1 transform function
* Copyright (C) 2013 Jussi Kivilinna
*
* Based on sha1.c:
* Copyright (C) 1998, 2001, 2002, 2003, 2008 Free Software Foundation, Inc.
*
* This file is part of Libgcrypt.
*
* Libgcrypt is free software; you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as
* published by the Free Software Foundation; either version 2.1 of
* the License, or (at your option) any later version.
*
* Libgcrypt is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this program; if not, see .
*/
/*
* Intel SSSE3 accelerated SHA-1 implementation based on white paper:
* "Improving the Performance of the Secure Hash Algorithm (SHA-1)"
* http://software.intel.com/en-us/articles/improving-the-performance-of-the-secure-hash-algorithm-1
*/
#ifdef __x86_64__
#include
#if (defined(HAVE_COMPATIBLE_GCC_AMD64_PLATFORM_AS) || \
defined(HAVE_COMPATIBLE_GCC_WIN64_PLATFORM_AS)) && \
defined(HAVE_GCC_INLINE_ASM_AVX) && defined(USE_SHA1)
#ifdef __PIC__
# define RIP (%rip)
#else
# define RIP
#endif
#ifdef HAVE_COMPATIBLE_GCC_AMD64_PLATFORM_AS
# define ELF(...) __VA_ARGS__
#else
# define ELF(...) /*_*/
#endif
/* Context structure */
#define state_h0 0
#define state_h1 4
#define state_h2 8
#define state_h3 12
#define state_h4 16
/* Constants */
.text
#define K1 0x5A827999
#define K2 0x6ED9EBA1
#define K3 0x8F1BBCDC
#define K4 0xCA62C1D6
.align 16
.LK_XMM:
.LK1: .long K1, K1, K1, K1
.LK2: .long K2, K2, K2, K2
.LK3: .long K3, K3, K3, K3
.LK4: .long K4, K4, K4, K4
.Lbswap_shufb_ctl:
.long 0x00010203, 0x04050607, 0x08090a0b, 0x0c0d0e0f
/* Register macros */
#define RSTATE %r8
#define RDATA %r9
#define ROLDSTACK %r10
#define RNBLKS %r11
#define a %eax
#define b %ebx
#define c %ecx
#define d %edx
#define e %edi
#define RT0 %esi
#define RT1 %ebp
#define Wtmp0 %xmm0
#define Wtmp1 %xmm1
#define W0 %xmm2
#define W1 %xmm3
#define W2 %xmm4
#define W3 %xmm5
#define W4 %xmm6
#define W5 %xmm7
#define W6 %xmm8
#define W7 %xmm9
#define BSWAP_REG %xmm10
/* Round function macros. */
#define WK(i) (((i) & 15) * 4)(%rsp)
#define R_F1(a,b,c,d,e,i) \
movl c, RT0; \
addl WK(i), e; \
xorl d, RT0; \
movl a, RT1; \
andl b, RT0; \
shldl $30, b, b; \
xorl d, RT0; \
leal (RT0,e), e; \
shldl $5, RT1, RT1; \
addl RT1, e;
#define R_F2(a,b,c,d,e,i) \
movl c, RT0; \
addl WK(i), e; \
xorl b, RT0; \
shldl $30, b, b; \
xorl d, RT0; \
movl a, RT1; \
leal (RT0,e), e; \
shldl $5, RT1, RT1; \
addl RT1, e;
#define R_F3(a,b,c,d,e,i) \
movl c, RT0; \
movl b, RT1; \
xorl b, RT0; \
andl c, RT1; \
andl d, RT0; \
addl RT1, e; \
addl WK(i), e; \
shldl $30, b, b; \
movl a, RT1; \
leal (RT0,e), e; \
shldl $5, RT1, RT1; \
addl RT1, e;
#define R_F4(a,b,c,d,e,i) R_F2(a,b,c,d,e,i)
#define R(a,b,c,d,e,f,i) \
R_##f(a,b,c,d,e,i)
/* Input expansion macros. */
#define W_PRECALC_00_15_0(i, W, tmp0) \
vmovdqu (4*(i))(RDATA), tmp0;
#define W_PRECALC_00_15_1(i, W, tmp0) \
vpshufb BSWAP_REG, tmp0, W;
#define W_PRECALC_00_15_2(i, W, tmp0) \
vpaddd (.LK_XMM + ((i)/20)*16) RIP, W, tmp0;
#define W_PRECALC_00_15_3(i, W, tmp0) \
vmovdqa tmp0, WK(i&~3);
#define W_PRECALC_16_31_0(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1) \
vpalignr $8, W_m16, W_m12, W; \
vpsrldq $4, W_m04, tmp0; \
vpxor W_m08, W, W;
#define W_PRECALC_16_31_1(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1) \
vpxor W_m16, tmp0, tmp0; \
vpxor tmp0, W, W; \
vpslld $1, W, tmp0; \
vpslldq $12, W, tmp1; \
vpsrld $31, W, W;
#define W_PRECALC_16_31_2(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1) \
vpor W, tmp0, tmp0; \
vpsrld $30, tmp1, W; \
vpslld $2, tmp1, tmp1;
#define W_PRECALC_16_31_3(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1) \
vpxor W, tmp0, tmp0; \
vpxor tmp1, tmp0, W; \
vpaddd (.LK_XMM + ((i)/20)*16) RIP, W, tmp0; \
vmovdqa tmp0, WK((i)&~3);
#define W_PRECALC_32_79_0(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0) \
vpxor W_m28, W, W; \
vpalignr $8, W_m08, W_m04, tmp0;
#define W_PRECALC_32_79_1(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0) \
vpxor W_m16, W, W; \
vpxor tmp0, W, W;
#define W_PRECALC_32_79_2(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0) \
vpsrld $30, W, tmp0; \
vpslld $2, W, W;
#define W_PRECALC_32_79_3(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0) \
vpor W, tmp0, W; \
vpaddd (.LK_XMM + ((i)/20)*16) RIP, W, tmp0; \
vmovdqa tmp0, WK((i)&~3);
/*
* Transform nblks*64 bytes (nblks*16 32-bit words) at DATA.
*
* unsigned int
* _gcry_sha1_transform_amd64_avx (void *ctx, const unsigned char *data,
* size_t nblks)
*/
.globl _gcry_sha1_transform_amd64_avx
ELF(.type _gcry_sha1_transform_amd64_avx,@function)
.align 16
_gcry_sha1_transform_amd64_avx:
/* input:
* %rdi: ctx, CTX
* %rsi: data (64*nblks bytes)
* %rdx: nblks
*/
xorl %eax, %eax;
cmpq $0, %rdx;
jz .Lret;
vzeroupper;
movq %rdx, RNBLKS;
movq %rdi, RSTATE;
movq %rsi, RDATA;
pushq %rbx;
pushq %rbp;
movq %rsp, ROLDSTACK;
subq $(16*4), %rsp;
andq $(~31), %rsp;
/* Get the values of the chaining variables. */
movl state_h0(RSTATE), a;
movl state_h1(RSTATE), b;
movl state_h2(RSTATE), c;
movl state_h3(RSTATE), d;
movl state_h4(RSTATE), e;
vmovdqa .Lbswap_shufb_ctl RIP, BSWAP_REG;
/* Precalc 0-15. */
W_PRECALC_00_15_0(0, W0, Wtmp0);
W_PRECALC_00_15_1(1, W0, Wtmp0);
W_PRECALC_00_15_2(2, W0, Wtmp0);
W_PRECALC_00_15_3(3, W0, Wtmp0);
W_PRECALC_00_15_0(4, W7, Wtmp0);
W_PRECALC_00_15_1(5, W7, Wtmp0);
W_PRECALC_00_15_2(6, W7, Wtmp0);
W_PRECALC_00_15_3(7, W7, Wtmp0);
W_PRECALC_00_15_0(8, W6, Wtmp0);
W_PRECALC_00_15_1(9, W6, Wtmp0);
W_PRECALC_00_15_2(10, W6, Wtmp0);
W_PRECALC_00_15_3(11, W6, Wtmp0);
W_PRECALC_00_15_0(12, W5, Wtmp0);
W_PRECALC_00_15_1(13, W5, Wtmp0);
W_PRECALC_00_15_2(14, W5, Wtmp0);
W_PRECALC_00_15_3(15, W5, Wtmp0);
.align 8
.Loop:
addq $64, RDATA;
/* Transform 0-15 + Precalc 16-31. */
R( a, b, c, d, e, F1, 0 ); W_PRECALC_16_31_0(16, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
R( e, a, b, c, d, F1, 1 ); W_PRECALC_16_31_1(17, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
R( d, e, a, b, c, F1, 2 ); W_PRECALC_16_31_2(18, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
R( c, d, e, a, b, F1, 3 ); W_PRECALC_16_31_3(19, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
R( b, c, d, e, a, F1, 4 ); W_PRECALC_16_31_0(20, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
R( a, b, c, d, e, F1, 5 ); W_PRECALC_16_31_1(21, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
R( e, a, b, c, d, F1, 6 ); W_PRECALC_16_31_2(22, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
R( d, e, a, b, c, F1, 7 ); W_PRECALC_16_31_3(23, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
R( c, d, e, a, b, F1, 8 ); W_PRECALC_16_31_0(24, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
R( b, c, d, e, a, F1, 9 ); W_PRECALC_16_31_1(25, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
R( a, b, c, d, e, F1, 10 ); W_PRECALC_16_31_2(26, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
R( e, a, b, c, d, F1, 11 ); W_PRECALC_16_31_3(27, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
R( d, e, a, b, c, F1, 12 ); W_PRECALC_16_31_0(28, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
R( c, d, e, a, b, F1, 13 ); W_PRECALC_16_31_1(29, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
R( b, c, d, e, a, F1, 14 ); W_PRECALC_16_31_2(30, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
R( a, b, c, d, e, F1, 15 ); W_PRECALC_16_31_3(31, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
/* Transform 16-63 + Precalc 32-79. */
R( e, a, b, c, d, F1, 16 ); W_PRECALC_32_79_0(32, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( d, e, a, b, c, F1, 17 ); W_PRECALC_32_79_1(33, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( c, d, e, a, b, F1, 18 ); W_PRECALC_32_79_2(34, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( b, c, d, e, a, F1, 19 ); W_PRECALC_32_79_3(35, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( a, b, c, d, e, F2, 20 ); W_PRECALC_32_79_0(36, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( e, a, b, c, d, F2, 21 ); W_PRECALC_32_79_1(37, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( d, e, a, b, c, F2, 22 ); W_PRECALC_32_79_2(38, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( c, d, e, a, b, F2, 23 ); W_PRECALC_32_79_3(39, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( b, c, d, e, a, F2, 24 ); W_PRECALC_32_79_0(40, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( a, b, c, d, e, F2, 25 ); W_PRECALC_32_79_1(41, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( e, a, b, c, d, F2, 26 ); W_PRECALC_32_79_2(42, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( d, e, a, b, c, F2, 27 ); W_PRECALC_32_79_3(43, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( c, d, e, a, b, F2, 28 ); W_PRECALC_32_79_0(44, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( b, c, d, e, a, F2, 29 ); W_PRECALC_32_79_1(45, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( a, b, c, d, e, F2, 30 ); W_PRECALC_32_79_2(46, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( e, a, b, c, d, F2, 31 ); W_PRECALC_32_79_3(47, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( d, e, a, b, c, F2, 32 ); W_PRECALC_32_79_0(48, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0);
R( c, d, e, a, b, F2, 33 ); W_PRECALC_32_79_1(49, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0);
R( b, c, d, e, a, F2, 34 ); W_PRECALC_32_79_2(50, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0);
R( a, b, c, d, e, F2, 35 ); W_PRECALC_32_79_3(51, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0);
R( e, a, b, c, d, F2, 36 ); W_PRECALC_32_79_0(52, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0);
R( d, e, a, b, c, F2, 37 ); W_PRECALC_32_79_1(53, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0);
R( c, d, e, a, b, F2, 38 ); W_PRECALC_32_79_2(54, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0);
R( b, c, d, e, a, F2, 39 ); W_PRECALC_32_79_3(55, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0);
R( a, b, c, d, e, F3, 40 ); W_PRECALC_32_79_0(56, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0);
R( e, a, b, c, d, F3, 41 ); W_PRECALC_32_79_1(57, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0);
R( d, e, a, b, c, F3, 42 ); W_PRECALC_32_79_2(58, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0);
R( c, d, e, a, b, F3, 43 ); W_PRECALC_32_79_3(59, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0);
R( b, c, d, e, a, F3, 44 ); W_PRECALC_32_79_0(60, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0);
R( a, b, c, d, e, F3, 45 ); W_PRECALC_32_79_1(61, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0);
R( e, a, b, c, d, F3, 46 ); W_PRECALC_32_79_2(62, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0);
R( d, e, a, b, c, F3, 47 ); W_PRECALC_32_79_3(63, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0);
R( c, d, e, a, b, F3, 48 ); W_PRECALC_32_79_0(64, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( b, c, d, e, a, F3, 49 ); W_PRECALC_32_79_1(65, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( a, b, c, d, e, F3, 50 ); W_PRECALC_32_79_2(66, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( e, a, b, c, d, F3, 51 ); W_PRECALC_32_79_3(67, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( d, e, a, b, c, F3, 52 ); W_PRECALC_32_79_0(68, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( c, d, e, a, b, F3, 53 ); W_PRECALC_32_79_1(69, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( b, c, d, e, a, F3, 54 ); W_PRECALC_32_79_2(70, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( a, b, c, d, e, F3, 55 ); W_PRECALC_32_79_3(71, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( e, a, b, c, d, F3, 56 ); W_PRECALC_32_79_0(72, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( d, e, a, b, c, F3, 57 ); W_PRECALC_32_79_1(73, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( c, d, e, a, b, F3, 58 ); W_PRECALC_32_79_2(74, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( b, c, d, e, a, F3, 59 ); W_PRECALC_32_79_3(75, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( a, b, c, d, e, F4, 60 ); W_PRECALC_32_79_0(76, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( e, a, b, c, d, F4, 61 ); W_PRECALC_32_79_1(77, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( d, e, a, b, c, F4, 62 ); W_PRECALC_32_79_2(78, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( c, d, e, a, b, F4, 63 ); W_PRECALC_32_79_3(79, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
decq RNBLKS;
jz .Lend;
/* Transform 64-79 + Precalc 0-15 of next block. */
R( b, c, d, e, a, F4, 64 ); W_PRECALC_00_15_0(0, W0, Wtmp0);
R( a, b, c, d, e, F4, 65 ); W_PRECALC_00_15_1(1, W0, Wtmp0);
R( e, a, b, c, d, F4, 66 ); W_PRECALC_00_15_2(2, W0, Wtmp0);
R( d, e, a, b, c, F4, 67 ); W_PRECALC_00_15_3(3, W0, Wtmp0);
R( c, d, e, a, b, F4, 68 ); W_PRECALC_00_15_0(4, W7, Wtmp0);
R( b, c, d, e, a, F4, 69 ); W_PRECALC_00_15_1(5, W7, Wtmp0);
R( a, b, c, d, e, F4, 70 ); W_PRECALC_00_15_2(6, W7, Wtmp0);
R( e, a, b, c, d, F4, 71 ); W_PRECALC_00_15_3(7, W7, Wtmp0);
R( d, e, a, b, c, F4, 72 ); W_PRECALC_00_15_0(8, W6, Wtmp0);
R( c, d, e, a, b, F4, 73 ); W_PRECALC_00_15_1(9, W6, Wtmp0);
R( b, c, d, e, a, F4, 74 ); W_PRECALC_00_15_2(10, W6, Wtmp0);
R( a, b, c, d, e, F4, 75 ); W_PRECALC_00_15_3(11, W6, Wtmp0);
R( e, a, b, c, d, F4, 76 ); W_PRECALC_00_15_0(12, W5, Wtmp0);
R( d, e, a, b, c, F4, 77 ); W_PRECALC_00_15_1(13, W5, Wtmp0);
R( c, d, e, a, b, F4, 78 );
addl state_h0(RSTATE), a; W_PRECALC_00_15_2(14, W5, Wtmp0);
R( b, c, d, e, a, F4, 79 ); W_PRECALC_00_15_3(15, W5, Wtmp0);
/* Update the chaining variables. */
addl state_h3(RSTATE), d;
addl state_h2(RSTATE), c;
addl state_h1(RSTATE), b;
addl state_h4(RSTATE), e;
movl d, state_h3(RSTATE);
movl c, state_h2(RSTATE);
movl b, state_h1(RSTATE);
movl a, state_h0(RSTATE);
movl e, state_h4(RSTATE);
jmp .Loop;
.align 16
.Lend:
vzeroall;
/* Transform 64-79. */
R( b, c, d, e, a, F4, 64 );
R( a, b, c, d, e, F4, 65 );
R( e, a, b, c, d, F4, 66 );
R( d, e, a, b, c, F4, 67 );
R( c, d, e, a, b, F4, 68 );
R( b, c, d, e, a, F4, 69 );
R( a, b, c, d, e, F4, 70 );
R( e, a, b, c, d, F4, 71 );
R( d, e, a, b, c, F4, 72 );
R( c, d, e, a, b, F4, 73 );
R( b, c, d, e, a, F4, 74 );
R( a, b, c, d, e, F4, 75 );
R( e, a, b, c, d, F4, 76 );
R( d, e, a, b, c, F4, 77 );
R( c, d, e, a, b, F4, 78 );
addl state_h0(RSTATE), a;
R( b, c, d, e, a, F4, 79 );
/* Update the chaining variables. */
addl state_h3(RSTATE), d;
addl state_h2(RSTATE), c;
addl state_h1(RSTATE), b;
addl state_h4(RSTATE), e;
movl d, state_h3(RSTATE);
movl c, state_h2(RSTATE);
movl b, state_h1(RSTATE);
movl a, state_h0(RSTATE);
movl e, state_h4(RSTATE);
movq ROLDSTACK, %rsp;
popq %rbp;
popq %rbx;
/* burn_stack */
movl $(16*4 + 2*8 + 31), %eax;
.Lret:
ret;
+ELF(.size _gcry_sha1_transform_amd64_avx,
+ .-_gcry_sha1_transform_amd64_avx;)
#endif
#endif
diff --git a/cipher/sha1-avx-bmi2-amd64.S b/cipher/sha1-avx-bmi2-amd64.S
index 8292c3af..79ea24ef 100644
--- a/cipher/sha1-avx-bmi2-amd64.S
+++ b/cipher/sha1-avx-bmi2-amd64.S
@@ -1,424 +1,438 @@
/* sha1-avx-bmi2-amd64.S - Intel AVX/BMI2 accelerated SHA-1 transform function
* Copyright (C) 2013 Jussi Kivilinna
*
* Based on sha1.c:
* Copyright (C) 1998, 2001, 2002, 2003, 2008 Free Software Foundation, Inc.
*
* This file is part of Libgcrypt.
*
* Libgcrypt is free software; you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as
* published by the Free Software Foundation; either version 2.1 of
* the License, or (at your option) any later version.
*
* Libgcrypt is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this program; if not, see .
*/
/*
* Intel SSSE3 accelerated SHA-1 implementation based on white paper:
* "Improving the Performance of the Secure Hash Algorithm (SHA-1)"
* http://software.intel.com/en-us/articles/improving-the-performance-of-the-secure-hash-algorithm-1
*/
#ifdef __x86_64__
#include
#if (defined(HAVE_COMPATIBLE_GCC_AMD64_PLATFORM_AS) || \
defined(HAVE_COMPATIBLE_GCC_WIN64_PLATFORM_AS)) && \
defined(HAVE_GCC_INLINE_ASM_BMI2) && \
defined(HAVE_GCC_INLINE_ASM_AVX) && defined(USE_SHA1)
#ifdef __PIC__
# define RIP (%rip)
#else
# define RIP
#endif
#ifdef HAVE_COMPATIBLE_GCC_AMD64_PLATFORM_AS
# define ELF(...) __VA_ARGS__
#else
# define ELF(...) /*_*/
#endif
/* Context structure */
#define state_h0 0
#define state_h1 4
#define state_h2 8
#define state_h3 12
#define state_h4 16
/* Constants */
.text
-#define K1 0x5A827999
-#define K2 0x6ED9EBA1
-#define K3 0x8F1BBCDC
-#define K4 0xCA62C1D6
.align 16
-.LK_XMM:
-.LK1: .long K1, K1, K1, K1
-.LK2: .long K2, K2, K2, K2
-.LK3: .long K3, K3, K3, K3
-.LK4: .long K4, K4, K4, K4
-
.Lbswap_shufb_ctl:
.long 0x00010203, 0x04050607, 0x08090a0b, 0x0c0d0e0f
+.LK1: .long 0x5A827999
+.LK2: .long 0x6ED9EBA1
+.LK3: .long 0x8F1BBCDC
+.LK4: .long 0xCA62C1D6
+
/* Register macros */
#define RSTATE %r8
#define RDATA %r9
#define ROLDSTACK %r10
#define RNBLKS %r11
-#define a %eax
-#define b %ebx
-#define c %ecx
+#define a %esi
+#define b %edi
+#define c %ebp
#define d %edx
-#define e %edi
+#define e %ecx
+#define ne %ebx
-#define RT0 %esi
-#define RT1 %ebp
+#define RT0 %eax
+#define RT1 %r12d
#define Wtmp0 %xmm0
#define Wtmp1 %xmm1
#define W0 %xmm2
#define W1 %xmm3
#define W2 %xmm4
#define W3 %xmm5
#define W4 %xmm6
#define W5 %xmm7
#define W6 %xmm8
#define W7 %xmm9
#define BSWAP_REG %xmm10
+#define K1 %xmm11
+#define K2 %xmm12
+#define K3 %xmm13
+#define K4 %xmm14
+
/* Round function macros. */
#define WK(i) (((i) & 15) * 4)(%rsp)
#define R_F1(a,b,c,d,e,i) \
movl c, RT0; \
andn d, b, RT1; \
addl WK(i), e; \
andl b, RT0; \
rorxl $2, b, b; \
addl RT1, e; \
- leal (RT0,e), e; \
- rorxl $27, a, RT1; \
- addl RT1, e;
+ addl ne, a; \
+ leal (RT0,e), ne; \
+ rorxl $27, a, e;
#define R_F2(a,b,c,d,e,i) \
movl c, RT0; \
addl WK(i), e; \
xorl b, RT0; \
rorxl $2, b, b; \
xorl d, RT0; \
- leal (RT0,e), e; \
- rorxl $27, a, RT1; \
- addl RT1, e;
+ addl ne, a; \
+ leal (RT0,e), ne; \
+ rorxl $27, a, e;
#define R_F3(a,b,c,d,e,i) \
movl c, RT0; \
movl b, RT1; \
+ addl WK(i), e; \
xorl b, RT0; \
andl c, RT1; \
andl d, RT0; \
addl RT1, e; \
- addl WK(i), e; \
rorxl $2, b, b; \
- leal (RT0,e), e; \
- rorxl $27, a, RT1; \
- addl RT1, e;
+ addl ne, a; \
+ leal (RT0,e), ne; \
+ rorxl $27, a, e;
#define R_F4(a,b,c,d,e,i) R_F2(a,b,c,d,e,i)
#define R(a,b,c,d,e,f,i) \
R_##f(a,b,c,d,e,i)
/* Input expansion macros. */
#define W_PRECALC_00_15_0(i, W, tmp0) \
vmovdqu (4*(i))(RDATA), tmp0;
#define W_PRECALC_00_15_1(i, W, tmp0) \
vpshufb BSWAP_REG, tmp0, W;
-#define W_PRECALC_00_15_2(i, W, tmp0) \
- vpaddd (.LK_XMM + ((i)/20)*16) RIP, W, tmp0;
+#define W_PRECALC_00_15_2(i, W, tmp0, K) \
+ vpaddd K, W, tmp0;
#define W_PRECALC_00_15_3(i, W, tmp0) \
vmovdqa tmp0, WK(i&~3);
#define W_PRECALC_16_31_0(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1) \
vpalignr $8, W_m16, W_m12, W; \
vpsrldq $4, W_m04, tmp0; \
vpxor W_m08, W, W;
#define W_PRECALC_16_31_1(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1) \
vpxor W_m16, tmp0, tmp0; \
vpxor tmp0, W, W; \
vpslld $1, W, tmp0; \
vpslldq $12, W, tmp1; \
vpsrld $31, W, W;
#define W_PRECALC_16_31_2(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1) \
vpor W, tmp0, tmp0; \
vpsrld $30, tmp1, W; \
vpslld $2, tmp1, tmp1;
-#define W_PRECALC_16_31_3(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1) \
+#define W_PRECALC_16_31_3(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1, K) \
vpxor W, tmp0, tmp0; \
vpxor tmp1, tmp0, W; \
- vpaddd (.LK_XMM + ((i)/20)*16) RIP, W, tmp0; \
+ vpaddd K, W, tmp0; \
vmovdqa tmp0, WK((i)&~3);
#define W_PRECALC_32_79_0(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0) \
vpxor W_m28, W, W; \
vpalignr $8, W_m08, W_m04, tmp0;
#define W_PRECALC_32_79_1(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0) \
vpxor W_m16, W, W; \
vpxor tmp0, W, W;
#define W_PRECALC_32_79_2(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0) \
vpsrld $30, W, tmp0; \
vpslld $2, W, W;
-#define W_PRECALC_32_79_3(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0) \
+#define W_PRECALC_32_79_3(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0, K) \
vpor W, tmp0, W; \
- vpaddd (.LK_XMM + ((i)/20)*16) RIP, W, tmp0; \
+ vpaddd K, W, tmp0; \
vmovdqa tmp0, WK((i)&~3);
/*
* Transform nblks*64 bytes (nblks*16 32-bit words) at DATA.
*
* unsigned int
* _gcry_sha1_transform_amd64_avx_bmi2 (void *ctx, const unsigned char *data,
* size_t nblks)
*/
.globl _gcry_sha1_transform_amd64_avx_bmi2
ELF(.type _gcry_sha1_transform_amd64_avx_bmi2,@function)
.align 16
_gcry_sha1_transform_amd64_avx_bmi2:
/* input:
* %rdi: ctx, CTX
* %rsi: data (64*nblks bytes)
* %rdx: nblks
*/
xorl %eax, %eax;
cmpq $0, %rdx;
jz .Lret;
vzeroupper;
movq %rdx, RNBLKS;
movq %rdi, RSTATE;
movq %rsi, RDATA;
pushq %rbx;
pushq %rbp;
+ pushq %r12;
movq %rsp, ROLDSTACK;
subq $(16*4), %rsp;
andq $(~31), %rsp;
/* Get the values of the chaining variables. */
movl state_h0(RSTATE), a;
movl state_h1(RSTATE), b;
movl state_h2(RSTATE), c;
movl state_h3(RSTATE), d;
movl state_h4(RSTATE), e;
+ xorl ne, ne;
vmovdqa .Lbswap_shufb_ctl RIP, BSWAP_REG;
+ vpbroadcastd .LK1 RIP, K1;
+ vpbroadcastd .LK2 RIP, K2;
+ vpbroadcastd .LK3 RIP, K3;
+ vpbroadcastd .LK4 RIP, K4;
/* Precalc 0-15. */
W_PRECALC_00_15_0(0, W0, Wtmp0);
W_PRECALC_00_15_1(1, W0, Wtmp0);
- W_PRECALC_00_15_2(2, W0, Wtmp0);
+ W_PRECALC_00_15_2(2, W0, Wtmp0, K1);
W_PRECALC_00_15_3(3, W0, Wtmp0);
W_PRECALC_00_15_0(4, W7, Wtmp0);
W_PRECALC_00_15_1(5, W7, Wtmp0);
- W_PRECALC_00_15_2(6, W7, Wtmp0);
+ W_PRECALC_00_15_2(6, W7, Wtmp0, K1);
W_PRECALC_00_15_3(7, W7, Wtmp0);
W_PRECALC_00_15_0(8, W6, Wtmp0);
W_PRECALC_00_15_1(9, W6, Wtmp0);
- W_PRECALC_00_15_2(10, W6, Wtmp0);
+ W_PRECALC_00_15_2(10, W6, Wtmp0, K1);
W_PRECALC_00_15_3(11, W6, Wtmp0);
W_PRECALC_00_15_0(12, W5, Wtmp0);
W_PRECALC_00_15_1(13, W5, Wtmp0);
- W_PRECALC_00_15_2(14, W5, Wtmp0);
+ W_PRECALC_00_15_2(14, W5, Wtmp0, K1);
W_PRECALC_00_15_3(15, W5, Wtmp0);
.align 8
.Loop:
addq $64, RDATA;
/* Transform 0-15 + Precalc 16-31. */
R( a, b, c, d, e, F1, 0 ); W_PRECALC_16_31_0(16, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
R( e, a, b, c, d, F1, 1 ); W_PRECALC_16_31_1(17, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
R( d, e, a, b, c, F1, 2 ); W_PRECALC_16_31_2(18, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
- R( c, d, e, a, b, F1, 3 ); W_PRECALC_16_31_3(19, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
+ R( c, d, e, a, b, F1, 3 ); W_PRECALC_16_31_3(19, W4, W5, W6, W7, W0, Wtmp0, Wtmp1, K1);
R( b, c, d, e, a, F1, 4 ); W_PRECALC_16_31_0(20, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
R( a, b, c, d, e, F1, 5 ); W_PRECALC_16_31_1(21, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
R( e, a, b, c, d, F1, 6 ); W_PRECALC_16_31_2(22, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
- R( d, e, a, b, c, F1, 7 ); W_PRECALC_16_31_3(23, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
+ R( d, e, a, b, c, F1, 7 ); W_PRECALC_16_31_3(23, W3, W4, W5, W6, W7, Wtmp0, Wtmp1, K2);
R( c, d, e, a, b, F1, 8 ); W_PRECALC_16_31_0(24, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
R( b, c, d, e, a, F1, 9 ); W_PRECALC_16_31_1(25, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
R( a, b, c, d, e, F1, 10 ); W_PRECALC_16_31_2(26, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
- R( e, a, b, c, d, F1, 11 ); W_PRECALC_16_31_3(27, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
+ R( e, a, b, c, d, F1, 11 ); W_PRECALC_16_31_3(27, W2, W3, W4, W5, W6, Wtmp0, Wtmp1, K2);
R( d, e, a, b, c, F1, 12 ); W_PRECALC_16_31_0(28, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
R( c, d, e, a, b, F1, 13 ); W_PRECALC_16_31_1(29, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
R( b, c, d, e, a, F1, 14 ); W_PRECALC_16_31_2(30, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
- R( a, b, c, d, e, F1, 15 ); W_PRECALC_16_31_3(31, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
+ R( a, b, c, d, e, F1, 15 ); W_PRECALC_16_31_3(31, W1, W2, W3, W4, W5, Wtmp0, Wtmp1, K2);
/* Transform 16-63 + Precalc 32-79. */
R( e, a, b, c, d, F1, 16 ); W_PRECALC_32_79_0(32, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( d, e, a, b, c, F1, 17 ); W_PRECALC_32_79_1(33, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( c, d, e, a, b, F1, 18 ); W_PRECALC_32_79_2(34, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
- R( b, c, d, e, a, F1, 19 ); W_PRECALC_32_79_3(35, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
+ R( b, c, d, e, a, F1, 19 ); W_PRECALC_32_79_3(35, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0, K2);
R( a, b, c, d, e, F2, 20 ); W_PRECALC_32_79_0(36, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( e, a, b, c, d, F2, 21 ); W_PRECALC_32_79_1(37, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( d, e, a, b, c, F2, 22 ); W_PRECALC_32_79_2(38, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
- R( c, d, e, a, b, F2, 23 ); W_PRECALC_32_79_3(39, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
+ R( c, d, e, a, b, F2, 23 ); W_PRECALC_32_79_3(39, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0, K2);
R( b, c, d, e, a, F2, 24 ); W_PRECALC_32_79_0(40, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( a, b, c, d, e, F2, 25 ); W_PRECALC_32_79_1(41, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( e, a, b, c, d, F2, 26 ); W_PRECALC_32_79_2(42, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
- R( d, e, a, b, c, F2, 27 ); W_PRECALC_32_79_3(43, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
+ R( d, e, a, b, c, F2, 27 ); W_PRECALC_32_79_3(43, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0, K3);
R( c, d, e, a, b, F2, 28 ); W_PRECALC_32_79_0(44, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( b, c, d, e, a, F2, 29 ); W_PRECALC_32_79_1(45, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( a, b, c, d, e, F2, 30 ); W_PRECALC_32_79_2(46, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
- R( e, a, b, c, d, F2, 31 ); W_PRECALC_32_79_3(47, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
+ R( e, a, b, c, d, F2, 31 ); W_PRECALC_32_79_3(47, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0, K3);
R( d, e, a, b, c, F2, 32 ); W_PRECALC_32_79_0(48, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0);
R( c, d, e, a, b, F2, 33 ); W_PRECALC_32_79_1(49, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0);
R( b, c, d, e, a, F2, 34 ); W_PRECALC_32_79_2(50, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0);
- R( a, b, c, d, e, F2, 35 ); W_PRECALC_32_79_3(51, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0);
+ R( a, b, c, d, e, F2, 35 ); W_PRECALC_32_79_3(51, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0, K3);
R( e, a, b, c, d, F2, 36 ); W_PRECALC_32_79_0(52, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0);
R( d, e, a, b, c, F2, 37 ); W_PRECALC_32_79_1(53, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0);
R( c, d, e, a, b, F2, 38 ); W_PRECALC_32_79_2(54, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0);
- R( b, c, d, e, a, F2, 39 ); W_PRECALC_32_79_3(55, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0);
+ R( b, c, d, e, a, F2, 39 ); W_PRECALC_32_79_3(55, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0, K3);
R( a, b, c, d, e, F3, 40 ); W_PRECALC_32_79_0(56, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0);
R( e, a, b, c, d, F3, 41 ); W_PRECALC_32_79_1(57, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0);
R( d, e, a, b, c, F3, 42 ); W_PRECALC_32_79_2(58, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0);
- R( c, d, e, a, b, F3, 43 ); W_PRECALC_32_79_3(59, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0);
+ R( c, d, e, a, b, F3, 43 ); W_PRECALC_32_79_3(59, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0, K3);
R( b, c, d, e, a, F3, 44 ); W_PRECALC_32_79_0(60, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0);
R( a, b, c, d, e, F3, 45 ); W_PRECALC_32_79_1(61, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0);
R( e, a, b, c, d, F3, 46 ); W_PRECALC_32_79_2(62, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0);
- R( d, e, a, b, c, F3, 47 ); W_PRECALC_32_79_3(63, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0);
+ R( d, e, a, b, c, F3, 47 ); W_PRECALC_32_79_3(63, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0, K4);
R( c, d, e, a, b, F3, 48 ); W_PRECALC_32_79_0(64, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( b, c, d, e, a, F3, 49 ); W_PRECALC_32_79_1(65, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( a, b, c, d, e, F3, 50 ); W_PRECALC_32_79_2(66, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
- R( e, a, b, c, d, F3, 51 ); W_PRECALC_32_79_3(67, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
+ R( e, a, b, c, d, F3, 51 ); W_PRECALC_32_79_3(67, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0, K4);
R( d, e, a, b, c, F3, 52 ); W_PRECALC_32_79_0(68, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( c, d, e, a, b, F3, 53 ); W_PRECALC_32_79_1(69, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( b, c, d, e, a, F3, 54 ); W_PRECALC_32_79_2(70, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
- R( a, b, c, d, e, F3, 55 ); W_PRECALC_32_79_3(71, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
+ R( a, b, c, d, e, F3, 55 ); W_PRECALC_32_79_3(71, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0, K4);
R( e, a, b, c, d, F3, 56 ); W_PRECALC_32_79_0(72, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( d, e, a, b, c, F3, 57 ); W_PRECALC_32_79_1(73, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( c, d, e, a, b, F3, 58 ); W_PRECALC_32_79_2(74, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
- R( b, c, d, e, a, F3, 59 ); W_PRECALC_32_79_3(75, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
+ R( b, c, d, e, a, F3, 59 ); W_PRECALC_32_79_3(75, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0, K4);
R( a, b, c, d, e, F4, 60 ); W_PRECALC_32_79_0(76, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( e, a, b, c, d, F4, 61 ); W_PRECALC_32_79_1(77, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( d, e, a, b, c, F4, 62 ); W_PRECALC_32_79_2(78, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
- R( c, d, e, a, b, F4, 63 ); W_PRECALC_32_79_3(79, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
+ R( c, d, e, a, b, F4, 63 ); W_PRECALC_32_79_3(79, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0, K4);
decq RNBLKS;
jz .Lend;
/* Transform 64-79 + Precalc 0-15 of next block. */
R( b, c, d, e, a, F4, 64 ); W_PRECALC_00_15_0(0, W0, Wtmp0);
R( a, b, c, d, e, F4, 65 ); W_PRECALC_00_15_1(1, W0, Wtmp0);
- R( e, a, b, c, d, F4, 66 ); W_PRECALC_00_15_2(2, W0, Wtmp0);
+ R( e, a, b, c, d, F4, 66 ); W_PRECALC_00_15_2(2, W0, Wtmp0, K1);
R( d, e, a, b, c, F4, 67 ); W_PRECALC_00_15_3(3, W0, Wtmp0);
R( c, d, e, a, b, F4, 68 ); W_PRECALC_00_15_0(4, W7, Wtmp0);
R( b, c, d, e, a, F4, 69 ); W_PRECALC_00_15_1(5, W7, Wtmp0);
- R( a, b, c, d, e, F4, 70 ); W_PRECALC_00_15_2(6, W7, Wtmp0);
+ R( a, b, c, d, e, F4, 70 ); W_PRECALC_00_15_2(6, W7, Wtmp0, K1);
R( e, a, b, c, d, F4, 71 ); W_PRECALC_00_15_3(7, W7, Wtmp0);
R( d, e, a, b, c, F4, 72 ); W_PRECALC_00_15_0(8, W6, Wtmp0);
R( c, d, e, a, b, F4, 73 ); W_PRECALC_00_15_1(9, W6, Wtmp0);
- R( b, c, d, e, a, F4, 74 ); W_PRECALC_00_15_2(10, W6, Wtmp0);
+ R( b, c, d, e, a, F4, 74 ); W_PRECALC_00_15_2(10, W6, Wtmp0, K1);
R( a, b, c, d, e, F4, 75 ); W_PRECALC_00_15_3(11, W6, Wtmp0);
R( e, a, b, c, d, F4, 76 ); W_PRECALC_00_15_0(12, W5, Wtmp0);
R( d, e, a, b, c, F4, 77 ); W_PRECALC_00_15_1(13, W5, Wtmp0);
R( c, d, e, a, b, F4, 78 );
- addl state_h0(RSTATE), a; W_PRECALC_00_15_2(14, W5, Wtmp0);
+ addl state_h0(RSTATE), a; W_PRECALC_00_15_2(14, W5, Wtmp0, K1);
R( b, c, d, e, a, F4, 79 ); W_PRECALC_00_15_3(15, W5, Wtmp0);
+ addl ne, a;
+ xorl ne, ne;
/* Update the chaining variables. */
addl state_h3(RSTATE), d;
addl state_h2(RSTATE), c;
addl state_h1(RSTATE), b;
addl state_h4(RSTATE), e;
movl d, state_h3(RSTATE);
movl c, state_h2(RSTATE);
movl b, state_h1(RSTATE);
movl a, state_h0(RSTATE);
movl e, state_h4(RSTATE);
jmp .Loop;
.align 16
.Lend:
vzeroall;
/* Transform 64-79. */
R( b, c, d, e, a, F4, 64 );
R( a, b, c, d, e, F4, 65 );
R( e, a, b, c, d, F4, 66 );
R( d, e, a, b, c, F4, 67 );
R( c, d, e, a, b, F4, 68 );
R( b, c, d, e, a, F4, 69 );
R( a, b, c, d, e, F4, 70 );
R( e, a, b, c, d, F4, 71 );
R( d, e, a, b, c, F4, 72 );
R( c, d, e, a, b, F4, 73 );
R( b, c, d, e, a, F4, 74 );
R( a, b, c, d, e, F4, 75 );
R( e, a, b, c, d, F4, 76 );
R( d, e, a, b, c, F4, 77 );
R( c, d, e, a, b, F4, 78 );
addl state_h0(RSTATE), a;
R( b, c, d, e, a, F4, 79 );
+ addl ne, a;
+ xorl ne, ne;
/* Update the chaining variables. */
addl state_h3(RSTATE), d;
addl state_h2(RSTATE), c;
addl state_h1(RSTATE), b;
addl state_h4(RSTATE), e;
movl d, state_h3(RSTATE);
movl c, state_h2(RSTATE);
movl b, state_h1(RSTATE);
movl a, state_h0(RSTATE);
movl e, state_h4(RSTATE);
movq ROLDSTACK, %rsp;
+ popq %r12;
popq %rbp;
popq %rbx;
/* burn_stack */
- movl $(16*4 + 2*8 + 31), %eax;
+ movl $(16*4 + 3*8 + 31), %eax;
.Lret:
ret;
+ELF(.size _gcry_sha1_transform_amd64_avx_bmi2,
+ .-_gcry_sha1_transform_amd64_avx_bmi2;)
#endif
#endif
diff --git a/cipher/sha1-avx2-bmi2-amd64.S b/cipher/sha1-avx2-bmi2-amd64.S
new file mode 100644
index 00000000..c666290f
--- /dev/null
+++ b/cipher/sha1-avx2-bmi2-amd64.S
@@ -0,0 +1,570 @@
+/* sha1-avx2-bmi2-amd64.S - Intel AVX2/BMI2 accelerated SHA-1 transform function
+ * Copyright (C) 2019 Jussi Kivilinna
+ *
+ * Based on sha1.c:
+ * Copyright (C) 1998, 2001, 2002, 2003, 2008 Free Software Foundation, Inc.
+ *
+ * This file is part of Libgcrypt.
+ *
+ * Libgcrypt is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU Lesser General Public License as
+ * published by the Free Software Foundation; either version 2.1 of
+ * the License, or (at your option) any later version.
+ *
+ * Libgcrypt is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU Lesser General Public License for more details.
+ *
+ * You should have received a copy of the GNU Lesser General Public
+ * License along with this program; if not, see .
+ */
+
+/*
+ * Intel SSSE3 accelerated SHA-1 implementation based on white paper:
+ * "Improving the Performance of the Secure Hash Algorithm (SHA-1)"
+ * http://software.intel.com/en-us/articles/improving-the-performance-of-the-secure-hash-algorithm-1
+ */
+
+#ifdef __x86_64__
+#include
+
+#if (defined(HAVE_COMPATIBLE_GCC_AMD64_PLATFORM_AS) || \
+ defined(HAVE_COMPATIBLE_GCC_WIN64_PLATFORM_AS)) && \
+ defined(HAVE_GCC_INLINE_ASM_BMI2) && defined(HAVE_GCC_INLINE_ASM_AVX) && \
+ defined(HAVE_GCC_INLINE_ASM_AVX2) && defined(USE_SHA1)
+
+#ifdef __PIC__
+# define RIP (%rip)
+#else
+# define RIP
+#endif
+
+
+#ifdef HAVE_COMPATIBLE_GCC_AMD64_PLATFORM_AS
+# define ELF(...) __VA_ARGS__
+#else
+# define ELF(...) /*_*/
+#endif
+
+
+/* Context structure */
+
+#define state_h0 0
+#define state_h1 4
+#define state_h2 8
+#define state_h3 12
+#define state_h4 16
+
+
+/* Constants */
+
+#define WK_STACK_WORDS (80 * 2)
+
+.text
+.align 16
+.Lbswap_shufb_ctl:
+ .long 0x00010203, 0x04050607, 0x08090a0b, 0x0c0d0e0f
+
+.LK1: .long 0x5A827999
+.LK2: .long 0x6ED9EBA1
+.LK3: .long 0x8F1BBCDC
+.LK4: .long 0xCA62C1D6
+
+
+/* Register macros */
+
+#define RSTATE %r8
+#define RDATA %r9
+#define ROLDSTACK %r10
+#define RNBLKS %r11
+
+#define a %eax
+#define b %ebx
+#define c %ecx
+#define d %edx
+#define e %edi
+#define ne %r12d
+
+#define RT0 %esi
+#define RT1 %ebp
+
+#define Wtmp0 %ymm0
+#define Wtmp1 %ymm1
+#define Wtmp0x %xmm0
+#define Wtmp1x %xmm1
+
+#define W0 %ymm2
+#define W1 %ymm3
+#define W2 %ymm4
+#define W3 %ymm5
+#define W4 %ymm6
+#define W5 %ymm7
+#define W6 %ymm8
+#define W7 %ymm9
+
+#define BSWAP_REG %ymm10
+
+#define K1 %ymm11
+#define K2 %ymm12
+#define K3 %ymm13
+#define K4 %ymm14
+
+
+/* Round function macros. */
+
+#define WK(i,block) ((block) * 16 + ((i) / 4) * 32 + ((i) % 4) * 4)(%rsp)
+#define PRE_WK(i) ((i) * 4 * 2)(%rsp)
+
+#define R_F1(a,b,c,d,e,i,block) \
+ movl c, RT0; \
+ andn d, b, RT1; \
+ addl WK(i,block), e; \
+ andl b, RT0; \
+ leal (a,ne), a; \
+ rorxl $2, b, b; \
+ addl RT1, e; \
+ rorxl $27, a, ne; \
+ addl RT0, e;
+
+#define R_F2(a,b,c,d,e,i,block) \
+ addl WK(i,block), e; \
+ movl c, RT0; \
+ xorl b, RT0; \
+ leal (a,ne), a; \
+ rorxl $2, b, b; \
+ xorl d, RT0; \
+ addl RT0, e; \
+ rorxl $27, a, ne;
+
+#define R_F3(a,b,c,d,e,i,block) \
+ movl c, RT0; \
+ addl WK(i,block), e; \
+ movl b, RT1; \
+ xorl b, RT0; \
+ leal (a,ne), a; \
+ rorxl $2, b, b; \
+ andl c, RT1; \
+ addl RT1, e; \
+ andl d, RT0; \
+ rorxl $27, a, ne; \
+ addl RT0, e;
+
+#define R_F4(a,b,c,d,e,i,block) R_F2(a,b,c,d,e,i,block)
+
+#define R(a,b,c,d,e,f,i,block) \
+ R_##f(a,b,c,d,e,i,block)
+
+
+/* Input expansion macros. */
+
+#define W_PRECALC_00_15_0(i, W, tmp0) \
+ vmovdqu (4*(i))(RDATA), tmp0##x; \
+ vinserti128 $1, (4*(i) + 64)(RDATA), tmp0, tmp0;
+
+#define W_PRECALC_00_15_1(i, W, tmp0) \
+ vpshufb BSWAP_REG, tmp0, W;
+
+#define W_PRECALC_00_15_2(i, W, tmp0, K) \
+ vpaddd K, W, tmp0;
+
+#define W_PRECALC_00_15_3(i, W, tmp0) \
+ vmovdqa tmp0, PRE_WK((i)&~3);
+
+#define W_PRECALC_16_31_0(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1) \
+ vpalignr $8, W_m16, W_m12, W; \
+ vpsrldq $4, W_m04, tmp0; \
+ vpxor W_m08, W, W;
+
+#define W_PRECALC_16_31_1(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1) \
+ vpxor W_m16, tmp0, tmp0; \
+ vpxor tmp0, W, W; \
+ vpslld $1, W, tmp0; \
+ vpslldq $12, W, tmp1; \
+ vpsrld $31, W, W;
+
+#define W_PRECALC_16_31_2(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1) \
+ vpor W, tmp0, tmp0; \
+ vpsrld $30, tmp1, W; \
+ vpslld $2, tmp1, tmp1;
+
+#define W_PRECALC_16_31_3(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1, K) \
+ vpxor W, tmp0, tmp0; \
+ vpxor tmp1, tmp0, W; \
+ vpaddd K, W, tmp0; \
+ vmovdqa tmp0, PRE_WK((i)&~3);
+
+#define W_PRECALC_32_79_0(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0) \
+ vpxor W_m28, W, W; \
+ vpalignr $8, W_m08, W_m04, tmp0;
+
+#define W_PRECALC_32_79_1(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0) \
+ vpxor W_m16, W, W; \
+ vpxor tmp0, W, W;
+
+#define W_PRECALC_32_79_2(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0) \
+ vpsrld $30, W, tmp0; \
+ vpslld $2, W, W;
+
+#define W_PRECALC_32_79_3(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0, K) \
+ vpor W, tmp0, W; \
+ vpaddd K, W, tmp0; \
+ vmovdqa tmp0, PRE_WK((i)&~3);
+
+
+/*
+ * Transform 2*nblks*64 bytes (2*nblks*16 32-bit words) at DATA.
+ *
+ * unsigned int
+ * _gcry_sha1_transform_amd64_avx2_bmi2 (void *ctx, const unsigned char *data,
+ * size_t nblks)
+ */
+.globl _gcry_sha1_transform_amd64_avx2_bmi2
+ELF(.type _gcry_sha1_transform_amd64_avx2_bmi2,@function)
+.align 16
+_gcry_sha1_transform_amd64_avx2_bmi2:
+ /* input:
+ * %rdi: ctx, CTX
+ * %rsi: data (64*nblks bytes)
+ * %rdx: nblks (multiple of 2, larger than 0)
+ */
+
+ vzeroupper;
+
+ movq %rdx, RNBLKS;
+ movq %rdi, RSTATE;
+ movq %rsi, RDATA;
+ pushq %rbx;
+ pushq %rbp;
+ pushq %r12;
+
+ movq %rsp, ROLDSTACK;
+
+ subq $(WK_STACK_WORDS*4), %rsp;
+ andq $(~63), %rsp;
+
+ /* Get the values of the chaining variables. */
+ movl state_h0(RSTATE), a;
+ movl state_h1(RSTATE), b;
+ movl state_h2(RSTATE), c;
+ movl state_h3(RSTATE), d;
+ movl state_h4(RSTATE), e;
+ xorl ne, ne;
+
+ vbroadcasti128 .Lbswap_shufb_ctl RIP, BSWAP_REG;
+ vpbroadcastd .LK1 RIP, K1;
+ vpbroadcastd .LK2 RIP, K2;
+ vpbroadcastd .LK3 RIP, K3;
+ vpbroadcastd .LK4 RIP, K4;
+
+ /* Precalc 0-31 for block 1 & 2. */
+ W_PRECALC_00_15_0(0, W0, Wtmp0);
+ W_PRECALC_00_15_1(1, W0, Wtmp0);
+ W_PRECALC_00_15_2(2, W0, Wtmp0, K1);
+ W_PRECALC_00_15_3(3, W0, Wtmp0);
+ W_PRECALC_00_15_0(4, W7, Wtmp0);
+ W_PRECALC_00_15_1(5, W7, Wtmp0);
+ W_PRECALC_00_15_2(6, W7, Wtmp0, K1);
+ W_PRECALC_00_15_3(7, W7, Wtmp0);
+ W_PRECALC_00_15_0(8, W6, Wtmp0);
+ W_PRECALC_00_15_1(9, W6, Wtmp0);
+ W_PRECALC_00_15_2(10, W6, Wtmp0, K1);
+ W_PRECALC_00_15_3(11, W6, Wtmp0);
+ W_PRECALC_00_15_0(12, W5, Wtmp0);
+ W_PRECALC_00_15_1(13, W5, Wtmp0);
+ W_PRECALC_00_15_2(14, W5, Wtmp0, K1);
+ W_PRECALC_00_15_3(15, W5, Wtmp0);
+ W_PRECALC_16_31_0(16, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
+ W_PRECALC_16_31_1(17, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
+ W_PRECALC_16_31_2(18, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
+ W_PRECALC_16_31_3(19, W4, W5, W6, W7, W0, Wtmp0, Wtmp1, K1);
+ W_PRECALC_16_31_0(20, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
+ W_PRECALC_16_31_1(21, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
+ W_PRECALC_16_31_2(22, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
+ W_PRECALC_16_31_3(23, W3, W4, W5, W6, W7, Wtmp0, Wtmp1, K2);
+ W_PRECALC_16_31_0(24, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
+ W_PRECALC_16_31_1(25, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
+ W_PRECALC_16_31_2(26, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
+ W_PRECALC_16_31_3(27, W2, W3, W4, W5, W6, Wtmp0, Wtmp1, K2);
+ W_PRECALC_16_31_0(28, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
+ W_PRECALC_16_31_1(29, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
+ W_PRECALC_16_31_2(30, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
+ W_PRECALC_16_31_3(31, W1, W2, W3, W4, W5, Wtmp0, Wtmp1, K2);
+
+.align 8
+.Loop:
+ addq $(2 * 64), RDATA;
+
+ /* Transform 0-15 for block 1 + Precalc 32-47 for block 1 & 2. */
+ R( a, b, c, d, e, F1, 0, 0 ); W_PRECALC_32_79_0(32, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
+ R( e, a, b, c, d, F1, 1, 0 ); W_PRECALC_32_79_1(33, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
+ R( d, e, a, b, c, F1, 2, 0 ); W_PRECALC_32_79_2(34, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
+ R( c, d, e, a, b, F1, 3, 0 ); W_PRECALC_32_79_3(35, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0, K2);
+ R( b, c, d, e, a, F1, 4, 0 ); W_PRECALC_32_79_0(36, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
+ R( a, b, c, d, e, F1, 5, 0 ); W_PRECALC_32_79_1(37, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
+ R( e, a, b, c, d, F1, 6, 0 ); W_PRECALC_32_79_2(38, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
+ R( d, e, a, b, c, F1, 7, 0 ); W_PRECALC_32_79_3(39, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0, K2);
+ R( c, d, e, a, b, F1, 8, 0 ); W_PRECALC_32_79_0(40, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
+ R( b, c, d, e, a, F1, 9, 0 ); W_PRECALC_32_79_1(41, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
+ R( a, b, c, d, e, F1, 10, 0 ); W_PRECALC_32_79_2(42, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
+ R( e, a, b, c, d, F1, 11, 0 ); W_PRECALC_32_79_3(43, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0, K3);
+ R( d, e, a, b, c, F1, 12, 0 ); W_PRECALC_32_79_0(44, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
+ R( c, d, e, a, b, F1, 13, 0 ); W_PRECALC_32_79_1(45, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
+ R( b, c, d, e, a, F1, 14, 0 ); W_PRECALC_32_79_2(46, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
+ R( a, b, c, d, e, F1, 15, 0 ); W_PRECALC_32_79_3(47, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0, K3);
+
+ /* Transform 16-47 for block 1 + Precalc 48-79 for block 1 & 2. */
+ R( e, a, b, c, d, F1, 16, 0 ); W_PRECALC_32_79_0(48, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0);
+ R( d, e, a, b, c, F1, 17, 0 ); W_PRECALC_32_79_1(49, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0);
+ R( c, d, e, a, b, F1, 18, 0 ); W_PRECALC_32_79_2(50, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0);
+ R( b, c, d, e, a, F1, 19, 0 ); W_PRECALC_32_79_3(51, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0, K3);
+ R( a, b, c, d, e, F2, 20, 0 ); W_PRECALC_32_79_0(52, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0);
+ R( e, a, b, c, d, F2, 21, 0 ); W_PRECALC_32_79_1(53, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0);
+ R( d, e, a, b, c, F2, 22, 0 ); W_PRECALC_32_79_2(54, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0);
+ R( c, d, e, a, b, F2, 23, 0 ); W_PRECALC_32_79_3(55, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0, K3);
+ R( b, c, d, e, a, F2, 24, 0 ); W_PRECALC_32_79_0(56, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0);
+ R( a, b, c, d, e, F2, 25, 0 ); W_PRECALC_32_79_1(57, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0);
+ R( e, a, b, c, d, F2, 26, 0 ); W_PRECALC_32_79_2(58, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0);
+ R( d, e, a, b, c, F2, 27, 0 ); W_PRECALC_32_79_3(59, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0, K3);
+ R( c, d, e, a, b, F2, 28, 0 ); W_PRECALC_32_79_0(60, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0);
+ R( b, c, d, e, a, F2, 29, 0 ); W_PRECALC_32_79_1(61, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0);
+ R( a, b, c, d, e, F2, 30, 0 ); W_PRECALC_32_79_2(62, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0);
+ R( e, a, b, c, d, F2, 31, 0 ); W_PRECALC_32_79_3(63, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0, K4);
+ R( d, e, a, b, c, F2, 32, 0 ); W_PRECALC_32_79_0(64, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
+ R( c, d, e, a, b, F2, 33, 0 ); W_PRECALC_32_79_1(65, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
+ R( b, c, d, e, a, F2, 34, 0 ); W_PRECALC_32_79_2(66, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
+ R( a, b, c, d, e, F2, 35, 0 ); W_PRECALC_32_79_3(67, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0, K4);
+ R( e, a, b, c, d, F2, 36, 0 ); W_PRECALC_32_79_0(68, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
+ R( d, e, a, b, c, F2, 37, 0 ); W_PRECALC_32_79_1(69, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
+ R( c, d, e, a, b, F2, 38, 0 ); W_PRECALC_32_79_2(70, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
+ R( b, c, d, e, a, F2, 39, 0 ); W_PRECALC_32_79_3(71, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0, K4);
+ R( a, b, c, d, e, F3, 40, 0 ); W_PRECALC_32_79_0(72, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
+ R( e, a, b, c, d, F3, 41, 0 ); W_PRECALC_32_79_1(73, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
+ R( d, e, a, b, c, F3, 42, 0 ); W_PRECALC_32_79_2(74, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
+ R( c, d, e, a, b, F3, 43, 0 ); W_PRECALC_32_79_3(75, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0, K4);
+ R( b, c, d, e, a, F3, 44, 0 ); W_PRECALC_32_79_0(76, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
+ R( a, b, c, d, e, F3, 45, 0 ); W_PRECALC_32_79_1(77, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
+ R( e, a, b, c, d, F3, 46, 0 ); W_PRECALC_32_79_2(78, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
+ R( d, e, a, b, c, F3, 47, 0 ); W_PRECALC_32_79_3(79, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0, K4);
+
+ /* Transform 48-79 for block 1. */
+ R( c, d, e, a, b, F3, 48, 0 );
+ R( b, c, d, e, a, F3, 49, 0 );
+ R( a, b, c, d, e, F3, 50, 0 );
+ R( e, a, b, c, d, F3, 51, 0 );
+ R( d, e, a, b, c, F3, 52, 0 );
+ R( c, d, e, a, b, F3, 53, 0 );
+ R( b, c, d, e, a, F3, 54, 0 );
+ R( a, b, c, d, e, F3, 55, 0 );
+ R( e, a, b, c, d, F3, 56, 0 );
+ R( d, e, a, b, c, F3, 57, 0 );
+ R( c, d, e, a, b, F3, 58, 0 );
+ R( b, c, d, e, a, F3, 59, 0 );
+ R( a, b, c, d, e, F4, 60, 0 );
+ R( e, a, b, c, d, F4, 61, 0 );
+ R( d, e, a, b, c, F4, 62, 0 );
+ R( c, d, e, a, b, F4, 63, 0 );
+ R( b, c, d, e, a, F4, 64, 0 );
+ R( a, b, c, d, e, F4, 65, 0 );
+ R( e, a, b, c, d, F4, 66, 0 );
+ R( d, e, a, b, c, F4, 67, 0 );
+ R( c, d, e, a, b, F4, 68, 0 );
+ R( b, c, d, e, a, F4, 69, 0 );
+ R( a, b, c, d, e, F4, 70, 0 );
+ R( e, a, b, c, d, F4, 71, 0 );
+ R( d, e, a, b, c, F4, 72, 0 );
+ R( c, d, e, a, b, F4, 73, 0 );
+ R( b, c, d, e, a, F4, 74, 0 );
+ R( a, b, c, d, e, F4, 75, 0 );
+ R( e, a, b, c, d, F4, 76, 0 );
+ R( d, e, a, b, c, F4, 77, 0 );
+ R( c, d, e, a, b, F4, 78, 0 );
+ addl state_h0(RSTATE), a;
+ R( b, c, d, e, a, F4, 79, 0 );
+ addl ne, a;
+ xorl ne, ne;
+
+ /* Update the chaining variables. */
+ addl state_h3(RSTATE), d;
+ addl state_h2(RSTATE), c;
+ addl state_h1(RSTATE), b;
+ addl state_h4(RSTATE), e;
+
+ movl d, state_h3(RSTATE);
+ movl c, state_h2(RSTATE);
+ movl b, state_h1(RSTATE);
+ movl a, state_h0(RSTATE);
+ movl e, state_h4(RSTATE);
+
+ /* Transform 0-47 for block 2. */
+ R( a, b, c, d, e, F1, 0, 1 );
+ R( e, a, b, c, d, F1, 1, 1 );
+ R( d, e, a, b, c, F1, 2, 1 );
+ R( c, d, e, a, b, F1, 3, 1 );
+ R( b, c, d, e, a, F1, 4, 1 );
+ R( a, b, c, d, e, F1, 5, 1 );
+ R( e, a, b, c, d, F1, 6, 1 );
+ R( d, e, a, b, c, F1, 7, 1 );
+ R( c, d, e, a, b, F1, 8, 1 );
+ R( b, c, d, e, a, F1, 9, 1 );
+ R( a, b, c, d, e, F1, 10, 1 );
+ R( e, a, b, c, d, F1, 11, 1 );
+ R( d, e, a, b, c, F1, 12, 1 );
+ R( c, d, e, a, b, F1, 13, 1 );
+ R( b, c, d, e, a, F1, 14, 1 );
+ R( a, b, c, d, e, F1, 15, 1 );
+ R( e, a, b, c, d, F1, 16, 1 );
+ R( d, e, a, b, c, F1, 17, 1 );
+ R( c, d, e, a, b, F1, 18, 1 );
+ R( b, c, d, e, a, F1, 19, 1 );
+ R( a, b, c, d, e, F2, 20, 1 );
+ R( e, a, b, c, d, F2, 21, 1 );
+ R( d, e, a, b, c, F2, 22, 1 );
+ R( c, d, e, a, b, F2, 23, 1 );
+ R( b, c, d, e, a, F2, 24, 1 );
+ R( a, b, c, d, e, F2, 25, 1 );
+ R( e, a, b, c, d, F2, 26, 1 );
+ R( d, e, a, b, c, F2, 27, 1 );
+ R( c, d, e, a, b, F2, 28, 1 );
+ R( b, c, d, e, a, F2, 29, 1 );
+ R( a, b, c, d, e, F2, 30, 1 );
+ R( e, a, b, c, d, F2, 31, 1 );
+ R( d, e, a, b, c, F2, 32, 1 );
+ R( c, d, e, a, b, F2, 33, 1 );
+ R( b, c, d, e, a, F2, 34, 1 );
+ R( a, b, c, d, e, F2, 35, 1 );
+ R( e, a, b, c, d, F2, 36, 1 );
+ R( d, e, a, b, c, F2, 37, 1 );
+ R( c, d, e, a, b, F2, 38, 1 );
+ R( b, c, d, e, a, F2, 39, 1 );
+ R( a, b, c, d, e, F3, 40, 1 );
+ R( e, a, b, c, d, F3, 41, 1 );
+ R( d, e, a, b, c, F3, 42, 1 );
+ R( c, d, e, a, b, F3, 43, 1 );
+ R( b, c, d, e, a, F3, 44, 1 );
+ R( a, b, c, d, e, F3, 45, 1 );
+ R( e, a, b, c, d, F3, 46, 1 );
+ R( d, e, a, b, c, F3, 47, 1 );
+
+ addq $-2, RNBLKS;
+ jz .Lend;
+
+ /* Transform 48-79 for block 2 + Precalc 0-31 for next two blocks. */
+ R( c, d, e, a, b, F3, 48, 1 ); W_PRECALC_00_15_0(0, W0, Wtmp0);
+ R( b, c, d, e, a, F3, 49, 1 ); W_PRECALC_00_15_1(1, W0, Wtmp0);
+ R( a, b, c, d, e, F3, 50, 1 ); W_PRECALC_00_15_2(2, W0, Wtmp0, K1);
+ R( e, a, b, c, d, F3, 51, 1 ); W_PRECALC_00_15_3(3, W0, Wtmp0);
+ R( d, e, a, b, c, F3, 52, 1 ); W_PRECALC_00_15_0(4, W7, Wtmp0);
+ R( c, d, e, a, b, F3, 53, 1 ); W_PRECALC_00_15_1(5, W7, Wtmp0);
+ R( b, c, d, e, a, F3, 54, 1 ); W_PRECALC_00_15_2(6, W7, Wtmp0, K1);
+ R( a, b, c, d, e, F3, 55, 1 ); W_PRECALC_00_15_3(7, W7, Wtmp0);
+ R( e, a, b, c, d, F3, 56, 1 ); W_PRECALC_00_15_0(8, W6, Wtmp0);
+ R( d, e, a, b, c, F3, 57, 1 ); W_PRECALC_00_15_1(9, W6, Wtmp0);
+ R( c, d, e, a, b, F3, 58, 1 ); W_PRECALC_00_15_2(10, W6, Wtmp0, K1);
+ R( b, c, d, e, a, F3, 59, 1 ); W_PRECALC_00_15_3(11, W6, Wtmp0);
+ R( a, b, c, d, e, F4, 60, 1 ); W_PRECALC_00_15_0(12, W5, Wtmp0);
+ R( e, a, b, c, d, F4, 61, 1 ); W_PRECALC_00_15_1(13, W5, Wtmp0);
+ R( d, e, a, b, c, F4, 62, 1 ); W_PRECALC_00_15_2(14, W5, Wtmp0, K1);
+ R( c, d, e, a, b, F4, 63, 1 ); W_PRECALC_00_15_3(15, W5, Wtmp0);
+ R( b, c, d, e, a, F4, 64, 1 ); W_PRECALC_16_31_0(16, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
+ R( a, b, c, d, e, F4, 65, 1 ); W_PRECALC_16_31_1(17, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
+ R( e, a, b, c, d, F4, 66, 1 ); W_PRECALC_16_31_2(18, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
+ R( d, e, a, b, c, F4, 67, 1 ); W_PRECALC_16_31_3(19, W4, W5, W6, W7, W0, Wtmp0, Wtmp1, K1);
+ R( c, d, e, a, b, F4, 68, 1 ); W_PRECALC_16_31_0(20, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
+ R( b, c, d, e, a, F4, 69, 1 ); W_PRECALC_16_31_1(21, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
+ R( a, b, c, d, e, F4, 70, 1 ); W_PRECALC_16_31_2(22, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
+ R( e, a, b, c, d, F4, 71, 1 ); W_PRECALC_16_31_3(23, W3, W4, W5, W6, W7, Wtmp0, Wtmp1, K2);
+ R( d, e, a, b, c, F4, 72, 1 ); W_PRECALC_16_31_0(24, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
+ R( c, d, e, a, b, F4, 73, 1 ); W_PRECALC_16_31_1(25, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
+ R( b, c, d, e, a, F4, 74, 1 ); W_PRECALC_16_31_2(26, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
+ R( a, b, c, d, e, F4, 75, 1 ); W_PRECALC_16_31_3(27, W2, W3, W4, W5, W6, Wtmp0, Wtmp1, K2);
+ R( e, a, b, c, d, F4, 76, 1 ); W_PRECALC_16_31_0(28, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
+ R( d, e, a, b, c, F4, 77, 1 ); W_PRECALC_16_31_1(29, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
+ R( c, d, e, a, b, F4, 78, 1 ); W_PRECALC_16_31_2(30, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
+ addl state_h0(RSTATE), a; W_PRECALC_16_31_3(31, W1, W2, W3, W4, W5, Wtmp0, Wtmp1, K2);
+ R( b, c, d, e, a, F4, 79, 1 );
+ addl ne, a;
+ xorl ne, ne;
+
+ /* Update the chaining variables. */
+ addl state_h3(RSTATE), d;
+ addl state_h2(RSTATE), c;
+ addl state_h1(RSTATE), b;
+ addl state_h4(RSTATE), e;
+
+ movl d, state_h3(RSTATE);
+ movl c, state_h2(RSTATE);
+ movl b, state_h1(RSTATE);
+ movl a, state_h0(RSTATE);
+ movl e, state_h4(RSTATE);
+
+ jmp .Loop;
+
+.align 16
+.Lend:
+ vzeroall;
+
+ /* Transform 48-79 for block 2. */
+ R( c, d, e, a, b, F3, 48, 1 );
+ R( b, c, d, e, a, F3, 49, 1 );
+ R( a, b, c, d, e, F3, 50, 1 );
+ R( e, a, b, c, d, F3, 51, 1 );
+ R( d, e, a, b, c, F3, 52, 1 );
+ R( c, d, e, a, b, F3, 53, 1 );
+ R( b, c, d, e, a, F3, 54, 1 );
+ R( a, b, c, d, e, F3, 55, 1 );
+ R( e, a, b, c, d, F3, 56, 1 );
+ R( d, e, a, b, c, F3, 57, 1 );
+ R( c, d, e, a, b, F3, 58, 1 );
+ R( b, c, d, e, a, F3, 59, 1 );
+ R( a, b, c, d, e, F4, 60, 1 );
+ R( e, a, b, c, d, F4, 61, 1 );
+ R( d, e, a, b, c, F4, 62, 1 );
+ R( c, d, e, a, b, F4, 63, 1 );
+ R( b, c, d, e, a, F4, 64, 1 );
+ R( a, b, c, d, e, F4, 65, 1 );
+ R( e, a, b, c, d, F4, 66, 1 );
+ R( d, e, a, b, c, F4, 67, 1 );
+ R( c, d, e, a, b, F4, 68, 1 );
+ R( b, c, d, e, a, F4, 69, 1 );
+ R( a, b, c, d, e, F4, 70, 1 );
+ R( e, a, b, c, d, F4, 71, 1 );
+ R( d, e, a, b, c, F4, 72, 1 );
+ R( c, d, e, a, b, F4, 73, 1 );
+ R( b, c, d, e, a, F4, 74, 1 );
+ R( a, b, c, d, e, F4, 75, 1 );
+ R( e, a, b, c, d, F4, 76, 1 );
+ R( d, e, a, b, c, F4, 77, 1 );
+ R( c, d, e, a, b, F4, 78, 1 );
+ addl state_h0(RSTATE), a;
+ R( b, c, d, e, a, F4, 79, 1 );
+ addl ne, a;
+ xorl ne, ne;
+
+ /* Update the chaining variables. */
+ addl state_h3(RSTATE), d;
+ addl state_h2(RSTATE), c;
+ addl state_h1(RSTATE), b;
+ addl state_h4(RSTATE), e;
+
+ movl d, state_h3(RSTATE);
+ movl c, state_h2(RSTATE);
+ movl b, state_h1(RSTATE);
+ movl a, state_h0(RSTATE);
+ movl e, state_h4(RSTATE);
+
+ movq ROLDSTACK, %rsp;
+
+ popq %r12;
+ popq %rbp;
+ popq %rbx;
+
+ /* burn_stack */
+ movl $((WK_STACK_WORDS)*4 + 3*8 + 31), %eax;
+
+ ret;
+ELF(.size _gcry_sha1_transform_amd64_avx2_bmi2,
+ .-_gcry_sha1_transform_amd64_avx2_bmi2;)
+
+#endif
+#endif
diff --git a/cipher/sha1-ssse3-amd64.S b/cipher/sha1-ssse3-amd64.S
index 2b439476..421bebec 100644
--- a/cipher/sha1-ssse3-amd64.S
+++ b/cipher/sha1-ssse3-amd64.S
@@ -1,434 +1,436 @@
/* sha1-ssse3-amd64.S - Intel SSSE3 accelerated SHA-1 transform function
* Copyright (C) 2013 Jussi Kivilinna
*
* Based on sha1.c:
* Copyright (C) 1998, 2001, 2002, 2003, 2008 Free Software Foundation, Inc.
*
* This file is part of Libgcrypt.
*
* Libgcrypt is free software; you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as
* published by the Free Software Foundation; either version 2.1 of
* the License, or (at your option) any later version.
*
* Libgcrypt is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this program; if not, see .
*/
/*
* Intel SSSE3 accelerated SHA-1 implementation based on white paper:
* "Improving the Performance of the Secure Hash Algorithm (SHA-1)"
* http://software.intel.com/en-us/articles/improving-the-performance-of-the-secure-hash-algorithm-1
*/
#ifdef __x86_64__
#include
#if (defined(HAVE_COMPATIBLE_GCC_AMD64_PLATFORM_AS) || \
defined(HAVE_COMPATIBLE_GCC_WIN64_PLATFORM_AS)) && \
defined(HAVE_GCC_INLINE_ASM_SSSE3) && defined(USE_SHA1)
#ifdef __PIC__
# define RIP (%rip)
#else
# define RIP
#endif
#ifdef HAVE_COMPATIBLE_GCC_AMD64_PLATFORM_AS
# define ELF(...) __VA_ARGS__
#else
# define ELF(...) /*_*/
#endif
/* Context structure */
#define state_h0 0
#define state_h1 4
#define state_h2 8
#define state_h3 12
#define state_h4 16
/* Constants */
.text
#define K1 0x5A827999
#define K2 0x6ED9EBA1
#define K3 0x8F1BBCDC
#define K4 0xCA62C1D6
.align 16
.LK_XMM:
.LK1: .long K1, K1, K1, K1
.LK2: .long K2, K2, K2, K2
.LK3: .long K3, K3, K3, K3
.LK4: .long K4, K4, K4, K4
.Lbswap_shufb_ctl:
.long 0x00010203, 0x04050607, 0x08090a0b, 0x0c0d0e0f
/* Register macros */
#define RSTATE %r8
#define RDATA %r9
#define ROLDSTACK %r10
#define RNBLKS %r11
#define a %eax
#define b %ebx
#define c %ecx
#define d %edx
#define e %edi
#define RT0 %esi
#define RT1 %ebp
#define Wtmp0 %xmm0
#define Wtmp1 %xmm1
#define W0 %xmm2
#define W1 %xmm3
#define W2 %xmm4
#define W3 %xmm5
#define W4 %xmm6
#define W5 %xmm7
#define W6 %xmm8
#define W7 %xmm9
#define BSWAP_REG %xmm10
/* Round function macros. */
#define WK(i) (((i) & 15) * 4)(%rsp)
#define R_F1(a,b,c,d,e,i) \
movl c, RT0; \
addl WK(i), e; \
xorl d, RT0; \
movl a, RT1; \
andl b, RT0; \
roll $30, b; \
xorl d, RT0; \
leal (RT0,e), e; \
roll $5, RT1; \
addl RT1, e;
#define R_F2(a,b,c,d,e,i) \
movl c, RT0; \
addl WK(i), e; \
xorl b, RT0; \
roll $30, b; \
xorl d, RT0; \
movl a, RT1; \
leal (RT0,e), e; \
roll $5, RT1; \
addl RT1, e;
#define R_F3(a,b,c,d,e,i) \
movl c, RT0; \
movl b, RT1; \
xorl b, RT0; \
andl c, RT1; \
andl d, RT0; \
addl RT1, e; \
addl WK(i), e; \
roll $30, b; \
movl a, RT1; \
leal (RT0,e), e; \
roll $5, RT1; \
addl RT1, e;
#define R_F4(a,b,c,d,e,i) R_F2(a,b,c,d,e,i)
#define R(a,b,c,d,e,f,i) \
R_##f(a,b,c,d,e,i)
/* Input expansion macros. */
#define W_PRECALC_00_15_0(i, W, tmp0) \
movdqu (4*(i))(RDATA), tmp0;
#define W_PRECALC_00_15_1(i, W, tmp0) \
pshufb BSWAP_REG, tmp0; \
movdqa tmp0, W;
#define W_PRECALC_00_15_2(i, W, tmp0) \
paddd (.LK_XMM + ((i)/20)*16) RIP, tmp0;
#define W_PRECALC_00_15_3(i, W, tmp0) \
movdqa tmp0, WK(i&~3);
#define W_PRECALC_16_31_0(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1) \
movdqa W_m12, W; \
palignr $8, W_m16, W; \
movdqa W_m04, tmp0; \
psrldq $4, tmp0; \
pxor W_m08, W;
#define W_PRECALC_16_31_1(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1) \
pxor W_m16, tmp0; \
pxor tmp0, W; \
movdqa W, tmp1; \
movdqa W, tmp0; \
pslldq $12, tmp1;
#define W_PRECALC_16_31_2(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1) \
psrld $31, W; \
pslld $1, tmp0; \
por W, tmp0; \
movdqa tmp1, W; \
psrld $30, tmp1; \
pslld $2, W;
#define W_PRECALC_16_31_3(i, W, W_m04, W_m08, W_m12, W_m16, tmp0, tmp1) \
pxor W, tmp0; \
pxor tmp1, tmp0; \
movdqa tmp0, W; \
paddd (.LK_XMM + ((i)/20)*16) RIP, tmp0; \
movdqa tmp0, WK((i)&~3);
#define W_PRECALC_32_79_0(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0) \
movdqa W_m04, tmp0; \
pxor W_m28, W; \
palignr $8, W_m08, tmp0;
#define W_PRECALC_32_79_1(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0) \
pxor W_m16, W; \
pxor tmp0, W; \
movdqa W, tmp0;
#define W_PRECALC_32_79_2(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0) \
psrld $30, W; \
pslld $2, tmp0; \
por W, tmp0;
#define W_PRECALC_32_79_3(i, W, W_m04, W_m08, W_m12, W_m16, W_m20, W_m24, W_m28, tmp0) \
movdqa tmp0, W; \
paddd (.LK_XMM + ((i)/20)*16) RIP, tmp0; \
movdqa tmp0, WK((i)&~3);
#define CLEAR_REG(reg) pxor reg, reg;
/*
* Transform nblks*64 bytes (nblks*16 32-bit words) at DATA.
*
* unsigned int
* _gcry_sha1_transform_amd64_ssse3 (void *ctx, const unsigned char *data,
* size_t nblks)
*/
.globl _gcry_sha1_transform_amd64_ssse3
ELF(.type _gcry_sha1_transform_amd64_ssse3,@function)
.align 16
_gcry_sha1_transform_amd64_ssse3:
/* input:
* %rdi: ctx, CTX
* %rsi: data (64*nblks bytes)
* %rdx: nblks
*/
xorl %eax, %eax;
cmpq $0, %rdx;
jz .Lret;
movq %rdx, RNBLKS;
movq %rdi, RSTATE;
movq %rsi, RDATA;
pushq %rbx;
pushq %rbp;
movq %rsp, ROLDSTACK;
subq $(16*4), %rsp;
andq $(~31), %rsp;
/* Get the values of the chaining variables. */
movl state_h0(RSTATE), a;
movl state_h1(RSTATE), b;
movl state_h2(RSTATE), c;
movl state_h3(RSTATE), d;
movl state_h4(RSTATE), e;
movdqa .Lbswap_shufb_ctl RIP, BSWAP_REG;
/* Precalc 0-15. */
W_PRECALC_00_15_0(0, W0, Wtmp0);
W_PRECALC_00_15_1(1, W0, Wtmp0);
W_PRECALC_00_15_2(2, W0, Wtmp0);
W_PRECALC_00_15_3(3, W0, Wtmp0);
W_PRECALC_00_15_0(4, W7, Wtmp0);
W_PRECALC_00_15_1(5, W7, Wtmp0);
W_PRECALC_00_15_2(6, W7, Wtmp0);
W_PRECALC_00_15_3(7, W7, Wtmp0);
W_PRECALC_00_15_0(8, W6, Wtmp0);
W_PRECALC_00_15_1(9, W6, Wtmp0);
W_PRECALC_00_15_2(10, W6, Wtmp0);
W_PRECALC_00_15_3(11, W6, Wtmp0);
W_PRECALC_00_15_0(12, W5, Wtmp0);
W_PRECALC_00_15_1(13, W5, Wtmp0);
W_PRECALC_00_15_2(14, W5, Wtmp0);
W_PRECALC_00_15_3(15, W5, Wtmp0);
.align 8
.Loop:
addq $64, RDATA;
/* Transform 0-15 + Precalc 16-31. */
R( a, b, c, d, e, F1, 0 ); W_PRECALC_16_31_0(16, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
R( e, a, b, c, d, F1, 1 ); W_PRECALC_16_31_1(17, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
R( d, e, a, b, c, F1, 2 ); W_PRECALC_16_31_2(18, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
R( c, d, e, a, b, F1, 3 ); W_PRECALC_16_31_3(19, W4, W5, W6, W7, W0, Wtmp0, Wtmp1);
R( b, c, d, e, a, F1, 4 ); W_PRECALC_16_31_0(20, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
R( a, b, c, d, e, F1, 5 ); W_PRECALC_16_31_1(21, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
R( e, a, b, c, d, F1, 6 ); W_PRECALC_16_31_2(22, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
R( d, e, a, b, c, F1, 7 ); W_PRECALC_16_31_3(23, W3, W4, W5, W6, W7, Wtmp0, Wtmp1);
R( c, d, e, a, b, F1, 8 ); W_PRECALC_16_31_0(24, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
R( b, c, d, e, a, F1, 9 ); W_PRECALC_16_31_1(25, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
R( a, b, c, d, e, F1, 10 ); W_PRECALC_16_31_2(26, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
R( e, a, b, c, d, F1, 11 ); W_PRECALC_16_31_3(27, W2, W3, W4, W5, W6, Wtmp0, Wtmp1);
R( d, e, a, b, c, F1, 12 ); W_PRECALC_16_31_0(28, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
R( c, d, e, a, b, F1, 13 ); W_PRECALC_16_31_1(29, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
R( b, c, d, e, a, F1, 14 ); W_PRECALC_16_31_2(30, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
R( a, b, c, d, e, F1, 15 ); W_PRECALC_16_31_3(31, W1, W2, W3, W4, W5, Wtmp0, Wtmp1);
/* Transform 16-63 + Precalc 32-79. */
R( e, a, b, c, d, F1, 16 ); W_PRECALC_32_79_0(32, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( d, e, a, b, c, F1, 17 ); W_PRECALC_32_79_1(33, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( c, d, e, a, b, F1, 18 ); W_PRECALC_32_79_2(34, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( b, c, d, e, a, F1, 19 ); W_PRECALC_32_79_3(35, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( a, b, c, d, e, F2, 20 ); W_PRECALC_32_79_0(36, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( e, a, b, c, d, F2, 21 ); W_PRECALC_32_79_1(37, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( d, e, a, b, c, F2, 22 ); W_PRECALC_32_79_2(38, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( c, d, e, a, b, F2, 23 ); W_PRECALC_32_79_3(39, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( b, c, d, e, a, F2, 24 ); W_PRECALC_32_79_0(40, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( a, b, c, d, e, F2, 25 ); W_PRECALC_32_79_1(41, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( e, a, b, c, d, F2, 26 ); W_PRECALC_32_79_2(42, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( d, e, a, b, c, F2, 27 ); W_PRECALC_32_79_3(43, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( c, d, e, a, b, F2, 28 ); W_PRECALC_32_79_0(44, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( b, c, d, e, a, F2, 29 ); W_PRECALC_32_79_1(45, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( a, b, c, d, e, F2, 30 ); W_PRECALC_32_79_2(46, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( e, a, b, c, d, F2, 31 ); W_PRECALC_32_79_3(47, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( d, e, a, b, c, F2, 32 ); W_PRECALC_32_79_0(48, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0);
R( c, d, e, a, b, F2, 33 ); W_PRECALC_32_79_1(49, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0);
R( b, c, d, e, a, F2, 34 ); W_PRECALC_32_79_2(50, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0);
R( a, b, c, d, e, F2, 35 ); W_PRECALC_32_79_3(51, W4, W5, W6, W7, W0, W1, W2, W3, Wtmp0);
R( e, a, b, c, d, F2, 36 ); W_PRECALC_32_79_0(52, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0);
R( d, e, a, b, c, F2, 37 ); W_PRECALC_32_79_1(53, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0);
R( c, d, e, a, b, F2, 38 ); W_PRECALC_32_79_2(54, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0);
R( b, c, d, e, a, F2, 39 ); W_PRECALC_32_79_3(55, W3, W4, W5, W6, W7, W0, W1, W2, Wtmp0);
R( a, b, c, d, e, F3, 40 ); W_PRECALC_32_79_0(56, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0);
R( e, a, b, c, d, F3, 41 ); W_PRECALC_32_79_1(57, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0);
R( d, e, a, b, c, F3, 42 ); W_PRECALC_32_79_2(58, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0);
R( c, d, e, a, b, F3, 43 ); W_PRECALC_32_79_3(59, W2, W3, W4, W5, W6, W7, W0, W1, Wtmp0);
R( b, c, d, e, a, F3, 44 ); W_PRECALC_32_79_0(60, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0);
R( a, b, c, d, e, F3, 45 ); W_PRECALC_32_79_1(61, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0);
R( e, a, b, c, d, F3, 46 ); W_PRECALC_32_79_2(62, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0);
R( d, e, a, b, c, F3, 47 ); W_PRECALC_32_79_3(63, W1, W2, W3, W4, W5, W6, W7, W0, Wtmp0);
R( c, d, e, a, b, F3, 48 ); W_PRECALC_32_79_0(64, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( b, c, d, e, a, F3, 49 ); W_PRECALC_32_79_1(65, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( a, b, c, d, e, F3, 50 ); W_PRECALC_32_79_2(66, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( e, a, b, c, d, F3, 51 ); W_PRECALC_32_79_3(67, W0, W1, W2, W3, W4, W5, W6, W7, Wtmp0);
R( d, e, a, b, c, F3, 52 ); W_PRECALC_32_79_0(68, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( c, d, e, a, b, F3, 53 ); W_PRECALC_32_79_1(69, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( b, c, d, e, a, F3, 54 ); W_PRECALC_32_79_2(70, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( a, b, c, d, e, F3, 55 ); W_PRECALC_32_79_3(71, W7, W0, W1, W2, W3, W4, W5, W6, Wtmp0);
R( e, a, b, c, d, F3, 56 ); W_PRECALC_32_79_0(72, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( d, e, a, b, c, F3, 57 ); W_PRECALC_32_79_1(73, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( c, d, e, a, b, F3, 58 ); W_PRECALC_32_79_2(74, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( b, c, d, e, a, F3, 59 ); W_PRECALC_32_79_3(75, W6, W7, W0, W1, W2, W3, W4, W5, Wtmp0);
R( a, b, c, d, e, F4, 60 ); W_PRECALC_32_79_0(76, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( e, a, b, c, d, F4, 61 ); W_PRECALC_32_79_1(77, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( d, e, a, b, c, F4, 62 ); W_PRECALC_32_79_2(78, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
R( c, d, e, a, b, F4, 63 ); W_PRECALC_32_79_3(79, W5, W6, W7, W0, W1, W2, W3, W4, Wtmp0);
decq RNBLKS;
jz .Lend;
/* Transform 64-79 + Precalc 0-15 of next block. */
R( b, c, d, e, a, F4, 64 ); W_PRECALC_00_15_0(0, W0, Wtmp0);
R( a, b, c, d, e, F4, 65 ); W_PRECALC_00_15_1(1, W0, Wtmp0);
R( e, a, b, c, d, F4, 66 ); W_PRECALC_00_15_2(2, W0, Wtmp0);
R( d, e, a, b, c, F4, 67 ); W_PRECALC_00_15_3(3, W0, Wtmp0);
R( c, d, e, a, b, F4, 68 ); W_PRECALC_00_15_0(4, W7, Wtmp0);
R( b, c, d, e, a, F4, 69 ); W_PRECALC_00_15_1(5, W7, Wtmp0);
R( a, b, c, d, e, F4, 70 ); W_PRECALC_00_15_2(6, W7, Wtmp0);
R( e, a, b, c, d, F4, 71 ); W_PRECALC_00_15_3(7, W7, Wtmp0);
R( d, e, a, b, c, F4, 72 ); W_PRECALC_00_15_0(8, W6, Wtmp0);
R( c, d, e, a, b, F4, 73 ); W_PRECALC_00_15_1(9, W6, Wtmp0);
R( b, c, d, e, a, F4, 74 ); W_PRECALC_00_15_2(10, W6, Wtmp0);
R( a, b, c, d, e, F4, 75 ); W_PRECALC_00_15_3(11, W6, Wtmp0);
R( e, a, b, c, d, F4, 76 ); W_PRECALC_00_15_0(12, W5, Wtmp0);
R( d, e, a, b, c, F4, 77 ); W_PRECALC_00_15_1(13, W5, Wtmp0);
R( c, d, e, a, b, F4, 78 );
addl state_h0(RSTATE), a; W_PRECALC_00_15_2(14, W5, Wtmp0);
R( b, c, d, e, a, F4, 79 ); W_PRECALC_00_15_3(15, W5, Wtmp0);
/* Update the chaining variables. */
addl state_h3(RSTATE), d;
addl state_h2(RSTATE), c;
addl state_h1(RSTATE), b;
addl state_h4(RSTATE), e;
movl d, state_h3(RSTATE);
movl c, state_h2(RSTATE);
movl b, state_h1(RSTATE);
movl a, state_h0(RSTATE);
movl e, state_h4(RSTATE);
jmp .Loop;
.align 16
.Lend:
/* Transform 64-79 + Clear XMM registers. */
R( b, c, d, e, a, F4, 64 ); CLEAR_REG(BSWAP_REG);
R( a, b, c, d, e, F4, 65 ); CLEAR_REG(Wtmp0);
R( e, a, b, c, d, F4, 66 ); CLEAR_REG(Wtmp1);
R( d, e, a, b, c, F4, 67 ); CLEAR_REG(W0);
R( c, d, e, a, b, F4, 68 ); CLEAR_REG(W1);
R( b, c, d, e, a, F4, 69 ); CLEAR_REG(W2);
R( a, b, c, d, e, F4, 70 ); CLEAR_REG(W3);
R( e, a, b, c, d, F4, 71 ); CLEAR_REG(W4);
R( d, e, a, b, c, F4, 72 ); CLEAR_REG(W5);
R( c, d, e, a, b, F4, 73 ); CLEAR_REG(W6);
R( b, c, d, e, a, F4, 74 ); CLEAR_REG(W7);
R( a, b, c, d, e, F4, 75 );
R( e, a, b, c, d, F4, 76 );
R( d, e, a, b, c, F4, 77 );
R( c, d, e, a, b, F4, 78 );
addl state_h0(RSTATE), a;
R( b, c, d, e, a, F4, 79 );
/* Update the chaining variables. */
addl state_h3(RSTATE), d;
addl state_h2(RSTATE), c;
addl state_h1(RSTATE), b;
addl state_h4(RSTATE), e;
movl d, state_h3(RSTATE);
movl c, state_h2(RSTATE);
movl b, state_h1(RSTATE);
movl a, state_h0(RSTATE);
movl e, state_h4(RSTATE);
movq ROLDSTACK, %rsp;
popq %rbp;
popq %rbx;
/* burn_stack */
movl $(16*4 + 2*8 + 31), %eax;
.Lret:
ret;
+ELF(.size _gcry_sha1_transform_amd64_ssse3,
+ .-_gcry_sha1_transform_amd64_ssse3;)
#endif
#endif
diff --git a/cipher/sha1.c b/cipher/sha1.c
index 76c486c7..affabfb0 100644
--- a/cipher/sha1.c
+++ b/cipher/sha1.c
@@ -1,664 +1,711 @@
/* sha1.c - SHA1 hash function
* Copyright (C) 1998, 2001, 2002, 2003, 2008 Free Software Foundation, Inc.
*
* This file is part of Libgcrypt.
*
* Libgcrypt is free software; you can redistribute it and/or modify
* it under the terms of the GNU Lesser General Public License as
* published by the Free Software Foundation; either version 2.1 of
* the License, or (at your option) any later version.
*
* Libgcrypt is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this program; if not, see .
*/
/* Test vectors:
*
* "abc"
* A999 3E36 4706 816A BA3E 2571 7850 C26C 9CD0 D89D
*
* "abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq"
* 8498 3E44 1C3B D26E BAAE 4AA1 F951 29E5 E546 70F1
*/
#include
#include
#include
#include
#ifdef HAVE_STDINT_H
# include
#endif
#include "g10lib.h"
#include "bithelp.h"
#include "bufhelp.h"
#include "cipher.h"
#include "sha1.h"
/* USE_SSSE3 indicates whether to compile with Intel SSSE3 code. */
#undef USE_SSSE3
#if defined(__x86_64__) && defined(HAVE_GCC_INLINE_ASM_SSSE3) && \
(defined(HAVE_COMPATIBLE_GCC_AMD64_PLATFORM_AS) || \
defined(HAVE_COMPATIBLE_GCC_WIN64_PLATFORM_AS))
# define USE_SSSE3 1
#endif
/* USE_AVX indicates whether to compile with Intel AVX code. */
#undef USE_AVX
#if defined(__x86_64__) && defined(HAVE_GCC_INLINE_ASM_AVX) && \
(defined(HAVE_COMPATIBLE_GCC_AMD64_PLATFORM_AS) || \
defined(HAVE_COMPATIBLE_GCC_WIN64_PLATFORM_AS))
# define USE_AVX 1
#endif
/* USE_BMI2 indicates whether to compile with Intel AVX/BMI2 code. */
#undef USE_BMI2
#if defined(__x86_64__) && defined(HAVE_GCC_INLINE_ASM_AVX) && \
defined(HAVE_GCC_INLINE_ASM_BMI2) && \
(defined(HAVE_COMPATIBLE_GCC_AMD64_PLATFORM_AS) || \
defined(HAVE_COMPATIBLE_GCC_WIN64_PLATFORM_AS))
# define USE_BMI2 1
#endif
+/* USE_AVX2 indicates whether to compile with Intel AVX2/BMI2 code. */
+#undef USE_AVX2
+#if defined(USE_BMI2) && defined(HAVE_GCC_INLINE_ASM_AVX2)
+# define USE_AVX2 1
+#endif
+
/* USE_SHAEXT indicates whether to compile with Intel SHA Extension code. */
#undef USE_SHAEXT
#if defined(HAVE_GCC_INLINE_ASM_SHAEXT) && \
defined(HAVE_GCC_INLINE_ASM_SSE41) && \
defined(ENABLE_SHAEXT_SUPPORT)
# define USE_SHAEXT 1
#endif
/* USE_NEON indicates whether to enable ARM NEON assembly code. */
#undef USE_NEON
#ifdef ENABLE_NEON_SUPPORT
# if defined(HAVE_ARM_ARCH_V6) && defined(__ARMEL__) \
&& defined(HAVE_COMPATIBLE_GCC_ARM_PLATFORM_AS) \
&& defined(HAVE_GCC_INLINE_ASM_NEON)
# define USE_NEON 1
# endif
#endif
/* USE_ARM_CE indicates whether to enable ARMv8 Crypto Extension assembly
* code. */
#undef USE_ARM_CE
#ifdef ENABLE_ARM_CRYPTO_SUPPORT
# if defined(HAVE_ARM_ARCH_V6) && defined(__ARMEL__) \
&& defined(HAVE_COMPATIBLE_GCC_ARM_PLATFORM_AS) \
&& defined(HAVE_GCC_INLINE_ASM_AARCH32_CRYPTO)
# define USE_ARM_CE 1
# elif defined(__AARCH64EL__) \
&& defined(HAVE_COMPATIBLE_GCC_AARCH64_PLATFORM_AS) \
&& defined(HAVE_GCC_INLINE_ASM_AARCH64_CRYPTO)
# define USE_ARM_CE 1
# endif
#endif
/* A macro to test whether P is properly aligned for an u32 type.
Note that config.h provides a suitable replacement for uintptr_t if
it does not exist in stdint.h. */
/* #if __GNUC__ >= 2 */
/* # define U32_ALIGNED_P(p) (!(((uintptr_t)p) % __alignof__ (u32))) */
/* #else */
/* # define U32_ALIGNED_P(p) (!(((uintptr_t)p) % sizeof (u32))) */
/* #endif */
/* Assembly implementations use SystemV ABI, ABI conversion and additional
* stack to store XMM6-XMM15 needed on Win64. */
#undef ASM_FUNC_ABI
#undef ASM_EXTRA_STACK
#if defined(USE_SSSE3) || defined(USE_AVX) || defined(USE_BMI2) || \
defined(USE_SHAEXT)
# ifdef HAVE_COMPATIBLE_GCC_WIN64_PLATFORM_AS
# define ASM_FUNC_ABI __attribute__((sysv_abi))
# define ASM_EXTRA_STACK (10 * 16 + sizeof(void *) * 4)
# else
# define ASM_FUNC_ABI
# define ASM_EXTRA_STACK 0
# endif
#endif
#ifdef USE_SSSE3
unsigned int
_gcry_sha1_transform_amd64_ssse3 (void *state, const unsigned char *data,
size_t nblks) ASM_FUNC_ABI;
static unsigned int
do_sha1_transform_amd64_ssse3 (void *ctx, const unsigned char *data,
size_t nblks)
{
SHA1_CONTEXT *hd = ctx;
return _gcry_sha1_transform_amd64_ssse3 (&hd->h0, data, nblks)
+ ASM_EXTRA_STACK;
}
#endif
#ifdef USE_AVX
unsigned int
_gcry_sha1_transform_amd64_avx (void *state, const unsigned char *data,
size_t nblks) ASM_FUNC_ABI;
static unsigned int
do_sha1_transform_amd64_avx (void *ctx, const unsigned char *data,
size_t nblks)
{
SHA1_CONTEXT *hd = ctx;
return _gcry_sha1_transform_amd64_avx (&hd->h0, data, nblks)
+ ASM_EXTRA_STACK;
}
#endif
#ifdef USE_BMI2
unsigned int
_gcry_sha1_transform_amd64_avx_bmi2 (void *state, const unsigned char *data,
size_t nblks) ASM_FUNC_ABI;
static unsigned int
do_sha1_transform_amd64_avx_bmi2 (void *ctx, const unsigned char *data,
size_t nblks)
{
SHA1_CONTEXT *hd = ctx;
return _gcry_sha1_transform_amd64_avx_bmi2 (&hd->h0, data, nblks)
+ ASM_EXTRA_STACK;
}
-#endif
+
+#ifdef USE_AVX2
+unsigned int
+_gcry_sha1_transform_amd64_avx2_bmi2 (void *state, const unsigned char *data,
+ size_t nblks) ASM_FUNC_ABI;
+
+static unsigned int
+do_sha1_transform_amd64_avx2_bmi2 (void *ctx, const unsigned char *data,
+ size_t nblks)
+{
+ SHA1_CONTEXT *hd = ctx;
+
+ /* AVX2/BMI2 function only handles pair of blocks so nblks needs to be
+ * multiple of 2 and function does not handle zero nblks. Use AVX/BMI2
+ * code to handle these cases. */
+
+ if (nblks <= 1)
+ return do_sha1_transform_amd64_avx_bmi2 (ctx, data, nblks);
+
+ if (nblks & 1)
+ {
+ (void)_gcry_sha1_transform_amd64_avx_bmi2 (&hd->h0, data, 1);
+ nblks--;
+ data += 64;
+ }
+
+ return _gcry_sha1_transform_amd64_avx2_bmi2 (&hd->h0, data, nblks)
+ + ASM_EXTRA_STACK;
+}
+#endif /* USE_AVX2 */
+#endif /* USE_BMI2 */
#ifdef USE_SHAEXT
/* Does not need ASM_FUNC_ABI */
unsigned int
_gcry_sha1_transform_intel_shaext (void *state, const unsigned char *data,
size_t nblks);
static unsigned int
do_sha1_transform_intel_shaext (void *ctx, const unsigned char *data,
size_t nblks)
{
SHA1_CONTEXT *hd = ctx;
return _gcry_sha1_transform_intel_shaext (&hd->h0, data, nblks);
}
#endif
#ifdef USE_NEON
unsigned int
_gcry_sha1_transform_armv7_neon (void *state, const unsigned char *data,
size_t nblks);
static unsigned int
do_sha1_transform_armv7_neon (void *ctx, const unsigned char *data,
size_t nblks)
{
SHA1_CONTEXT *hd = ctx;
return _gcry_sha1_transform_armv7_neon (&hd->h0, data, nblks);
}
#endif
#ifdef USE_ARM_CE
unsigned int
_gcry_sha1_transform_armv8_ce (void *state, const unsigned char *data,
size_t nblks);
static unsigned int
do_sha1_transform_armv8_ce (void *ctx, const unsigned char *data,
size_t nblks)
{
SHA1_CONTEXT *hd = ctx;
return _gcry_sha1_transform_armv8_ce (&hd->h0, data, nblks);
}
#endif
static unsigned int
do_transform_generic (void *c, const unsigned char *data, size_t nblks);
static void
sha1_init (void *context, unsigned int flags)
{
SHA1_CONTEXT *hd = context;
unsigned int features = _gcry_get_hw_features ();
(void)flags;
hd->h0 = 0x67452301;
hd->h1 = 0xefcdab89;
hd->h2 = 0x98badcfe;
hd->h3 = 0x10325476;
hd->h4 = 0xc3d2e1f0;
hd->bctx.nblocks = 0;
hd->bctx.nblocks_high = 0;
hd->bctx.count = 0;
hd->bctx.blocksize = 64;
/* Order of feature checks is important here; last match will be
* selected. Keep slower implementations at the top and faster at
* the bottom. */
hd->bctx.bwrite = do_transform_generic;
#ifdef USE_SSSE3
if ((features & HWF_INTEL_SSSE3) != 0)
hd->bctx.bwrite = do_sha1_transform_amd64_ssse3;
#endif
#ifdef USE_AVX
/* AVX implementation uses SHLD which is known to be slow on non-Intel CPUs.
* Therefore use this implementation on Intel CPUs only. */
if ((features & HWF_INTEL_AVX) && (features & HWF_INTEL_FAST_SHLD))
hd->bctx.bwrite = do_sha1_transform_amd64_avx;
#endif
#ifdef USE_BMI2
if ((features & HWF_INTEL_AVX) && (features & HWF_INTEL_BMI2))
hd->bctx.bwrite = do_sha1_transform_amd64_avx_bmi2;
#endif
+#ifdef USE_AVX2
+ if ((features & HWF_INTEL_AVX2) && (features & HWF_INTEL_AVX) &&
+ (features & HWF_INTEL_BMI2))
+ hd->bctx.bwrite = do_sha1_transform_amd64_avx2_bmi2;
+#endif
#ifdef USE_SHAEXT
if ((features & HWF_INTEL_SHAEXT) && (features & HWF_INTEL_SSE4_1))
hd->bctx.bwrite = do_sha1_transform_intel_shaext;
#endif
#ifdef USE_NEON
if ((features & HWF_ARM_NEON) != 0)
hd->bctx.bwrite = do_sha1_transform_armv7_neon;
#endif
#ifdef USE_ARM_CE
if ((features & HWF_ARM_SHA1) != 0)
hd->bctx.bwrite = do_sha1_transform_armv8_ce;
#endif
(void)features;
}
/*
* Initialize the context HD. This is used to prepare the use of
* _gcry_sha1_mixblock. WARNING: This is a special purpose function
* for exclusive use by random-csprng.c.
*/
void
_gcry_sha1_mixblock_init (SHA1_CONTEXT *hd)
{
sha1_init (hd, 0);
}
/* Round function macros. */
#define K1 0x5A827999L
#define K2 0x6ED9EBA1L
#define K3 0x8F1BBCDCL
#define K4 0xCA62C1D6L
#define F1(x,y,z) ( z ^ ( x & ( y ^ z ) ) )
#define F2(x,y,z) ( x ^ y ^ z )
#define F3(x,y,z) ( ( x & y ) | ( z & ( x | y ) ) )
#define F4(x,y,z) ( x ^ y ^ z )
#define M(i) ( tm = x[ i &0x0f] \
^ x[(i-14)&0x0f] \
^ x[(i-8) &0x0f] \
^ x[(i-3) &0x0f], \
(x[i&0x0f] = rol(tm, 1)))
#define R(a,b,c,d,e,f,k,m) do { e += rol( a, 5 ) \
+ f( b, c, d ) \
+ k \
+ m; \
b = rol( b, 30 ); \
} while(0)
/*
* Transform NBLOCKS of each 64 bytes (16 32-bit words) at DATA.
*/
static unsigned int
do_transform_generic (void *ctx, const unsigned char *data, size_t nblks)
{
SHA1_CONTEXT *hd = ctx;
do
{
const u32 *idata = (const void *)data;
u32 a, b, c, d, e; /* Local copies of the chaining variables. */
u32 tm; /* Helper. */
u32 x[16]; /* The array we work on. */
#define I(i) (x[i] = buf_get_be32(idata + i))
/* Get the values of the chaining variables. */
a = hd->h0;
b = hd->h1;
c = hd->h2;
d = hd->h3;
e = hd->h4;
/* Transform. */
R( a, b, c, d, e, F1, K1, I( 0) );
R( e, a, b, c, d, F1, K1, I( 1) );
R( d, e, a, b, c, F1, K1, I( 2) );
R( c, d, e, a, b, F1, K1, I( 3) );
R( b, c, d, e, a, F1, K1, I( 4) );
R( a, b, c, d, e, F1, K1, I( 5) );
R( e, a, b, c, d, F1, K1, I( 6) );
R( d, e, a, b, c, F1, K1, I( 7) );
R( c, d, e, a, b, F1, K1, I( 8) );
R( b, c, d, e, a, F1, K1, I( 9) );
R( a, b, c, d, e, F1, K1, I(10) );
R( e, a, b, c, d, F1, K1, I(11) );
R( d, e, a, b, c, F1, K1, I(12) );
R( c, d, e, a, b, F1, K1, I(13) );
R( b, c, d, e, a, F1, K1, I(14) );
R( a, b, c, d, e, F1, K1, I(15) );
R( e, a, b, c, d, F1, K1, M(16) );
R( d, e, a, b, c, F1, K1, M(17) );
R( c, d, e, a, b, F1, K1, M(18) );
R( b, c, d, e, a, F1, K1, M(19) );
R( a, b, c, d, e, F2, K2, M(20) );
R( e, a, b, c, d, F2, K2, M(21) );
R( d, e, a, b, c, F2, K2, M(22) );
R( c, d, e, a, b, F2, K2, M(23) );
R( b, c, d, e, a, F2, K2, M(24) );
R( a, b, c, d, e, F2, K2, M(25) );
R( e, a, b, c, d, F2, K2, M(26) );
R( d, e, a, b, c, F2, K2, M(27) );
R( c, d, e, a, b, F2, K2, M(28) );
R( b, c, d, e, a, F2, K2, M(29) );
R( a, b, c, d, e, F2, K2, M(30) );
R( e, a, b, c, d, F2, K2, M(31) );
R( d, e, a, b, c, F2, K2, M(32) );
R( c, d, e, a, b, F2, K2, M(33) );
R( b, c, d, e, a, F2, K2, M(34) );
R( a, b, c, d, e, F2, K2, M(35) );
R( e, a, b, c, d, F2, K2, M(36) );
R( d, e, a, b, c, F2, K2, M(37) );
R( c, d, e, a, b, F2, K2, M(38) );
R( b, c, d, e, a, F2, K2, M(39) );
R( a, b, c, d, e, F3, K3, M(40) );
R( e, a, b, c, d, F3, K3, M(41) );
R( d, e, a, b, c, F3, K3, M(42) );
R( c, d, e, a, b, F3, K3, M(43) );
R( b, c, d, e, a, F3, K3, M(44) );
R( a, b, c, d, e, F3, K3, M(45) );
R( e, a, b, c, d, F3, K3, M(46) );
R( d, e, a, b, c, F3, K3, M(47) );
R( c, d, e, a, b, F3, K3, M(48) );
R( b, c, d, e, a, F3, K3, M(49) );
R( a, b, c, d, e, F3, K3, M(50) );
R( e, a, b, c, d, F3, K3, M(51) );
R( d, e, a, b, c, F3, K3, M(52) );
R( c, d, e, a, b, F3, K3, M(53) );
R( b, c, d, e, a, F3, K3, M(54) );
R( a, b, c, d, e, F3, K3, M(55) );
R( e, a, b, c, d, F3, K3, M(56) );
R( d, e, a, b, c, F3, K3, M(57) );
R( c, d, e, a, b, F3, K3, M(58) );
R( b, c, d, e, a, F3, K3, M(59) );
R( a, b, c, d, e, F4, K4, M(60) );
R( e, a, b, c, d, F4, K4, M(61) );
R( d, e, a, b, c, F4, K4, M(62) );
R( c, d, e, a, b, F4, K4, M(63) );
R( b, c, d, e, a, F4, K4, M(64) );
R( a, b, c, d, e, F4, K4, M(65) );
R( e, a, b, c, d, F4, K4, M(66) );
R( d, e, a, b, c, F4, K4, M(67) );
R( c, d, e, a, b, F4, K4, M(68) );
R( b, c, d, e, a, F4, K4, M(69) );
R( a, b, c, d, e, F4, K4, M(70) );
R( e, a, b, c, d, F4, K4, M(71) );
R( d, e, a, b, c, F4, K4, M(72) );
R( c, d, e, a, b, F4, K4, M(73) );
R( b, c, d, e, a, F4, K4, M(74) );
R( a, b, c, d, e, F4, K4, M(75) );
R( e, a, b, c, d, F4, K4, M(76) );
R( d, e, a, b, c, F4, K4, M(77) );
R( c, d, e, a, b, F4, K4, M(78) );
R( b, c, d, e, a, F4, K4, M(79) );
/* Update the chaining variables. */
hd->h0 += a;
hd->h1 += b;
hd->h2 += c;
hd->h3 += d;
hd->h4 += e;
data += 64;
}
while (--nblks);
return 88+4*sizeof(void*);
}
/*
* Apply the SHA-1 transform function on the buffer BLOCKOF64BYTE
* which must have a length 64 bytes. BLOCKOF64BYTE must be 32-bit
* aligned. Updates the 20 bytes in BLOCKOF64BYTE with its mixed
* content. Returns the number of bytes which should be burned on the
* stack. You need to use _gcry_sha1_mixblock_init to initialize the
* context.
* WARNING: This is a special purpose function for exclusive use by
* random-csprng.c.
*/
unsigned int
_gcry_sha1_mixblock (SHA1_CONTEXT *hd, void *blockof64byte)
{
u32 *p = blockof64byte;
unsigned int nburn;
nburn = (*hd->bctx.bwrite) (hd, blockof64byte, 1);
p[0] = hd->h0;
p[1] = hd->h1;
p[2] = hd->h2;
p[3] = hd->h3;
p[4] = hd->h4;
return nburn;
}
/* The routine final terminates the computation and
* returns the digest.
* The handle is prepared for a new cycle, but adding bytes to the
* handle will the destroy the returned buffer.
* Returns: 20 bytes representing the digest.
*/
static void
sha1_final(void *context)
{
SHA1_CONTEXT *hd = context;
u32 t, th, msb, lsb;
unsigned char *p;
unsigned int burn;
_gcry_md_block_write (hd, NULL, 0); /* flush */;
t = hd->bctx.nblocks;
if (sizeof t == sizeof hd->bctx.nblocks)
th = hd->bctx.nblocks_high;
else
th = hd->bctx.nblocks >> 32;
/* multiply by 64 to make a byte count */
lsb = t << 6;
msb = (th << 6) | (t >> 26);
/* add the count */
t = lsb;
if( (lsb += hd->bctx.count) < t )
msb++;
/* multiply by 8 to make a bit count */
t = lsb;
lsb <<= 3;
msb <<= 3;
msb |= t >> 29;
if( hd->bctx.count < 56 ) /* enough room */
{
hd->bctx.buf[hd->bctx.count++] = 0x80; /* pad */
- while( hd->bctx.count < 56 )
- hd->bctx.buf[hd->bctx.count++] = 0; /* pad */
+ if (hd->bctx.count < 56)
+ memset (&hd->bctx.buf[hd->bctx.count], 0, 56 - hd->bctx.count);
+ hd->bctx.count = 56;
+
+ /* append the 64 bit count */
+ buf_put_be32(hd->bctx.buf + 56, msb);
+ buf_put_be32(hd->bctx.buf + 60, lsb);
+ burn = (*hd->bctx.bwrite) ( hd, hd->bctx.buf, 1 );
}
else /* need one extra block */
{
hd->bctx.buf[hd->bctx.count++] = 0x80; /* pad character */
- while( hd->bctx.count < 64 )
- hd->bctx.buf[hd->bctx.count++] = 0;
- _gcry_md_block_write(hd, NULL, 0); /* flush */;
- memset(hd->bctx.buf, 0, 56 ); /* fill next block with zeroes */
+ /* fill pad and next block with zeroes */
+ memset (&hd->bctx.buf[hd->bctx.count], 0, 64 - hd->bctx.count + 56);
+ hd->bctx.count = 64 + 56;
+
+ /* append the 64 bit count */
+ buf_put_be32(hd->bctx.buf + 64 + 56, msb);
+ buf_put_be32(hd->bctx.buf + 64 + 60, lsb);
+ burn = (*hd->bctx.bwrite) ( hd, hd->bctx.buf, 2 );
}
- /* append the 64 bit count */
- buf_put_be32(hd->bctx.buf + 56, msb);
- buf_put_be32(hd->bctx.buf + 60, lsb);
- burn = (*hd->bctx.bwrite) ( hd, hd->bctx.buf, 1 );
- _gcry_burn_stack (burn);
p = hd->bctx.buf;
#define X(a) do { buf_put_be32(p, hd->h##a); p += 4; } while(0)
X(0);
X(1);
X(2);
X(3);
X(4);
#undef X
+ _gcry_burn_stack (burn);
}
static unsigned char *
sha1_read( void *context )
{
SHA1_CONTEXT *hd = context;
return hd->bctx.buf;
}
/****************
* Shortcut functions which puts the hash value of the supplied buffer
* into outbuf which must have a size of 20 bytes.
*/
void
_gcry_sha1_hash_buffer (void *outbuf, const void *buffer, size_t length)
{
SHA1_CONTEXT hd;
sha1_init (&hd, 0);
_gcry_md_block_write (&hd, buffer, length);
sha1_final (&hd);
memcpy (outbuf, hd.bctx.buf, 20);
}
/* Variant of the above shortcut function using a multiple buffers. */
void
_gcry_sha1_hash_buffers (void *outbuf, const gcry_buffer_t *iov, int iovcnt)
{
SHA1_CONTEXT hd;
sha1_init (&hd, 0);
for (;iovcnt > 0; iov++, iovcnt--)
_gcry_md_block_write (&hd,
(const char*)iov[0].data + iov[0].off, iov[0].len);
sha1_final (&hd);
memcpy (outbuf, hd.bctx.buf, 20);
}
/*
Self-test section.
*/
static gpg_err_code_t
selftests_sha1 (int extended, selftest_report_func_t report)
{
const char *what;
const char *errtxt;
what = "short string";
errtxt = _gcry_hash_selftest_check_one
(GCRY_MD_SHA1, 0,
"abc", 3,
"\xA9\x99\x3E\x36\x47\x06\x81\x6A\xBA\x3E"
"\x25\x71\x78\x50\xC2\x6C\x9C\xD0\xD8\x9D", 20);
if (errtxt)
goto failed;
if (extended)
{
what = "long string";
errtxt = _gcry_hash_selftest_check_one
(GCRY_MD_SHA1, 0,
"abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq", 56,
"\x84\x98\x3E\x44\x1C\x3B\xD2\x6E\xBA\xAE"
"\x4A\xA1\xF9\x51\x29\xE5\xE5\x46\x70\xF1", 20);
if (errtxt)
goto failed;
what = "one million \"a\"";
errtxt = _gcry_hash_selftest_check_one
(GCRY_MD_SHA1, 1,
NULL, 0,
"\x34\xAA\x97\x3C\xD4\xC4\xDA\xA4\xF6\x1E"
"\xEB\x2B\xDB\xAD\x27\x31\x65\x34\x01\x6F", 20);
if (errtxt)
goto failed;
}
return 0; /* Succeeded. */
failed:
if (report)
report ("digest", GCRY_MD_SHA1, what, errtxt);
return GPG_ERR_SELFTEST_FAILED;
}
/* Run a full self-test for ALGO and return 0 on success. */
static gpg_err_code_t
run_selftests (int algo, int extended, selftest_report_func_t report)
{
gpg_err_code_t ec;
switch (algo)
{
case GCRY_MD_SHA1:
ec = selftests_sha1 (extended, report);
break;
default:
ec = GPG_ERR_DIGEST_ALGO;
break;
}
return ec;
}
static unsigned char asn[15] = /* Object ID is 1.3.14.3.2.26 */
{ 0x30, 0x21, 0x30, 0x09, 0x06, 0x05, 0x2b, 0x0e, 0x03,
0x02, 0x1a, 0x05, 0x00, 0x04, 0x14 };
static gcry_md_oid_spec_t oid_spec_sha1[] =
{
/* iso.member-body.us.rsadsi.pkcs.pkcs-1.5 (sha1WithRSAEncryption) */
{ "1.2.840.113549.1.1.5" },
/* iso.member-body.us.x9-57.x9cm.3 (dsaWithSha1)*/
{ "1.2.840.10040.4.3" },
/* from NIST's OIW (sha1) */
{ "1.3.14.3.2.26" },
/* from NIST OIW (sha-1WithRSAEncryption) */
{ "1.3.14.3.2.29" },
/* iso.member-body.us.ansi-x9-62.signatures.ecdsa-with-sha1 */
{ "1.2.840.10045.4.1" },
{ NULL },
};
gcry_md_spec_t _gcry_digest_spec_sha1 =
{
GCRY_MD_SHA1, {0, 1},
"SHA1", asn, DIM (asn), oid_spec_sha1, 20,
sha1_init, _gcry_md_block_write, sha1_final, sha1_read, NULL,
_gcry_sha1_hash_buffer, _gcry_sha1_hash_buffers,
sizeof (SHA1_CONTEXT),
run_selftests
};
diff --git a/configure.ac b/configure.ac
index bb3c666f..0a931f95 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1,2785 +1,2786 @@
# Configure.ac script for Libgcrypt
# Copyright (C) 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2006,
# 2007, 2008, 2009, 2011 Free Software Foundation, Inc.
# Copyright (C) 2012-2017 g10 Code GmbH
#
# This file is part of Libgcrypt.
#
# Libgcrypt is free software; you can redistribute it and/or modify
# it under the terms of the GNU Lesser General Public License as
# published by the Free Software Foundation; either version 2.1 of
# the License, or (at your option) any later version.
#
# Libgcrypt is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU Lesser General Public License for more details.
#
# You should have received a copy of the GNU Lesser General Public
# License along with this program; if not, see .
# (Process this file with autoconf to produce a configure script.)
AC_REVISION($Revision$)
AC_PREREQ(2.60)
min_automake_version="1.14"
# To build a release you need to create a tag with the version number
# (git tag -s libgcrypt-n.m.k) and run "./autogen.sh --force". Please
# bump the version number immediately after the release and do another
# commit and push so that the git magic is able to work. See below
# for the LT versions.
m4_define([mym4_package],[libgcrypt])
m4_define([mym4_major], [1])
m4_define([mym4_minor], [9])
m4_define([mym4_micro], [0])
# Below is m4 magic to extract and compute the git revision number,
# the decimalized short revision number, a beta version string and a
# flag indicating a development version (mym4_isbeta). Note that the
# m4 processing is done by autoconf and not during the configure run.
m4_define([mym4_verslist], m4_split(m4_esyscmd([./autogen.sh --find-version] \
mym4_package mym4_major mym4_minor mym4_micro),[:]))
m4_define([mym4_isbeta], m4_argn(2, mym4_verslist))
m4_define([mym4_version], m4_argn(4, mym4_verslist))
m4_define([mym4_revision], m4_argn(7, mym4_verslist))
m4_define([mym4_revision_dec], m4_argn(8, mym4_verslist))
m4_esyscmd([echo ]mym4_version[>VERSION])
AC_INIT([mym4_package],[mym4_version], [https://bugs.gnupg.org])
# LT Version numbers, remember to change them just *before* a release.
# (Code changed: REVISION++)
# (Interfaces added/removed/changed: CURRENT++, REVISION=0)
# (Interfaces added: AGE++)
# (Interfaces removed: AGE=0)
#
# (Interfaces removed: CURRENT++, AGE=0, REVISION=0)
# (Interfaces added: CURRENT++, AGE++, REVISION=0)
# (No interfaces changed: REVISION++)
LIBGCRYPT_LT_CURRENT=23
LIBGCRYPT_LT_AGE=3
LIBGCRYPT_LT_REVISION=0
################################################
AC_SUBST(LIBGCRYPT_LT_CURRENT)
AC_SUBST(LIBGCRYPT_LT_AGE)
AC_SUBST(LIBGCRYPT_LT_REVISION)
# If the API is changed in an incompatible way: increment the next counter.
#
# 1.6: ABI and API change but the change is to most users irrelevant
# and thus the API version number has not been incremented.
LIBGCRYPT_CONFIG_API_VERSION=1
# If you change the required gpg-error version, please remove
# unnecessary error code defines in src/gcrypt-int.h.
NEED_GPG_ERROR_VERSION=1.25
AC_CONFIG_AUX_DIR([build-aux])
AC_CONFIG_SRCDIR([src/libgcrypt.vers])
AM_INIT_AUTOMAKE([serial-tests dist-bzip2])
AC_CONFIG_HEADER(config.h)
AC_CONFIG_MACRO_DIR([m4])
AC_CONFIG_LIBOBJ_DIR([compat])
AC_CANONICAL_HOST
AM_MAINTAINER_MODE
AM_SILENT_RULES
AC_ARG_VAR(SYSROOT,[locate config scripts also below that directory])
AH_TOP([
#ifndef _GCRYPT_CONFIG_H_INCLUDED
#define _GCRYPT_CONFIG_H_INCLUDED
/* Enable gpg-error's strerror macro for W32CE. */
#define GPG_ERR_ENABLE_ERRNO_MACROS 1
])
AH_BOTTOM([
#define _GCRYPT_IN_LIBGCRYPT 1
/* If the configure check for endianness has been disabled, get it from
OS macros. This is intended for making fat binary builds on OS X. */
#ifdef DISABLED_ENDIAN_CHECK
# if defined(__BIG_ENDIAN__)
# define WORDS_BIGENDIAN 1
# elif defined(__LITTLE_ENDIAN__)
# undef WORDS_BIGENDIAN
# else
# error "No endianness found"
# endif
#endif /*DISABLED_ENDIAN_CHECK*/
/* We basically use the original Camellia source. Make sure the symbols
properly prefixed. */
#define CAMELLIA_EXT_SYM_PREFIX _gcry_
#endif /*_GCRYPT_CONFIG_H_INCLUDED*/
])
AH_VERBATIM([_REENTRANT],
[/* To allow the use of Libgcrypt in multithreaded programs we have to use
special features from the library. */
#ifndef _REENTRANT
# define _REENTRANT 1
#endif
])
######################
## Basic checks. ### (we need some results later on (e.g. $GCC)
######################
AC_PROG_MAKE_SET
missing_dir=`cd $ac_aux_dir && pwd`
AM_MISSING_PROG(ACLOCAL, aclocal, $missing_dir)
AM_MISSING_PROG(AUTOCONF, autoconf, $missing_dir)
AM_MISSING_PROG(AUTOMAKE, automake, $missing_dir)
AM_MISSING_PROG(AUTOHEADER, autoheader, $missing_dir)
# AM_MISSING_PROG(MAKEINFO, makeinfo, $missing_dir)
AC_PROG_CC
AC_PROG_CPP
AM_PROG_CC_C_O
AM_PROG_AS
AC_ISC_POSIX
AC_PROG_INSTALL
AC_PROG_AWK
AC_GNU_SOURCE
# Taken from mpfr-4.0.1, then modified for LDADD_FOR_TESTS_KLUDGE
dnl Under Linux, make sure that the old dtags are used if LD_LIBRARY_PATH
dnl is defined. The issue is that with the new dtags, LD_LIBRARY_PATH has
dnl the precedence over the run path, so that if a compatible MPFR library
dnl is installed in some directory from $LD_LIBRARY_PATH, then the tested
dnl MPFR library will be this library instead of the MPFR library from the
dnl build tree. Other OS with the same issue might be added later.
dnl
dnl References:
dnl https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=859732
dnl http://lists.gnu.org/archive/html/libtool/2017-05/msg00000.html
dnl
dnl We need to check whether --disable-new-dtags is supported as alternate
dnl linkers may be used (e.g., with tcc: CC=tcc LD=tcc).
dnl
case $host in
*-*-linux*)
if test -n "$LD_LIBRARY_PATH"; then
saved_LDFLAGS="$LDFLAGS"
LDADD_FOR_TESTS_KLUDGE="-Wl,--disable-new-dtags"
LDFLAGS="$LDFLAGS $LDADD_FOR_TESTS_KLUDGE"
AC_MSG_CHECKING(whether --disable-new-dtags is supported by the linker)
AC_LINK_IFELSE([AC_LANG_SOURCE([[
int main (void) { return 0; }
]])],
[AC_MSG_RESULT(yes (use it since LD_LIBRARY_PATH is set))],
[AC_MSG_RESULT(no)
LDADD_FOR_TESTS_KLUDGE=""
])
LDFLAGS="$saved_LDFLAGS"
fi
;;
esac
AC_SUBST([LDADD_FOR_TESTS_KLUDGE])
VERSION_NUMBER=m4_esyscmd(printf "0x%02x%02x%02x" mym4_major \
mym4_minor mym4_micro)
AC_SUBST(VERSION_NUMBER)
# We need to compile and run a program on the build machine. A
# comment in libgpg-error says that the AC_PROG_CC_FOR_BUILD macro in
# the AC archive is broken for autoconf 2.57. Given that there is no
# newer version of that macro, we assume that it is also broken for
# autoconf 2.61 and thus we use a simple but usually sufficient
# approach.
AC_MSG_CHECKING(for cc for build)
if test "$cross_compiling" = "yes"; then
CC_FOR_BUILD="${CC_FOR_BUILD-cc}"
else
CC_FOR_BUILD="${CC_FOR_BUILD-$CC}"
fi
AC_MSG_RESULT($CC_FOR_BUILD)
AC_ARG_VAR(CC_FOR_BUILD,[build system C compiler])
LT_PREREQ([2.2.6])
LT_INIT([win32-dll disable-static])
LT_LANG([Windows Resource])
##########################
## General definitions. ##
##########################
# Used by libgcrypt-config
LIBGCRYPT_CONFIG_LIBS="-lgcrypt"
LIBGCRYPT_CONFIG_CFLAGS=""
LIBGCRYPT_CONFIG_HOST="$host"
# Definitions for symmetric ciphers.
available_ciphers="arcfour blowfish cast5 des aes twofish serpent rfc2268 seed"
available_ciphers="$available_ciphers camellia idea salsa20 gost28147 chacha20"
enabled_ciphers=""
# Definitions for public-key ciphers.
available_pubkey_ciphers="dsa elgamal rsa ecc"
enabled_pubkey_ciphers=""
# Definitions for message digests.
available_digests="crc gostr3411-94 md2 md4 md5 rmd160 sha1 sha256 sha512"
available_digests="$available_digests sha3 tiger whirlpool stribog blake2"
available_digests="$available_digests sm3"
enabled_digests=""
# Definitions for kdfs (optional ones)
available_kdfs="s2k pkdf2 scrypt"
enabled_kdfs=""
# Definitions for random modules.
available_random_modules="linux egd unix"
auto_random_modules="$available_random_modules"
# Supported thread backends.
LIBGCRYPT_THREAD_MODULES=""
# Other definitions.
have_w32_system=no
have_w32ce_system=no
have_pthread=no
# Setup some stuff depending on host.
case "${host}" in
*-*-mingw32*)
ac_cv_have_dev_random=no
have_w32_system=yes
case "${host}" in
*-mingw32ce*)
have_w32ce_system=yes
available_random_modules="w32ce"
;;
*)
available_random_modules="w32"
;;
esac
AC_DEFINE(USE_ONLY_8DOT3,1,
[set this to limit filenames to the 8.3 format])
AC_DEFINE(HAVE_DRIVE_LETTERS,1,
[defined if we must run on a stupid file system])
AC_DEFINE(HAVE_DOSISH_SYSTEM,1,
[defined if we run on some of the PCDOS like systems
(DOS, Windoze. OS/2) with special properties like
no file modes])
;;
i?86-emx-os2 | i?86-*-os2*emx)
# OS/2 with the EMX environment
ac_cv_have_dev_random=no
AC_DEFINE(HAVE_DRIVE_LETTERS)
AC_DEFINE(HAVE_DOSISH_SYSTEM)
;;
i?86-*-msdosdjgpp*)
# DOS with the DJGPP environment
ac_cv_have_dev_random=no
AC_DEFINE(HAVE_DRIVE_LETTERS)
AC_DEFINE(HAVE_DOSISH_SYSTEM)
;;
*-*-hpux*)
if test -z "$GCC" ; then
CFLAGS="$CFLAGS -Ae -D_HPUX_SOURCE"
fi
;;
*-dec-osf4*)
if test -z "$GCC" ; then
# Suppress all warnings
# to get rid of the unsigned/signed char mismatch warnings.
CFLAGS="$CFLAGS -w"
fi
;;
m68k-atari-mint)
;;
*-apple-darwin*)
AC_DEFINE(_DARWIN_C_SOURCE, 900000L,
Expose all libc features (__DARWIN_C_FULL).)
;;
*)
;;
esac
if test "$have_w32_system" = yes; then
AC_DEFINE(HAVE_W32_SYSTEM,1, [Defined if we run on a W32 API based system])
if test "$have_w32ce_system" = yes; then
AC_DEFINE(HAVE_W32CE_SYSTEM,1,[Defined if we run on WindowsCE])
fi
fi
AM_CONDITIONAL(HAVE_W32_SYSTEM, test "$have_w32_system" = yes)
AM_CONDITIONAL(HAVE_W32CE_SYSTEM, test "$have_w32ce_system" = yes)
# A printable OS Name is sometimes useful.
case "${host}" in
*-*-mingw32ce*)
PRINTABLE_OS_NAME="W32CE"
;;
*-*-mingw32*)
PRINTABLE_OS_NAME="W32"
;;
i?86-emx-os2 | i?86-*-os2*emx )
PRINTABLE_OS_NAME="OS/2"
;;
i?86-*-msdosdjgpp*)
PRINTABLE_OS_NAME="MSDOS/DJGPP"
;;
*-linux*)
PRINTABLE_OS_NAME="GNU/Linux"
;;
*)
PRINTABLE_OS_NAME=`uname -s || echo "Unknown"`
;;
esac
NAME_OF_DEV_RANDOM="/dev/random"
NAME_OF_DEV_URANDOM="/dev/urandom"
AC_ARG_ENABLE(endian-check,
AC_HELP_STRING([--disable-endian-check],
[disable the endian check and trust the OS provided macros]),
endiancheck=$enableval,endiancheck=yes)
if test x"$endiancheck" = xyes ; then
AC_C_BIGENDIAN
else
AC_DEFINE(DISABLED_ENDIAN_CHECK,1,[configure did not test for endianness])
fi
AC_CHECK_SIZEOF(unsigned short, 2)
AC_CHECK_SIZEOF(unsigned int, 4)
AC_CHECK_SIZEOF(unsigned long, 4)
AC_CHECK_SIZEOF(unsigned long long, 0)
AC_CHECK_SIZEOF(void *, 0)
AC_TYPE_UINTPTR_T
if test "$ac_cv_sizeof_unsigned_short" = "0" \
|| test "$ac_cv_sizeof_unsigned_int" = "0" \
|| test "$ac_cv_sizeof_unsigned_long" = "0"; then
AC_MSG_WARN([Hmmm, something is wrong with the sizes - using defaults]);
fi
# Ensure that we have UINT64_C before we bother to check for uint64_t
AC_CACHE_CHECK([for UINT64_C],[gnupg_cv_uint64_c_works],
AC_COMPILE_IFELSE([AC_LANG_PROGRAM([[#include ]],
[[uint64_t foo=UINT64_C(42);]])],
gnupg_cv_uint64_c_works=yes,gnupg_cv_uint64_c_works=no))
if test "$gnupg_cv_uint64_c_works" = "yes" ; then
AC_CHECK_SIZEOF(uint64_t)
fi
# Do we have any 64-bit data types?
if test "$ac_cv_sizeof_unsigned_int" != "8" \
&& test "$ac_cv_sizeof_unsigned_long" != "8" \
&& test "$ac_cv_sizeof_unsigned_long_long" != "8" \
&& test "$ac_cv_sizeof_uint64_t" != "8"; then
AC_MSG_ERROR([[
***
*** No 64-bit integer type available.
*** It is not possible to build Libgcrypt on this platform.
***]])
fi
# If not specified otherwise, all available algorithms will be
# included.
default_ciphers="$available_ciphers"
default_pubkey_ciphers="$available_pubkey_ciphers"
default_digests="$available_digests"
default_kdfs="$available_kdfs"
# Blacklist MD2 by default
default_digests=`echo $default_digests | sed -e 's/md2//g'`
# Substitutions to set generated files in a Emacs buffer to read-only.
AC_SUBST(emacs_local_vars_begin, ['Local Variables:'])
AC_SUBST(emacs_local_vars_read_only, ['buffer-read-only: t'])
AC_SUBST(emacs_local_vars_end, ['End:'])
############################
## Command line switches. ##
############################
# Implementation of the --enable-ciphers switch.
AC_ARG_ENABLE(ciphers,
AC_HELP_STRING([--enable-ciphers=ciphers],
[select the symmetric ciphers to include]),
[enabled_ciphers=`echo $enableval | tr ',:' ' ' | tr '[A-Z]' '[a-z]'`],
[enabled_ciphers=""])
if test "x$enabled_ciphers" = "x" \
-o "$enabled_ciphers" = "yes" \
-o "$enabled_ciphers" = "no"; then
enabled_ciphers=$default_ciphers
fi
AC_MSG_CHECKING([which symmetric ciphers to include])
for cipher in $enabled_ciphers; do
LIST_MEMBER($cipher, $available_ciphers)
if test "$found" = "0"; then
AC_MSG_ERROR([unsupported cipher "$cipher" specified])
fi
done
AC_MSG_RESULT([$enabled_ciphers])
# Implementation of the --enable-pubkey-ciphers switch.
AC_ARG_ENABLE(pubkey-ciphers,
AC_HELP_STRING([--enable-pubkey-ciphers=ciphers],
[select the public-key ciphers to include]),
[enabled_pubkey_ciphers=`echo $enableval | tr ',:' ' ' | tr '[A-Z]' '[a-z]'`],
[enabled_pubkey_ciphers=""])
if test "x$enabled_pubkey_ciphers" = "x" \
-o "$enabled_pubkey_ciphers" = "yes" \
-o "$enabled_pubkey_ciphers" = "no"; then
enabled_pubkey_ciphers=$default_pubkey_ciphers
fi
AC_MSG_CHECKING([which public-key ciphers to include])
for cipher in $enabled_pubkey_ciphers; do
LIST_MEMBER($cipher, $available_pubkey_ciphers)
if test "$found" = "0"; then
AC_MSG_ERROR([unsupported public-key cipher specified])
fi
done
AC_MSG_RESULT([$enabled_pubkey_ciphers])
# Implementation of the --enable-digests switch.
AC_ARG_ENABLE(digests,
AC_HELP_STRING([--enable-digests=digests],
[select the message digests to include]),
[enabled_digests=`echo $enableval | tr ',:' ' ' | tr '[A-Z]' '[a-z]'`],
[enabled_digests=""])
if test "x$enabled_digests" = "x" \
-o "$enabled_digests" = "yes" \
-o "$enabled_digests" = "no"; then
enabled_digests=$default_digests
fi
AC_MSG_CHECKING([which message digests to include])
for digest in $enabled_digests; do
LIST_MEMBER($digest, $available_digests)
if test "$found" = "0"; then
AC_MSG_ERROR([unsupported message digest specified])
fi
done
AC_MSG_RESULT([$enabled_digests])
# Implementation of the --enable-kdfs switch.
AC_ARG_ENABLE(kdfs,
AC_HELP_STRING([--enable-kfds=kdfs],
[select the KDFs to include]),
[enabled_kdfs=`echo $enableval | tr ',:' ' ' | tr '[A-Z]' '[a-z]'`],
[enabled_kdfs=""])
if test "x$enabled_kdfs" = "x" \
-o "$enabled_kdfs" = "yes" \
-o "$enabled_kdfs" = "no"; then
enabled_kdfs=$default_kdfs
fi
AC_MSG_CHECKING([which key derivation functions to include])
for kdf in $enabled_kdfs; do
LIST_MEMBER($kdf, $available_kdfs)
if test "$found" = "0"; then
AC_MSG_ERROR([unsupported key derivation function specified])
fi
done
AC_MSG_RESULT([$enabled_kdfs])
# Implementation of the --enable-random switch.
AC_ARG_ENABLE(random,
AC_HELP_STRING([--enable-random=name],
[select which random number generator to use]),
[random=`echo $enableval | tr '[A-Z]' '[a-z]'`],
[])
if test "x$random" = "x" -o "$random" = "yes" -o "$random" = "no"; then
random=default
fi
AC_MSG_CHECKING([which random module to use])
if test "$random" != "default" -a "$random" != "auto"; then
LIST_MEMBER($random, $available_random_modules)
if test "$found" = "0"; then
AC_MSG_ERROR([unsupported random module specified])
fi
fi
AC_MSG_RESULT($random)
# Implementation of the --disable-dev-random switch.
AC_MSG_CHECKING([whether use of /dev/random is requested])
AC_ARG_ENABLE(dev-random,
[ --disable-dev-random disable the use of dev random],
try_dev_random=$enableval, try_dev_random=yes)
AC_MSG_RESULT($try_dev_random)
# Implementation of the --with-egd-socket switch.
AC_ARG_WITH(egd-socket,
[ --with-egd-socket=NAME Use NAME for the EGD socket)],
egd_socket_name="$withval", egd_socket_name="" )
AC_DEFINE_UNQUOTED(EGD_SOCKET_NAME, "$egd_socket_name",
[Define if you don't want the default EGD socket name.
For details see cipher/rndegd.c])
# Implementation of the --enable-random-daemon
AC_MSG_CHECKING([whether the experimental random daemon is requested])
AC_ARG_ENABLE([random-daemon],
AC_HELP_STRING([--enable-random-daemon],
[Build and support the experimental gcryptrnd]),
[use_random_daemon=$enableval],
[use_random_daemon=no])
AC_MSG_RESULT($use_random_daemon)
if test x$use_random_daemon = xyes ; then
AC_DEFINE(USE_RANDOM_DAEMON,1,
[Define to support the experimental random daemon])
fi
AM_CONDITIONAL(USE_RANDOM_DAEMON, test x$use_random_daemon = xyes)
# Implementation of --disable-asm.
AC_MSG_CHECKING([whether MPI assembler modules are requested])
AC_ARG_ENABLE([asm],
AC_HELP_STRING([--disable-asm],
[Disable MPI assembler modules]),
[try_asm_modules=$enableval],
[try_asm_modules=yes])
AC_MSG_RESULT($try_asm_modules)
# Implementation of the --enable-m-guard switch.
AC_MSG_CHECKING([whether memory guard is requested])
AC_ARG_ENABLE(m-guard,
AC_HELP_STRING([--enable-m-guard],
[Enable memory guard facility]),
[use_m_guard=$enableval], [use_m_guard=no])
AC_MSG_RESULT($use_m_guard)
if test "$use_m_guard" = yes ; then
AC_DEFINE(M_GUARD,1,[Define to use the (obsolete) malloc guarding feature])
fi
# Implementation of the --enable-large-data-tests switch.
AC_MSG_CHECKING([whether to run large data tests])
AC_ARG_ENABLE(large-data-tests,
AC_HELP_STRING([--enable-large-data-tests],
[Enable the real long ruinning large data tests]),
large_data_tests=$enableval,large_data_tests=no)
AC_MSG_RESULT($large_data_tests)
AC_SUBST(RUN_LARGE_DATA_TESTS, $large_data_tests)
# Implementation of the --with-capabilities switch.
# Check whether we want to use Linux capabilities
AC_MSG_CHECKING([whether use of capabilities is requested])
AC_ARG_WITH(capabilities,
AC_HELP_STRING([--with-capabilities],
[Use linux capabilities [default=no]]),
[use_capabilities="$withval"],[use_capabilities=no])
AC_MSG_RESULT($use_capabilities)
# Implementation of the --enable-hmac-binary-check.
AC_MSG_CHECKING([whether a HMAC binary check is requested])
AC_ARG_ENABLE(hmac-binary-check,
AC_HELP_STRING([--enable-hmac-binary-check],
[Enable library integrity check]),
[use_hmac_binary_check=$enableval],
[use_hmac_binary_check=no])
AC_MSG_RESULT($use_hmac_binary_check)
if test "$use_hmac_binary_check" = yes ; then
AC_DEFINE(ENABLE_HMAC_BINARY_CHECK,1,
[Define to support an HMAC based integrity check])
fi
# Implementation of the --disable-jent-support switch.
AC_MSG_CHECKING([whether jitter entropy support is requested])
AC_ARG_ENABLE(jent-support,
AC_HELP_STRING([--disable-jent-support],
[Disable support for the Jitter entropy collector]),
jentsupport=$enableval,jentsupport=yes)
AC_MSG_RESULT($jentsupport)
# Implementation of the --disable-padlock-support switch.
AC_MSG_CHECKING([whether padlock support is requested])
AC_ARG_ENABLE(padlock-support,
AC_HELP_STRING([--disable-padlock-support],
[Disable support for the PadLock Engine of VIA processors]),
padlocksupport=$enableval,padlocksupport=yes)
AC_MSG_RESULT($padlocksupport)
# Implementation of the --disable-aesni-support switch.
AC_MSG_CHECKING([whether AESNI support is requested])
AC_ARG_ENABLE(aesni-support,
AC_HELP_STRING([--disable-aesni-support],
[Disable support for the Intel AES-NI instructions]),
aesnisupport=$enableval,aesnisupport=yes)
AC_MSG_RESULT($aesnisupport)
# Implementation of the --disable-shaext-support switch.
AC_MSG_CHECKING([whether SHAEXT support is requested])
AC_ARG_ENABLE(shaext-support,
AC_HELP_STRING([--disable-shaext-support],
[Disable support for the Intel SHAEXT instructions]),
shaextsupport=$enableval,shaextsupport=yes)
AC_MSG_RESULT($shaextsupport)
# Implementation of the --disable-pclmul-support switch.
AC_MSG_CHECKING([whether PCLMUL support is requested])
AC_ARG_ENABLE(pclmul-support,
AC_HELP_STRING([--disable-pclmul-support],
[Disable support for the Intel PCLMUL instructions]),
pclmulsupport=$enableval,pclmulsupport=yes)
AC_MSG_RESULT($pclmulsupport)
# Implementation of the --disable-sse41-support switch.
AC_MSG_CHECKING([whether SSE4.1 support is requested])
AC_ARG_ENABLE(sse41-support,
AC_HELP_STRING([--disable-sse41-support],
[Disable support for the Intel SSE4.1 instructions]),
sse41support=$enableval,sse41support=yes)
AC_MSG_RESULT($sse41support)
# Implementation of the --disable-drng-support switch.
AC_MSG_CHECKING([whether DRNG support is requested])
AC_ARG_ENABLE(drng-support,
AC_HELP_STRING([--disable-drng-support],
[Disable support for the Intel DRNG (RDRAND instruction)]),
drngsupport=$enableval,drngsupport=yes)
AC_MSG_RESULT($drngsupport)
# Implementation of the --disable-avx-support switch.
AC_MSG_CHECKING([whether AVX support is requested])
AC_ARG_ENABLE(avx-support,
AC_HELP_STRING([--disable-avx-support],
[Disable support for the Intel AVX instructions]),
avxsupport=$enableval,avxsupport=yes)
AC_MSG_RESULT($avxsupport)
# Implementation of the --disable-avx2-support switch.
AC_MSG_CHECKING([whether AVX2 support is requested])
AC_ARG_ENABLE(avx2-support,
AC_HELP_STRING([--disable-avx2-support],
[Disable support for the Intel AVX2 instructions]),
avx2support=$enableval,avx2support=yes)
AC_MSG_RESULT($avx2support)
# Implementation of the --disable-neon-support switch.
AC_MSG_CHECKING([whether NEON support is requested])
AC_ARG_ENABLE(neon-support,
AC_HELP_STRING([--disable-neon-support],
[Disable support for the ARM NEON instructions]),
neonsupport=$enableval,neonsupport=yes)
AC_MSG_RESULT($neonsupport)
# Implementation of the --disable-arm-crypto-support switch.
AC_MSG_CHECKING([whether ARMv8 Crypto Extension support is requested])
AC_ARG_ENABLE(arm-crypto-support,
AC_HELP_STRING([--disable-arm-crypto-support],
[Disable support for the ARMv8 Crypto Extension instructions]),
armcryptosupport=$enableval,armcryptosupport=yes)
AC_MSG_RESULT($armcryptosupport)
# Implementation of the --disable-O-flag-munging switch.
AC_MSG_CHECKING([whether a -O flag munging is requested])
AC_ARG_ENABLE([O-flag-munging],
AC_HELP_STRING([--disable-O-flag-munging],
[Disable modification of the cc -O flag]),
[enable_o_flag_munging=$enableval],
[enable_o_flag_munging=yes])
AC_MSG_RESULT($enable_o_flag_munging)
AM_CONDITIONAL(ENABLE_O_FLAG_MUNGING, test "$enable_o_flag_munging" = "yes")
# Implementation of the --disable-amd64-as-feature-detection switch.
AC_MSG_CHECKING([whether to enable AMD64 as(1) feature detection])
AC_ARG_ENABLE(amd64-as-feature-detection,
AC_HELP_STRING([--disable-amd64-as-feature-detection],
[Disable the auto-detection of AMD64 as(1) features]),
amd64_as_feature_detection=$enableval,
amd64_as_feature_detection=yes)
AC_MSG_RESULT($amd64_as_feature_detection)
AC_DEFINE_UNQUOTED(PRINTABLE_OS_NAME, "$PRINTABLE_OS_NAME",
[A human readable text with the name of the OS])
# For some systems we know that we have ld_version scripts.
# Use it then as default.
have_ld_version_script=no
case "${host}" in
*-*-linux*)
have_ld_version_script=yes
;;
*-*-gnu*)
have_ld_version_script=yes
;;
esac
AC_ARG_ENABLE([ld-version-script],
AC_HELP_STRING([--enable-ld-version-script],
[enable/disable use of linker version script.
(default is system dependent)]),
[have_ld_version_script=$enableval],
[ : ] )
AM_CONDITIONAL(HAVE_LD_VERSION_SCRIPT, test "$have_ld_version_script" = "yes")
AC_DEFINE_UNQUOTED(NAME_OF_DEV_RANDOM, "$NAME_OF_DEV_RANDOM",
[defined to the name of the strong random device])
AC_DEFINE_UNQUOTED(NAME_OF_DEV_URANDOM, "$NAME_OF_DEV_URANDOM",
[defined to the name of the weaker random device])
###############################
#### Checks for libraries. ####
###############################
#
# gpg-error is required.
#
AM_PATH_GPG_ERROR("$NEED_GPG_ERROR_VERSION")
if test "x$GPG_ERROR_LIBS" = "x"; then
AC_MSG_ERROR([libgpg-error is needed.
See ftp://ftp.gnupg.org/gcrypt/libgpg-error/ .])
fi
AC_DEFINE(GPG_ERR_SOURCE_DEFAULT, GPG_ERR_SOURCE_GCRYPT,
[The default error source for libgcrypt.])
#
# Check whether the GNU Pth library is available. We require this
# to build the optional gcryptrnd program.
#
AC_ARG_WITH(pth-prefix,
AC_HELP_STRING([--with-pth-prefix=PFX],
[prefix where GNU Pth is installed (optional)]),
pth_config_prefix="$withval", pth_config_prefix="")
if test x$pth_config_prefix != x ; then
PTH_CONFIG="$pth_config_prefix/bin/pth-config"
fi
if test "$use_random_daemon" = "yes"; then
AC_PATH_PROG(PTH_CONFIG, pth-config, no)
if test "$PTH_CONFIG" = "no"; then
AC_MSG_WARN([[
***
*** To build the Libgcrypt's random number daemon
*** we need the support of the GNU Portable Threads Library.
*** Download it from ftp://ftp.gnu.org/gnu/pth/
*** On a Debian GNU/Linux system you might want to try
*** apt-get install libpth-dev
***]])
else
GNUPG_PTH_VERSION_CHECK([1.3.7])
if test $have_pth = yes; then
PTH_CFLAGS=`$PTH_CONFIG --cflags`
PTH_LIBS=`$PTH_CONFIG --ldflags`
PTH_LIBS="$PTH_LIBS `$PTH_CONFIG --libs --all`"
AC_DEFINE(USE_GNU_PTH, 1,
[Defined if the GNU Portable Thread Library should be used])
AC_DEFINE(HAVE_PTH, 1,
[Defined if the GNU Pth is available])
fi
fi
fi
AC_SUBST(PTH_CFLAGS)
AC_SUBST(PTH_LIBS)
#
# Check whether pthreads is available
#
if test "$have_w32_system" != yes; then
AC_CHECK_LIB(pthread,pthread_create,have_pthread=yes)
if test "$have_pthread" = yes; then
AC_DEFINE(HAVE_PTHREAD, 1 ,[Define if we have pthread.])
fi
fi
# Solaris needs -lsocket and -lnsl. Unisys system includes
# gethostbyname in libsocket but needs libnsl for socket.
AC_SEARCH_LIBS(setsockopt, [socket], ,
[AC_SEARCH_LIBS(setsockopt, [socket], , , [-lnsl])])
AC_SEARCH_LIBS(setsockopt, [nsl])
##################################
#### Checks for header files. ####
##################################
AC_HEADER_STDC
AC_CHECK_HEADERS(unistd.h sys/select.h sys/msg.h)
INSERT_SYS_SELECT_H=
if test x"$ac_cv_header_sys_select_h" = xyes; then
INSERT_SYS_SELECT_H=" include "
fi
AC_SUBST(INSERT_SYS_SELECT_H)
##########################################
#### Checks for typedefs, structures, ####
#### and compiler characteristics. ####
##########################################
AC_C_CONST
AC_C_INLINE
AC_TYPE_SIZE_T
AC_TYPE_SIGNAL
AC_DECL_SYS_SIGLIST
AC_TYPE_PID_T
GNUPG_CHECK_TYPEDEF(byte, HAVE_BYTE_TYPEDEF)
GNUPG_CHECK_TYPEDEF(ushort, HAVE_USHORT_TYPEDEF)
GNUPG_CHECK_TYPEDEF(ulong, HAVE_ULONG_TYPEDEF)
GNUPG_CHECK_TYPEDEF(u16, HAVE_U16_TYPEDEF)
GNUPG_CHECK_TYPEDEF(u32, HAVE_U32_TYPEDEF)
gl_TYPE_SOCKLEN_T
case "${host}" in
*-*-mingw32*)
# socklen_t may or may not be defined depending on what headers
# are included. To be safe we use int as this is the actual type.
FALLBACK_SOCKLEN_T="typedef int gcry_socklen_t;"
;;
*)
if test ".$gl_cv_socklen_t_equiv" = "."; then
FALLBACK_SOCKLEN_T="typedef socklen_t gcry_socklen_t;"
else
FALLBACK_SOCKLEN_T="typedef ${gl_cv_socklen_t_equiv} gcry_socklen_t;"
fi
esac
AC_SUBST(FALLBACK_SOCKLEN_T)
#
# Check for __builtin_bswap32 intrinsic.
#
AC_CACHE_CHECK(for __builtin_bswap32,
[gcry_cv_have_builtin_bswap32],
[gcry_cv_have_builtin_bswap32=no
AC_LINK_IFELSE([AC_LANG_PROGRAM([],
[int x = 0; int y = __builtin_bswap32(x); return y;])],
[gcry_cv_have_builtin_bswap32=yes])])
if test "$gcry_cv_have_builtin_bswap32" = "yes" ; then
AC_DEFINE(HAVE_BUILTIN_BSWAP32,1,
[Defined if compiler has '__builtin_bswap32' intrinsic])
fi
#
# Check for __builtin_bswap64 intrinsic.
#
AC_CACHE_CHECK(for __builtin_bswap64,
[gcry_cv_have_builtin_bswap64],
[gcry_cv_have_builtin_bswap64=no
AC_LINK_IFELSE([AC_LANG_PROGRAM([],
[long long x = 0; long long y = __builtin_bswap64(x); return y;])],
[gcry_cv_have_builtin_bswap64=yes])])
if test "$gcry_cv_have_builtin_bswap64" = "yes" ; then
AC_DEFINE(HAVE_BUILTIN_BSWAP64,1,
[Defined if compiler has '__builtin_bswap64' intrinsic])
fi
#
# Check for __builtin_ctz intrinsic.
#
AC_CACHE_CHECK(for __builtin_ctz,
[gcry_cv_have_builtin_ctz],
[gcry_cv_have_builtin_ctz=no
AC_LINK_IFELSE([AC_LANG_PROGRAM([],
[unsigned int x = 0; int y = __builtin_ctz(x); return y;])],
[gcry_cv_have_builtin_ctz=yes])])
if test "$gcry_cv_have_builtin_ctz" = "yes" ; then
AC_DEFINE(HAVE_BUILTIN_CTZ, 1,
[Defined if compiler has '__builtin_ctz' intrinsic])
fi
#
# Check for __sync_synchronize intrinsic.
#
AC_CACHE_CHECK(for __sync_synchronize,
[gcry_cv_have_sync_synchronize],
[gcry_cv_have_sync_synchronize=no
AC_LINK_IFELSE([AC_LANG_PROGRAM([],
[__sync_synchronize(); return 0;])],
[gcry_cv_have_sync_synchronize=yes])])
if test "$gcry_cv_have_sync_synchronize" = "yes" ; then
AC_DEFINE(HAVE_SYNC_SYNCHRONIZE, 1,
[Defined if compiler has '__sync_synchronize' intrinsic])
fi
#
# Check for VLA support (variable length arrays).
#
AC_CACHE_CHECK(whether the variable length arrays are supported,
[gcry_cv_have_vla],
[gcry_cv_have_vla=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[void f1(char *, int);
char foo(int i) {
char b[(i < 0 ? 0 : i) + 1];
f1(b, sizeof b); return b[0];}]])],
[gcry_cv_have_vla=yes])])
if test "$gcry_cv_have_vla" = "yes" ; then
AC_DEFINE(HAVE_VLA,1, [Defined if variable length arrays are supported])
fi
#
# Check for ELF visibility support.
#
AC_CACHE_CHECK(whether the visibility attribute is supported,
gcry_cv_visibility_attribute,
[gcry_cv_visibility_attribute=no
AC_LANG_CONFTEST([AC_LANG_SOURCE(
[[int foo __attribute__ ((visibility ("hidden"))) = 1;
int bar __attribute__ ((visibility ("protected"))) = 1;
]])])
if ${CC-cc} -Werror -S conftest.c -o conftest.s \
1>&AS_MESSAGE_LOG_FD 2>&AS_MESSAGE_LOG_FD ; then
if grep '\.hidden.*foo' conftest.s >/dev/null 2>&1 ; then
if grep '\.protected.*bar' conftest.s >/dev/null 2>&1; then
gcry_cv_visibility_attribute=yes
fi
fi
fi
])
if test "$gcry_cv_visibility_attribute" = "yes"; then
AC_CACHE_CHECK(for broken visibility attribute,
gcry_cv_broken_visibility_attribute,
[gcry_cv_broken_visibility_attribute=yes
AC_LANG_CONFTEST([AC_LANG_SOURCE(
[[int foo (int x);
int bar (int x) __asm__ ("foo")
__attribute__ ((visibility ("hidden")));
int bar (int x) { return x; }
]])])
if ${CC-cc} -Werror -S conftest.c -o conftest.s \
1>&AS_MESSAGE_LOG_FD 2>&AS_MESSAGE_LOG_FD ; then
if grep '\.hidden@<:@ _@:>@foo' conftest.s >/dev/null 2>&1;
then
gcry_cv_broken_visibility_attribute=no
fi
fi
])
fi
if test "$gcry_cv_visibility_attribute" = "yes"; then
AC_CACHE_CHECK(for broken alias attribute,
gcry_cv_broken_alias_attribute,
[gcry_cv_broken_alias_attribute=yes
AC_LANG_CONFTEST([AC_LANG_SOURCE(
[[extern int foo (int x) __asm ("xyzzy");
int bar (int x) { return x; }
extern __typeof (bar) foo __attribute ((weak, alias ("bar")));
extern int dfoo;
extern __typeof (dfoo) dfoo __asm ("abccb");
int dfoo = 1;
]])])
if ${CC-cc} -Werror -S conftest.c -o conftest.s \
1>&AS_MESSAGE_LOG_FD 2>&AS_MESSAGE_LOG_FD ; then
if grep 'xyzzy' conftest.s >/dev/null 2>&1 && \
grep 'abccb' conftest.s >/dev/null 2>&1; then
gcry_cv_broken_alias_attribute=no
fi
fi
])
fi
if test "$gcry_cv_visibility_attribute" = "yes"; then
AC_CACHE_CHECK(if gcc supports -fvisibility=hidden,
gcry_cv_gcc_has_f_visibility,
[gcry_cv_gcc_has_f_visibility=no
_gcc_cflags_save=$CFLAGS
CFLAGS="-fvisibility=hidden"
AC_COMPILE_IFELSE([AC_LANG_PROGRAM([],[])],
gcry_cv_gcc_has_f_visibility=yes)
CFLAGS=$_gcc_cflags_save;
])
fi
if test "$gcry_cv_visibility_attribute" = "yes" \
&& test "$gcry_cv_broken_visibility_attribute" != "yes" \
&& test "$gcry_cv_broken_alias_attribute" != "yes" \
&& test "$gcry_cv_gcc_has_f_visibility" = "yes"
then
AC_DEFINE(GCRY_USE_VISIBILITY, 1,
[Define to use the GNU C visibility attribute.])
CFLAGS="$CFLAGS -fvisibility=hidden"
fi
# Following attribute tests depend on warnings to cause compile to fail,
# so set -Werror temporarily.
_gcc_cflags_save=$CFLAGS
CFLAGS="$CFLAGS -Werror"
#
# Check whether the compiler supports the GCC style aligned attribute
#
AC_CACHE_CHECK([whether the GCC style aligned attribute is supported],
[gcry_cv_gcc_attribute_aligned],
[gcry_cv_gcc_attribute_aligned=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[struct { int a; } foo __attribute__ ((aligned (16)));]])],
[gcry_cv_gcc_attribute_aligned=yes])])
if test "$gcry_cv_gcc_attribute_aligned" = "yes" ; then
AC_DEFINE(HAVE_GCC_ATTRIBUTE_ALIGNED,1,
[Defined if a GCC style "__attribute__ ((aligned (n))" is supported])
fi
#
# Check whether the compiler supports the GCC style packed attribute
#
AC_CACHE_CHECK([whether the GCC style packed attribute is supported],
[gcry_cv_gcc_attribute_packed],
[gcry_cv_gcc_attribute_packed=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[struct foolong_s { long b; } __attribute__ ((packed));
struct foo_s { char a; struct foolong_s b; }
__attribute__ ((packed));
enum bar {
FOO = 1 / (sizeof(struct foo_s) == (sizeof(char) + sizeof(long))),
};]])],
[gcry_cv_gcc_attribute_packed=yes])])
if test "$gcry_cv_gcc_attribute_packed" = "yes" ; then
AC_DEFINE(HAVE_GCC_ATTRIBUTE_PACKED,1,
[Defined if a GCC style "__attribute__ ((packed))" is supported])
fi
#
# Check whether the compiler supports the GCC style may_alias attribute
#
AC_CACHE_CHECK([whether the GCC style may_alias attribute is supported],
[gcry_cv_gcc_attribute_may_alias],
[gcry_cv_gcc_attribute_may_alias=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[typedef struct foo_s { int a; }
__attribute__ ((may_alias)) foo_t;]])],
[gcry_cv_gcc_attribute_may_alias=yes])])
if test "$gcry_cv_gcc_attribute_may_alias" = "yes" ; then
AC_DEFINE(HAVE_GCC_ATTRIBUTE_MAY_ALIAS,1,
[Defined if a GCC style "__attribute__ ((may_alias))" is supported])
fi
# Restore flags.
CFLAGS=$_gcc_cflags_save;
#
# Check whether the compiler supports 'asm' or '__asm__' keyword for
# assembler blocks.
#
AC_CACHE_CHECK([whether 'asm' assembler keyword is supported],
[gcry_cv_have_asm],
[gcry_cv_have_asm=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[void a(void) { asm("":::"memory"); }]])],
[gcry_cv_have_asm=yes])])
AC_CACHE_CHECK([whether '__asm__' assembler keyword is supported],
[gcry_cv_have___asm__],
[gcry_cv_have___asm__=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[void a(void) { __asm__("":::"memory"); }]])],
[gcry_cv_have___asm__=yes])])
if test "$gcry_cv_have_asm" = "no" ; then
if test "$gcry_cv_have___asm__" = "yes" ; then
AC_DEFINE(asm,__asm__,
[Define to supported assembler block keyword, if plain 'asm' was not
supported])
fi
fi
#
# Check whether the compiler supports inline assembly memory barrier.
#
if test "$gcry_cv_have_asm" = "no" ; then
if test "$gcry_cv_have___asm__" = "yes" ; then
AC_CACHE_CHECK([whether inline assembly memory barrier is supported],
[gcry_cv_have_asm_volatile_memory],
[gcry_cv_have_asm_volatile_memory=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[void a(void) { __asm__ volatile("":::"memory"); }]])],
[gcry_cv_have_asm_volatile_memory=yes])])
fi
else
AC_CACHE_CHECK([whether inline assembly memory barrier is supported],
[gcry_cv_have_asm_volatile_memory],
[gcry_cv_have_asm_volatile_memory=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[void a(void) { asm volatile("":::"memory"); }]])],
[gcry_cv_have_asm_volatile_memory=yes])])
fi
if test "$gcry_cv_have_asm_volatile_memory" = "yes" ; then
AC_DEFINE(HAVE_GCC_ASM_VOLATILE_MEMORY,1,
[Define if inline asm memory barrier is supported])
fi
#
# Check whether GCC assembler supports features needed for our ARM
# implementations. This needs to be done before setting up the
# assembler stuff.
#
AC_CACHE_CHECK([whether GCC assembler is compatible for ARM assembly implementations],
[gcry_cv_gcc_arm_platform_as_ok],
[gcry_cv_gcc_arm_platform_as_ok=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[__asm__(
/* Test if assembler supports UAL syntax. */
".syntax unified\n\t"
".arm\n\t" /* our assembly code is in ARM mode */
/* Following causes error if assembler ignored '.syntax unified'. */
"asmfunc:\n\t"
"add %r0, %r0, %r4, ror #12;\n\t"
/* Test if '.type' and '.size' are supported. */
".size asmfunc,.-asmfunc;\n\t"
".type asmfunc,%function;\n\t"
);]])],
[gcry_cv_gcc_arm_platform_as_ok=yes])])
if test "$gcry_cv_gcc_arm_platform_as_ok" = "yes" ; then
AC_DEFINE(HAVE_COMPATIBLE_GCC_ARM_PLATFORM_AS,1,
[Defined if underlying assembler is compatible with ARM assembly implementations])
fi
#
# Check whether GCC assembler supports features needed for our ARMv8/Aarch64
# implementations. This needs to be done before setting up the
# assembler stuff.
#
AC_CACHE_CHECK([whether GCC assembler is compatible for ARMv8/Aarch64 assembly implementations],
[gcry_cv_gcc_aarch64_platform_as_ok],
[gcry_cv_gcc_aarch64_platform_as_ok=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[__asm__(
"asmfunc:\n\t"
"eor x0, x0, x30, ror #12;\n\t"
"add x0, x0, x30, asr #12;\n\t"
"eor v0.16b, v0.16b, v31.16b;\n\t"
);]])],
[gcry_cv_gcc_aarch64_platform_as_ok=yes])])
if test "$gcry_cv_gcc_aarch64_platform_as_ok" = "yes" ; then
AC_DEFINE(HAVE_COMPATIBLE_GCC_AARCH64_PLATFORM_AS,1,
[Defined if underlying assembler is compatible with ARMv8/Aarch64 assembly implementations])
fi
#
# Check whether underscores in symbols are required. This needs to be
# done before setting up the assembler stuff.
#
GNUPG_SYS_SYMBOL_UNDERSCORE()
#################################
#### ####
#### Setup assembler stuff. ####
#### Define mpi_cpu_arch. ####
#### ####
#################################
AC_ARG_ENABLE(mpi-path,
AC_HELP_STRING([--enable-mpi-path=EXTRA_PATH],
[prepend EXTRA_PATH to list of CPU specific optimizations]),
mpi_extra_path="$enableval",mpi_extra_path="")
AC_MSG_CHECKING(architecture and mpi assembler functions)
if test -f $srcdir/mpi/config.links ; then
. $srcdir/mpi/config.links
AC_CONFIG_LINKS("$mpi_ln_list")
ac_cv_mpi_sflags="$mpi_sflags"
AC_MSG_RESULT($mpi_cpu_arch)
else
AC_MSG_RESULT(failed)
AC_MSG_ERROR([mpi/config.links missing!])
fi
MPI_SFLAGS="$ac_cv_mpi_sflags"
AC_SUBST(MPI_SFLAGS)
AM_CONDITIONAL(MPI_MOD_ASM_MPIH_ADD1, test "$mpi_mod_asm_mpih_add1" = yes)
AM_CONDITIONAL(MPI_MOD_ASM_MPIH_SUB1, test "$mpi_mod_asm_mpih_sub1" = yes)
AM_CONDITIONAL(MPI_MOD_ASM_MPIH_MUL1, test "$mpi_mod_asm_mpih_mul1" = yes)
AM_CONDITIONAL(MPI_MOD_ASM_MPIH_MUL2, test "$mpi_mod_asm_mpih_mul2" = yes)
AM_CONDITIONAL(MPI_MOD_ASM_MPIH_MUL3, test "$mpi_mod_asm_mpih_mul3" = yes)
AM_CONDITIONAL(MPI_MOD_ASM_MPIH_LSHIFT, test "$mpi_mod_asm_mpih_lshift" = yes)
AM_CONDITIONAL(MPI_MOD_ASM_MPIH_RSHIFT, test "$mpi_mod_asm_mpih_rshift" = yes)
AM_CONDITIONAL(MPI_MOD_ASM_UDIV, test "$mpi_mod_asm_udiv" = yes)
AM_CONDITIONAL(MPI_MOD_ASM_UDIV_QRNND, test "$mpi_mod_asm_udiv_qrnnd" = yes)
AM_CONDITIONAL(MPI_MOD_C_MPIH_ADD1, test "$mpi_mod_c_mpih_add1" = yes)
AM_CONDITIONAL(MPI_MOD_C_MPIH_SUB1, test "$mpi_mod_c_mpih_sub1" = yes)
AM_CONDITIONAL(MPI_MOD_C_MPIH_MUL1, test "$mpi_mod_c_mpih_mul1" = yes)
AM_CONDITIONAL(MPI_MOD_C_MPIH_MUL2, test "$mpi_mod_c_mpih_mul2" = yes)
AM_CONDITIONAL(MPI_MOD_C_MPIH_MUL3, test "$mpi_mod_c_mpih_mul3" = yes)
AM_CONDITIONAL(MPI_MOD_C_MPIH_LSHIFT, test "$mpi_mod_c_mpih_lshift" = yes)
AM_CONDITIONAL(MPI_MOD_C_MPIH_RSHIFT, test "$mpi_mod_c_mpih_rshift" = yes)
AM_CONDITIONAL(MPI_MOD_C_UDIV, test "$mpi_mod_c_udiv" = yes)
AM_CONDITIONAL(MPI_MOD_C_UDIV_QRNND, test "$mpi_mod_c_udiv_qrnnd" = yes)
# Reset non applicable feature flags.
if test "$mpi_cpu_arch" != "x86" ; then
aesnisupport="n/a"
shaextsupport="n/a"
pclmulsupport="n/a"
sse41support="n/a"
avxsupport="n/a"
avx2support="n/a"
padlocksupport="n/a"
jentsupport="n/a"
drngsupport="n/a"
fi
if test "$mpi_cpu_arch" != "arm" ; then
if test "$mpi_cpu_arch" != "aarch64" ; then
neonsupport="n/a"
armcryptosupport="n/a"
fi
fi
#############################################
#### ####
#### Platform specific compiler checks. ####
#### ####
#############################################
# Following tests depend on warnings to cause compile to fail, so set -Werror
# temporarily.
_gcc_cflags_save=$CFLAGS
CFLAGS="$CFLAGS -Werror"
#
# Check whether compiler supports 'ms_abi' function attribute.
#
AC_CACHE_CHECK([whether compiler supports 'ms_abi' function attribute],
[gcry_cv_gcc_attribute_ms_abi],
[gcry_cv_gcc_attribute_ms_abi=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[int __attribute__ ((ms_abi)) proto(int);]])],
[gcry_cv_gcc_attribute_ms_abi=yes])])
if test "$gcry_cv_gcc_attribute_ms_abi" = "yes" ; then
AC_DEFINE(HAVE_GCC_ATTRIBUTE_MS_ABI,1,
[Defined if compiler supports "__attribute__ ((ms_abi))" function attribute])
fi
#
# Check whether compiler supports 'sysv_abi' function attribute.
#
AC_CACHE_CHECK([whether compiler supports 'sysv_abi' function attribute],
[gcry_cv_gcc_attribute_sysv_abi],
[gcry_cv_gcc_attribute_sysv_abi=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[int __attribute__ ((sysv_abi)) proto(int);]])],
[gcry_cv_gcc_attribute_sysv_abi=yes])])
if test "$gcry_cv_gcc_attribute_sysv_abi" = "yes" ; then
AC_DEFINE(HAVE_GCC_ATTRIBUTE_SYSV_ABI,1,
[Defined if compiler supports "__attribute__ ((sysv_abi))" function attribute])
fi
#
# Check whether default calling convention is 'ms_abi'.
#
if test "$gcry_cv_gcc_attribute_ms_abi" = "yes" ; then
AC_CACHE_CHECK([whether default calling convention is 'ms_abi'],
[gcry_cv_gcc_default_abi_is_ms_abi],
[gcry_cv_gcc_default_abi_is_ms_abi=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[void *test(void) {
void *(*def_func)(void) = test;
void *__attribute__((ms_abi))(*msabi_func)(void);
/* warning on SysV abi targets, passes on Windows based targets */
msabi_func = def_func;
return msabi_func;
}]])],
[gcry_cv_gcc_default_abi_is_ms_abi=yes])])
if test "$gcry_cv_gcc_default_abi_is_ms_abi" = "yes" ; then
AC_DEFINE(HAVE_GCC_DEFAULT_ABI_IS_MS_ABI,1,
[Defined if default calling convention is 'ms_abi'])
fi
fi
#
# Check whether default calling convention is 'sysv_abi'.
#
if test "$gcry_cv_gcc_attribute_sysv_abi" = "yes" ; then
AC_CACHE_CHECK([whether default calling convention is 'sysv_abi'],
[gcry_cv_gcc_default_abi_is_sysv_abi],
[gcry_cv_gcc_default_abi_is_sysv_abi=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[void *test(void) {
void *(*def_func)(void) = test;
void *__attribute__((sysv_abi))(*sysvabi_func)(void);
/* warning on MS ABI targets, passes on SysV ABI targets */
sysvabi_func = def_func;
return sysvabi_func;
}]])],
[gcry_cv_gcc_default_abi_is_sysv_abi=yes])])
if test "$gcry_cv_gcc_default_abi_is_sysv_abi" = "yes" ; then
AC_DEFINE(HAVE_GCC_DEFAULT_ABI_IS_SYSV_ABI,1,
[Defined if default calling convention is 'sysv_abi'])
fi
fi
# Restore flags.
CFLAGS=$_gcc_cflags_save;
#
# Check whether GCC inline assembler supports SSSE3 instructions
# This is required for the AES-NI instructions.
#
AC_CACHE_CHECK([whether GCC inline assembler supports SSSE3 instructions],
[gcry_cv_gcc_inline_asm_ssse3],
[if test "$mpi_cpu_arch" != "x86" ; then
gcry_cv_gcc_inline_asm_ssse3="n/a"
else
gcry_cv_gcc_inline_asm_ssse3=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[static unsigned char be_mask[16] __attribute__ ((aligned (16))) =
{ 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0 };
void a(void) {
__asm__("pshufb %[mask], %%xmm2\n\t"::[mask]"m"(*be_mask):);
}]])],
[gcry_cv_gcc_inline_asm_ssse3=yes])
fi])
if test "$gcry_cv_gcc_inline_asm_ssse3" = "yes" ; then
AC_DEFINE(HAVE_GCC_INLINE_ASM_SSSE3,1,
[Defined if inline assembler supports SSSE3 instructions])
fi
#
# Check whether GCC inline assembler supports PCLMUL instructions.
#
AC_CACHE_CHECK([whether GCC inline assembler supports PCLMUL instructions],
[gcry_cv_gcc_inline_asm_pclmul],
[if test "$mpi_cpu_arch" != "x86" ; then
gcry_cv_gcc_inline_asm_pclmul="n/a"
else
gcry_cv_gcc_inline_asm_pclmul=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[void a(void) {
__asm__("pclmulqdq \$0, %%xmm1, %%xmm3\n\t":::"cc");
}]])],
[gcry_cv_gcc_inline_asm_pclmul=yes])
fi])
if test "$gcry_cv_gcc_inline_asm_pclmul" = "yes" ; then
AC_DEFINE(HAVE_GCC_INLINE_ASM_PCLMUL,1,
[Defined if inline assembler supports PCLMUL instructions])
fi
#
# Check whether GCC inline assembler supports SHA Extensions instructions.
#
AC_CACHE_CHECK([whether GCC inline assembler supports SHA Extensions instructions],
[gcry_cv_gcc_inline_asm_shaext],
[if test "$mpi_cpu_arch" != "x86" ; then
gcry_cv_gcc_inline_asm_shaext="n/a"
else
gcry_cv_gcc_inline_asm_shaext=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[void a(void) {
__asm__("sha1rnds4 \$0, %%xmm1, %%xmm3\n\t":::"cc");
__asm__("sha1nexte %%xmm1, %%xmm3\n\t":::"cc");
__asm__("sha1msg1 %%xmm1, %%xmm3\n\t":::"cc");
__asm__("sha1msg2 %%xmm1, %%xmm3\n\t":::"cc");
__asm__("sha256rnds2 %%xmm0, %%xmm1, %%xmm3\n\t":::"cc");
__asm__("sha256msg1 %%xmm1, %%xmm3\n\t":::"cc");
__asm__("sha256msg2 %%xmm1, %%xmm3\n\t":::"cc");
}]])],
[gcry_cv_gcc_inline_asm_shaext=yes])
fi])
if test "$gcry_cv_gcc_inline_asm_shaext" = "yes" ; then
AC_DEFINE(HAVE_GCC_INLINE_ASM_SHAEXT,1,
[Defined if inline assembler supports SHA Extensions instructions])
fi
#
# Check whether GCC inline assembler supports SSE4.1 instructions.
#
AC_CACHE_CHECK([whether GCC inline assembler supports SSE4.1 instructions],
[gcry_cv_gcc_inline_asm_sse41],
[if test "$mpi_cpu_arch" != "x86" ; then
gcry_cv_gcc_inline_asm_sse41="n/a"
else
gcry_cv_gcc_inline_asm_sse41=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[void a(void) {
int i;
__asm__("pextrd \$2, %%xmm0, %[out]\n\t" : [out] "=m" (i));
}]])],
[gcry_cv_gcc_inline_asm_sse41=yes])
fi])
if test "$gcry_cv_gcc_inline_asm_sse41" = "yes" ; then
AC_DEFINE(HAVE_GCC_INLINE_ASM_SSE41,1,
[Defined if inline assembler supports SSE4.1 instructions])
fi
#
# Check whether GCC inline assembler supports AVX instructions
#
AC_CACHE_CHECK([whether GCC inline assembler supports AVX instructions],
[gcry_cv_gcc_inline_asm_avx],
[if test "$mpi_cpu_arch" != "x86" ; then
gcry_cv_gcc_inline_asm_avx="n/a"
else
gcry_cv_gcc_inline_asm_avx=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[void a(void) {
__asm__("xgetbv; vaesdeclast (%[mem]),%%xmm0,%%xmm7\n\t"::[mem]"r"(0):);
}]])],
[gcry_cv_gcc_inline_asm_avx=yes])
fi])
if test "$gcry_cv_gcc_inline_asm_avx" = "yes" ; then
AC_DEFINE(HAVE_GCC_INLINE_ASM_AVX,1,
[Defined if inline assembler supports AVX instructions])
fi
#
# Check whether GCC inline assembler supports AVX2 instructions
#
AC_CACHE_CHECK([whether GCC inline assembler supports AVX2 instructions],
[gcry_cv_gcc_inline_asm_avx2],
[if test "$mpi_cpu_arch" != "x86" ; then
gcry_cv_gcc_inline_asm_avx2="n/a"
else
gcry_cv_gcc_inline_asm_avx2=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[void a(void) {
__asm__("xgetbv; vpbroadcastb %%xmm7,%%ymm1\n\t":::"cc");
}]])],
[gcry_cv_gcc_inline_asm_avx2=yes])
fi])
if test "$gcry_cv_gcc_inline_asm_avx2" = "yes" ; then
AC_DEFINE(HAVE_GCC_INLINE_ASM_AVX2,1,
[Defined if inline assembler supports AVX2 instructions])
fi
#
# Check whether GCC inline assembler supports BMI2 instructions
#
AC_CACHE_CHECK([whether GCC inline assembler supports BMI2 instructions],
[gcry_cv_gcc_inline_asm_bmi2],
[if test "$mpi_cpu_arch" != "x86" ; then
gcry_cv_gcc_inline_asm_bmi2="n/a"
else
gcry_cv_gcc_inline_asm_bmi2=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[unsigned int a(unsigned int x, unsigned int y) {
unsigned int tmp1, tmp2;
asm ("rorxl %2, %1, %0"
: "=r" (tmp1)
: "rm0" (x), "J" (32 - ((23) & 31)));
asm ("andnl %2, %1, %0"
: "=r" (tmp2)
: "r0" (x), "rm" (y));
return tmp1 + tmp2;
}]])],
[gcry_cv_gcc_inline_asm_bmi2=yes])
fi])
if test "$gcry_cv_gcc_inline_asm_bmi2" = "yes" ; then
AC_DEFINE(HAVE_GCC_INLINE_ASM_BMI2,1,
[Defined if inline assembler supports BMI2 instructions])
fi
#
# Check whether GCC assembler needs "-Wa,--divide" to correctly handle
# constant division
#
if test $amd64_as_feature_detection = yes; then
AC_CACHE_CHECK([whether GCC assembler handles division correctly],
[gcry_cv_gcc_as_const_division_ok],
[gcry_cv_gcc_as_const_division_ok=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[__asm__("xorl \$(123456789/12345678), %ebp;\n\t");]])],
[gcry_cv_gcc_as_const_division_ok=yes])])
if test "$gcry_cv_gcc_as_const_division_ok" = "no" ; then
#
# Add '-Wa,--divide' to CPPFLAGS and try check again.
#
_gcc_cppflags_save="$CPPFLAGS"
CPPFLAGS="$CPPFLAGS -Wa,--divide"
AC_CACHE_CHECK([whether GCC assembler handles division correctly with "-Wa,--divide"],
[gcry_cv_gcc_as_const_division_with_wadivide_ok],
[gcry_cv_gcc_as_const_division_with_wadivide_ok=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[__asm__("xorl \$(123456789/12345678), %ebp;\n\t");]])],
[gcry_cv_gcc_as_const_division_with_wadivide_ok=yes])])
if test "$gcry_cv_gcc_as_const_division_with_wadivide_ok" = "no" ; then
# '-Wa,--divide' did not work, restore old flags.
CPPFLAGS="$_gcc_cppflags_save"
fi
fi
fi
#
# Check whether GCC assembler supports features needed for our amd64
# implementations
#
if test $amd64_as_feature_detection = yes; then
AC_CACHE_CHECK([whether GCC assembler is compatible for amd64 assembly implementations],
[gcry_cv_gcc_amd64_platform_as_ok],
[if test "$mpi_cpu_arch" != "x86" ; then
gcry_cv_gcc_amd64_platform_as_ok="n/a"
else
gcry_cv_gcc_amd64_platform_as_ok=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[__asm__(
/* Test if '.type' and '.size' are supported. */
/* These work only on ELF targets. */
"asmfunc:\n\t"
".size asmfunc,.-asmfunc;\n\t"
".type asmfunc,@function;\n\t"
/* Test if assembler allows use of '/' for constant division
* (Solaris/x86 issue). If previous constant division check
* and "-Wa,--divide" workaround failed, this causes assembly
* to be disable on this machine. */
"xorl \$(123456789/12345678), %ebp;\n\t"
);]])],
[gcry_cv_gcc_amd64_platform_as_ok=yes])
fi])
if test "$gcry_cv_gcc_amd64_platform_as_ok" = "yes" ; then
AC_DEFINE(HAVE_COMPATIBLE_GCC_AMD64_PLATFORM_AS,1,
[Defined if underlying assembler is compatible with amd64 assembly implementations])
fi
if test "$gcry_cv_gcc_amd64_platform_as_ok" = "no" &&
test "$gcry_cv_gcc_attribute_sysv_abi" = "yes" &&
test "$gcry_cv_gcc_default_abi_is_ms_abi" = "yes"; then
AC_CACHE_CHECK([whether GCC assembler is compatible for WIN64 assembly implementations],
[gcry_cv_gcc_win64_platform_as_ok],
[gcry_cv_gcc_win64_platform_as_ok=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[__asm__(
".globl asmfunc\n\t"
"asmfunc:\n\t"
"xorq \$(1234), %rbp;\n\t"
);]])],
[gcry_cv_gcc_win64_platform_as_ok=yes])])
if test "$gcry_cv_gcc_win64_platform_as_ok" = "yes" ; then
AC_DEFINE(HAVE_COMPATIBLE_GCC_WIN64_PLATFORM_AS,1,
[Defined if underlying assembler is compatible with WIN64 assembly implementations])
fi
fi
fi
#
# Check whether GCC assembler supports features needed for assembly
# implementations that use Intel syntax
#
AC_CACHE_CHECK([whether GCC assembler is compatible for Intel syntax assembly implementations],
[gcry_cv_gcc_platform_as_ok_for_intel_syntax],
[if test "$mpi_cpu_arch" != "x86" ; then
gcry_cv_gcc_platform_as_ok_for_intel_syntax="n/a"
else
gcry_cv_gcc_platform_as_ok_for_intel_syntax=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[__asm__(
".intel_syntax noprefix\n\t"
"pxor xmm1, xmm7;\n\t"
/* Intel syntax implementation also use GAS macros, so check
* for them here. */
"VAL_A = xmm4\n\t"
"VAL_B = xmm2\n\t"
".macro SET_VAL_A p1\n\t"
" VAL_A = \\\\p1 \n\t"
".endm\n\t"
".macro SET_VAL_B p1\n\t"
" VAL_B = \\\\p1 \n\t"
".endm\n\t"
"vmovdqa VAL_A, VAL_B;\n\t"
"SET_VAL_A eax\n\t"
"SET_VAL_B ebp\n\t"
"add VAL_A, VAL_B;\n\t"
"add VAL_B, 0b10101;\n\t"
);]])],
[gcry_cv_gcc_platform_as_ok_for_intel_syntax=yes])
fi])
if test "$gcry_cv_gcc_platform_as_ok_for_intel_syntax" = "yes" ; then
AC_DEFINE(HAVE_INTEL_SYNTAX_PLATFORM_AS,1,
[Defined if underlying assembler is compatible with Intel syntax assembly implementations])
fi
#
# Check whether compiler is configured for ARMv6 or newer architecture
#
AC_CACHE_CHECK([whether compiler is configured for ARMv6 or newer architecture],
[gcry_cv_cc_arm_arch_is_v6],
[if test "$mpi_cpu_arch" != "arm" ; then
gcry_cv_cc_arm_arch_is_v6="n/a"
else
gcry_cv_cc_arm_arch_is_v6=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[
#if defined(__arm__) && \
((defined(__ARM_ARCH) && __ARM_ARCH >= 6) \
|| defined(__ARM_ARCH_6__) || defined(__ARM_ARCH_6J__) \
|| defined(__ARM_ARCH_6Z__) || defined(__ARM_ARCH_6ZK__) \
|| defined(__ARM_ARCH_6K__) || defined(__ARM_ARCH_6T2__) \
|| defined(__ARM_ARCH_7__) || defined(__ARM_ARCH_7A__) \
|| defined(__ARM_ARCH_7R__) || defined(__ARM_ARCH_7M__) \
|| defined(__ARM_ARCH_7EM__))
/* empty */
#else
/* fail compile if not ARMv6. */
not_armv6 not_armv6 = (not_armv6)not_armv6;
#endif
]])],
[gcry_cv_cc_arm_arch_is_v6=yes])
fi])
if test "$gcry_cv_cc_arm_arch_is_v6" = "yes" ; then
AC_DEFINE(HAVE_ARM_ARCH_V6,1,
[Defined if ARM architecture is v6 or newer])
fi
#
# Check whether GCC inline assembler supports NEON instructions
#
AC_CACHE_CHECK([whether GCC inline assembler supports NEON instructions],
[gcry_cv_gcc_inline_asm_neon],
[if test "$mpi_cpu_arch" != "arm" ; then
gcry_cv_gcc_inline_asm_neon="n/a"
else
gcry_cv_gcc_inline_asm_neon=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[__asm__(
".syntax unified\n\t"
".arm\n\t"
".fpu neon\n\t"
"vld1.64 {%q0-%q1}, [%r0]!;\n\t"
"vrev64.8 %q0, %q3;\n\t"
"vadd.u64 %q0, %q1;\n\t"
"vadd.s64 %d3, %d2, %d3;\n\t"
);
]])],
[gcry_cv_gcc_inline_asm_neon=yes])
fi])
if test "$gcry_cv_gcc_inline_asm_neon" = "yes" ; then
AC_DEFINE(HAVE_GCC_INLINE_ASM_NEON,1,
[Defined if inline assembler supports NEON instructions])
fi
#
# Check whether GCC inline assembler supports AArch32 Crypto Extension instructions
#
AC_CACHE_CHECK([whether GCC inline assembler supports AArch32 Crypto Extension instructions],
[gcry_cv_gcc_inline_asm_aarch32_crypto],
[if test "$mpi_cpu_arch" != "arm" ; then
gcry_cv_gcc_inline_asm_aarch32_crypto="n/a"
else
gcry_cv_gcc_inline_asm_aarch32_crypto=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[__asm__(
".syntax unified\n\t"
".arch armv8-a\n\t"
".arm\n\t"
".fpu crypto-neon-fp-armv8\n\t"
"sha1h.32 q0, q0;\n\t"
"sha1c.32 q0, q0, q0;\n\t"
"sha1p.32 q0, q0, q0;\n\t"
"sha1su0.32 q0, q0, q0;\n\t"
"sha1su1.32 q0, q0;\n\t"
"sha256h.32 q0, q0, q0;\n\t"
"sha256h2.32 q0, q0, q0;\n\t"
"sha1p.32 q0, q0, q0;\n\t"
"sha256su0.32 q0, q0;\n\t"
"sha256su1.32 q0, q0, q15;\n\t"
"aese.8 q0, q0;\n\t"
"aesd.8 q0, q0;\n\t"
"aesmc.8 q0, q0;\n\t"
"aesimc.8 q0, q0;\n\t"
"vmull.p64 q0, d0, d0;\n\t"
);
]])],
[gcry_cv_gcc_inline_asm_aarch32_crypto=yes])
fi])
if test "$gcry_cv_gcc_inline_asm_aarch32_crypto" = "yes" ; then
AC_DEFINE(HAVE_GCC_INLINE_ASM_AARCH32_CRYPTO,1,
[Defined if inline assembler supports AArch32 Crypto Extension instructions])
fi
#
# Check whether GCC inline assembler supports AArch64 NEON instructions
#
AC_CACHE_CHECK([whether GCC inline assembler supports AArch64 NEON instructions],
[gcry_cv_gcc_inline_asm_aarch64_neon],
[if test "$mpi_cpu_arch" != "aarch64" ; then
gcry_cv_gcc_inline_asm_aarch64_neon="n/a"
else
gcry_cv_gcc_inline_asm_aarch64_neon=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[__asm__(
".cpu generic+simd\n\t"
"mov w0, \#42;\n\t"
"dup v0.8b, w0;\n\t"
"ld4 {v0.8b,v1.8b,v2.8b,v3.8b},[x0],\#32;\n\t"
);
]])],
[gcry_cv_gcc_inline_asm_aarch64_neon=yes])
fi])
if test "$gcry_cv_gcc_inline_asm_aarch64_neon" = "yes" ; then
AC_DEFINE(HAVE_GCC_INLINE_ASM_AARCH64_NEON,1,
[Defined if inline assembler supports AArch64 NEON instructions])
fi
#
# Check whether GCC inline assembler supports AArch64 Crypto Extension instructions
#
AC_CACHE_CHECK([whether GCC inline assembler supports AArch64 Crypto Extension instructions],
[gcry_cv_gcc_inline_asm_aarch64_crypto],
[if test "$mpi_cpu_arch" != "aarch64" ; then
gcry_cv_gcc_inline_asm_aarch64_crypto="n/a"
else
gcry_cv_gcc_inline_asm_aarch64_crypto=no
AC_COMPILE_IFELSE([AC_LANG_SOURCE(
[[__asm__(
".cpu generic+simd+crypto\n\t"
"mov w0, \#42;\n\t"
"dup v0.8b, w0;\n\t"
"ld4 {v0.8b,v1.8b,v2.8b,v3.8b},[x0],\#32;\n\t"
"sha1h s0, s0;\n\t"
"sha1c q0, s0, v0.4s;\n\t"
"sha1p q0, s0, v0.4s;\n\t"
"sha1su0 v0.4s, v0.4s, v0.4s;\n\t"
"sha1su1 v0.4s, v0.4s;\n\t"
"sha256h q0, q0, v0.4s;\n\t"
"sha256h2 q0, q0, v0.4s;\n\t"
"sha1p q0, s0, v0.4s;\n\t"
"sha256su0 v0.4s, v0.4s;\n\t"
"sha256su1 v0.4s, v0.4s, v31.4s;\n\t"
"aese v0.16b, v0.16b;\n\t"
"aesd v0.16b, v0.16b;\n\t"
"aesmc v0.16b, v0.16b;\n\t"
"aesimc v0.16b, v0.16b;\n\t"
"pmull v0.1q, v0.1d, v31.1d;\n\t"
"pmull2 v0.1q, v0.2d, v31.2d;\n\t"
);
]])],
[gcry_cv_gcc_inline_asm_aarch64_crypto=yes])
fi])
if test "$gcry_cv_gcc_inline_asm_aarch64_crypto" = "yes" ; then
AC_DEFINE(HAVE_GCC_INLINE_ASM_AARCH64_CRYPTO,1,
[Defined if inline assembler supports AArch64 Crypto Extension instructions])
fi
#######################################
#### Checks for library functions. ####
#######################################
AC_FUNC_VPRINTF
# We have replacements for these in src/missing-string.c
AC_CHECK_FUNCS(stpcpy strcasecmp)
# We have replacements for these in src/g10lib.h
AC_CHECK_FUNCS(strtoul memmove stricmp atexit raise)
# Other checks
AC_CHECK_FUNCS(strerror rand mmap getpagesize sysconf waitpid wait4)
AC_CHECK_FUNCS(gettimeofday getrusage gethrtime clock_gettime syslog)
AC_CHECK_FUNCS(syscall fcntl ftruncate flockfile)
AC_CHECK_FUNCS(explicit_bzero getentropy)
GNUPG_CHECK_MLOCK
#
# Replacement functions.
#
AC_REPLACE_FUNCS([getpid clock])
#
# Check whether it is necessary to link against libdl.
#
DL_LIBS=""
if test "$use_hmac_binary_check" = yes ; then
_gcry_save_libs="$LIBS"
LIBS=""
AC_SEARCH_LIBS(dlopen, c dl,,,)
DL_LIBS=$LIBS
LIBS="$_gcry_save_libs"
LIBGCRYPT_CONFIG_LIBS="${LIBGCRYPT_CONFIG_LIBS} ${DL_LIBS}"
fi
AC_SUBST(DL_LIBS)
#
# Check whether we can use Linux capabilities as requested.
#
if test "$use_capabilities" = "yes" ; then
use_capabilities=no
AC_CHECK_HEADERS(sys/capability.h)
if test "$ac_cv_header_sys_capability_h" = "yes" ; then
AC_CHECK_LIB(cap, cap_init, ac_need_libcap=1)
if test "$ac_cv_lib_cap_cap_init" = "yes"; then
AC_DEFINE(USE_CAPABILITIES,1,
[define if capabilities should be used])
LIBS="$LIBS -lcap"
use_capabilities=yes
fi
fi
if test "$use_capabilities" = "no" ; then
AC_MSG_WARN([[
***
*** The use of capabilities on this system is not possible.
*** You need a recent Linux kernel and some patches:
*** fcaps-2.2.9-990610.patch (kernel patch for 2.2.9)
*** fcap-module-990613.tar.gz (kernel module)
*** libcap-1.92.tar.gz (user mode library and utilities)
*** And you have to configure the kernel with CONFIG_VFS_CAP_PLUGIN
*** set (filesystems menu). Be warned: This code is *really* ALPHA.
***]])
fi
fi
# Check whether a random device is available.
if test "$try_dev_random" = yes ; then
AC_CACHE_CHECK(for random device, ac_cv_have_dev_random,
[if test -r "$NAME_OF_DEV_RANDOM" && test -r "$NAME_OF_DEV_URANDOM" ; then
ac_cv_have_dev_random=yes; else ac_cv_have_dev_random=no; fi])
if test "$ac_cv_have_dev_random" = yes; then
AC_DEFINE(HAVE_DEV_RANDOM,1,
[defined if the system supports a random device] )
fi
else
AC_MSG_CHECKING(for random device)
ac_cv_have_dev_random=no
AC_MSG_RESULT(has been disabled)
fi
# Figure out the random modules for this configuration.
if test "$random" = "default"; then
# Select default value.
if test "$ac_cv_have_dev_random" = yes; then
# Try Linuxish random device.
random_modules="linux"
else
case "${host}" in
*-*-mingw32ce*)
# WindowsCE random device.
random_modules="w32ce"
;;
*-*-mingw32*|*-*-cygwin*)
# Windows random device.
random_modules="w32"
;;
*)
# Build everything, allow to select at runtime.
random_modules="$auto_random_modules"
;;
esac
fi
else
if test "$random" = "auto"; then
# Build everything, allow to select at runtime.
random_modules="$auto_random_modules"
else
random_modules="$random"
fi
fi
#
# Other defines
#
if test mym4_isgit = "yes"; then
AC_DEFINE(IS_DEVELOPMENT_VERSION,1,
[Defined if this is not a regular release])
fi
AM_CONDITIONAL(CROSS_COMPILING, test x$cross_compiling = xyes)
# This is handy for debugging so the compiler doesn't rearrange
# things and eliminate variables.
AC_ARG_ENABLE(optimization,
AC_HELP_STRING([--disable-optimization],
[disable compiler optimization]),
[if test $enableval = no ; then
CFLAGS=`echo $CFLAGS | sed 's/-O[[0-9]]//'`
fi])
AC_MSG_NOTICE([checking for cc features])
# CFLAGS mangling when using gcc.
if test "$GCC" = yes; then
AC_MSG_CHECKING([if gcc supports -fno-delete-null-pointer-checks])
_gcc_cflags_save=$CFLAGS
CFLAGS="-fno-delete-null-pointer-checks"
AC_COMPILE_IFELSE([AC_LANG_PROGRAM([],[])],_gcc_wopt=yes,_gcc_wopt=no)
AC_MSG_RESULT($_gcc_wopt)
CFLAGS=$_gcc_cflags_save;
if test x"$_gcc_wopt" = xyes ; then
CFLAGS="$CFLAGS -fno-delete-null-pointer-checks"
fi
CFLAGS="$CFLAGS -Wall"
if test "$USE_MAINTAINER_MODE" = "yes"; then
CFLAGS="$CFLAGS -Wcast-align -Wshadow -Wstrict-prototypes"
CFLAGS="$CFLAGS -Wformat -Wno-format-y2k -Wformat-security"
# If -Wno-missing-field-initializers is supported we can enable a
# a bunch of really useful warnings.
AC_MSG_CHECKING([if gcc supports -Wno-missing-field-initializers])
_gcc_cflags_save=$CFLAGS
CFLAGS="-Wno-missing-field-initializers"
AC_COMPILE_IFELSE([AC_LANG_PROGRAM([],[])],_gcc_wopt=yes,_gcc_wopt=no)
AC_MSG_RESULT($_gcc_wopt)
CFLAGS=$_gcc_cflags_save;
if test x"$_gcc_wopt" = xyes ; then
CFLAGS="$CFLAGS -W -Wextra -Wbad-function-cast"
CFLAGS="$CFLAGS -Wwrite-strings"
CFLAGS="$CFLAGS -Wdeclaration-after-statement"
CFLAGS="$CFLAGS -Wno-missing-field-initializers"
CFLAGS="$CFLAGS -Wno-sign-compare"
fi
AC_MSG_CHECKING([if gcc supports -Wpointer-arith])
_gcc_cflags_save=$CFLAGS
CFLAGS="-Wpointer-arith"
AC_COMPILE_IFELSE([AC_LANG_PROGRAM([],[])],_gcc_wopt=yes,_gcc_wopt=no)
AC_MSG_RESULT($_gcc_wopt)
CFLAGS=$_gcc_cflags_save;
if test x"$_gcc_wopt" = xyes ; then
CFLAGS="$CFLAGS -Wpointer-arith"
fi
fi
fi
# Check whether as(1) supports a noeexecstack feature. This test
# includes an override option.
CL_AS_NOEXECSTACK
AC_SUBST(LIBGCRYPT_CONFIG_API_VERSION)
AC_SUBST(LIBGCRYPT_CONFIG_LIBS)
AC_SUBST(LIBGCRYPT_CONFIG_CFLAGS)
AC_SUBST(LIBGCRYPT_CONFIG_HOST)
AC_SUBST(LIBGCRYPT_THREAD_MODULES)
AC_CONFIG_COMMANDS([gcrypt-conf],[[
chmod +x src/libgcrypt-config
]],[[
prefix=$prefix
exec_prefix=$exec_prefix
libdir=$libdir
datadir=$datadir
DATADIRNAME=$DATADIRNAME
]])
#####################
#### Conclusion. ####
#####################
# Check that requested feature can actually be used and define
# ENABLE_foo_SUPPORT macros.
if test x"$aesnisupport" = xyes ; then
if test "$gcry_cv_gcc_inline_asm_ssse3" != "yes" ; then
aesnisupport="no (unsupported by compiler)"
fi
fi
if test x"$shaextsupport" = xyes ; then
if test "$gcry_cv_gcc_inline_asm_shaext" != "yes" ; then
shaextsupport="no (unsupported by compiler)"
fi
fi
if test x"$pclmulsupport" = xyes ; then
if test "$gcry_cv_gcc_inline_asm_pclmul" != "yes" ; then
pclmulsupport="no (unsupported by compiler)"
fi
fi
if test x"$sse41support" = xyes ; then
if test "$gcry_cv_gcc_inline_asm_sse41" != "yes" ; then
sse41support="no (unsupported by compiler)"
fi
fi
if test x"$avxsupport" = xyes ; then
if test "$gcry_cv_gcc_inline_asm_avx" != "yes" ; then
avxsupport="no (unsupported by compiler)"
fi
fi
if test x"$avx2support" = xyes ; then
if test "$gcry_cv_gcc_inline_asm_avx2" != "yes" ; then
avx2support="no (unsupported by compiler)"
fi
fi
if test x"$neonsupport" = xyes ; then
if test "$gcry_cv_gcc_inline_asm_neon" != "yes" ; then
if test "$gcry_cv_gcc_inline_asm_aarch64_neon" != "yes" ; then
neonsupport="no (unsupported by compiler)"
fi
fi
fi
if test x"$armcryptosupport" = xyes ; then
if test "$gcry_cv_gcc_inline_asm_aarch32_crypto" != "yes" ; then
if test "$gcry_cv_gcc_inline_asm_aarch64_crypto" != "yes" ; then
neonsupport="no (unsupported by compiler)"
fi
fi
fi
if test x"$aesnisupport" = xyes ; then
AC_DEFINE(ENABLE_AESNI_SUPPORT, 1,
[Enable support for Intel AES-NI instructions.])
fi
if test x"$shaextsupport" = xyes ; then
AC_DEFINE(ENABLE_SHAEXT_SUPPORT, 1,
[Enable support for Intel SHAEXT instructions.])
fi
if test x"$pclmulsupport" = xyes ; then
AC_DEFINE(ENABLE_PCLMUL_SUPPORT, 1,
[Enable support for Intel PCLMUL instructions.])
fi
if test x"$sse41support" = xyes ; then
AC_DEFINE(ENABLE_SSE41_SUPPORT, 1,
[Enable support for Intel SSE4.1 instructions.])
fi
if test x"$avxsupport" = xyes ; then
AC_DEFINE(ENABLE_AVX_SUPPORT,1,
[Enable support for Intel AVX instructions.])
fi
if test x"$avx2support" = xyes ; then
AC_DEFINE(ENABLE_AVX2_SUPPORT,1,
[Enable support for Intel AVX2 instructions.])
fi
if test x"$neonsupport" = xyes ; then
AC_DEFINE(ENABLE_NEON_SUPPORT,1,
[Enable support for ARM NEON instructions.])
fi
if test x"$armcryptosupport" = xyes ; then
AC_DEFINE(ENABLE_ARM_CRYPTO_SUPPORT,1,
[Enable support for ARMv8 Crypto Extension instructions.])
fi
if test x"$jentsupport" = xyes ; then
AC_DEFINE(ENABLE_JENT_SUPPORT, 1,
[Enable support for the jitter entropy collector.])
fi
if test x"$padlocksupport" = xyes ; then
AC_DEFINE(ENABLE_PADLOCK_SUPPORT, 1,
[Enable support for the PadLock engine.])
fi
if test x"$drngsupport" = xyes ; then
AC_DEFINE(ENABLE_DRNG_SUPPORT, 1,
[Enable support for Intel DRNG (RDRAND instruction).])
fi
# Define conditional sources and config.h symbols depending on the
# selected ciphers, pubkey-ciphers, digests, kdfs, and random modules.
LIST_MEMBER(arcfour, $enabled_ciphers)
if test "$found" = "1"; then
GCRYPT_CIPHERS="$GCRYPT_CIPHERS arcfour.lo"
AC_DEFINE(USE_ARCFOUR, 1, [Defined if this module should be included])
case "${host}" in
x86_64-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS arcfour-amd64.lo"
;;
esac
fi
LIST_MEMBER(blowfish, $enabled_ciphers)
if test "$found" = "1" ; then
GCRYPT_CIPHERS="$GCRYPT_CIPHERS blowfish.lo"
AC_DEFINE(USE_BLOWFISH, 1, [Defined if this module should be included])
case "${host}" in
x86_64-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS blowfish-amd64.lo"
;;
arm*-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS blowfish-arm.lo"
;;
esac
fi
LIST_MEMBER(cast5, $enabled_ciphers)
if test "$found" = "1" ; then
GCRYPT_CIPHERS="$GCRYPT_CIPHERS cast5.lo"
AC_DEFINE(USE_CAST5, 1, [Defined if this module should be included])
case "${host}" in
x86_64-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS cast5-amd64.lo"
;;
arm*-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS cast5-arm.lo"
;;
esac
fi
LIST_MEMBER(des, $enabled_ciphers)
if test "$found" = "1" ; then
GCRYPT_CIPHERS="$GCRYPT_CIPHERS des.lo"
AC_DEFINE(USE_DES, 1, [Defined if this module should be included])
case "${host}" in
x86_64-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS des-amd64.lo"
;;
esac
fi
LIST_MEMBER(aes, $enabled_ciphers)
if test "$found" = "1" ; then
GCRYPT_CIPHERS="$GCRYPT_CIPHERS rijndael.lo"
AC_DEFINE(USE_AES, 1, [Defined if this module should be included])
case "${host}" in
x86_64-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS rijndael-amd64.lo"
# Build with the SSSE3 implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS rijndael-ssse3-amd64.lo"
GCRYPT_CIPHERS="$GCRYPT_CIPHERS rijndael-ssse3-amd64-asm.lo"
;;
arm*-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS rijndael-arm.lo"
# Build with the ARMv8/AArch32 CE implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS rijndael-armv8-ce.lo"
GCRYPT_CIPHERS="$GCRYPT_CIPHERS rijndael-armv8-aarch32-ce.lo"
;;
aarch64-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS rijndael-aarch64.lo"
# Build with the ARMv8/AArch64 CE implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS rijndael-armv8-ce.lo"
GCRYPT_CIPHERS="$GCRYPT_CIPHERS rijndael-armv8-aarch64-ce.lo"
;;
esac
case "$mpi_cpu_arch" in
x86)
# Build with the AES-NI implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS rijndael-aesni.lo"
# Build with the Padlock implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS rijndael-padlock.lo"
;;
esac
fi
LIST_MEMBER(twofish, $enabled_ciphers)
if test "$found" = "1" ; then
GCRYPT_CIPHERS="$GCRYPT_CIPHERS twofish.lo"
AC_DEFINE(USE_TWOFISH, 1, [Defined if this module should be included])
case "${host}" in
x86_64-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS twofish-amd64.lo"
if test x"$avx2support" = xyes ; then
# Build with the AVX2 implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS twofish-avx2-amd64.lo"
fi
;;
arm*-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS twofish-arm.lo"
;;
aarch64-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS twofish-aarch64.lo"
;;
esac
fi
LIST_MEMBER(serpent, $enabled_ciphers)
if test "$found" = "1" ; then
GCRYPT_CIPHERS="$GCRYPT_CIPHERS serpent.lo"
AC_DEFINE(USE_SERPENT, 1, [Defined if this module should be included])
case "${host}" in
x86_64-*-*)
# Build with the SSE2 implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS serpent-sse2-amd64.lo"
;;
esac
if test x"$avx2support" = xyes ; then
# Build with the AVX2 implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS serpent-avx2-amd64.lo"
fi
if test x"$neonsupport" = xyes ; then
# Build with the NEON implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS serpent-armv7-neon.lo"
fi
fi
LIST_MEMBER(rfc2268, $enabled_ciphers)
if test "$found" = "1" ; then
GCRYPT_CIPHERS="$GCRYPT_CIPHERS rfc2268.lo"
AC_DEFINE(USE_RFC2268, 1, [Defined if this module should be included])
fi
LIST_MEMBER(seed, $enabled_ciphers)
if test "$found" = "1" ; then
GCRYPT_CIPHERS="$GCRYPT_CIPHERS seed.lo"
AC_DEFINE(USE_SEED, 1, [Defined if this module should be included])
fi
LIST_MEMBER(camellia, $enabled_ciphers)
if test "$found" = "1" ; then
GCRYPT_CIPHERS="$GCRYPT_CIPHERS camellia.lo camellia-glue.lo"
AC_DEFINE(USE_CAMELLIA, 1, [Defined if this module should be included])
case "${host}" in
arm*-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS camellia-arm.lo"
;;
aarch64-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS camellia-aarch64.lo"
;;
esac
if test x"$avxsupport" = xyes ; then
if test x"$aesnisupport" = xyes ; then
# Build with the AES-NI/AVX implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS camellia-aesni-avx-amd64.lo"
fi
fi
if test x"$avx2support" = xyes ; then
if test x"$aesnisupport" = xyes ; then
# Build with the AES-NI/AVX2 implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS camellia-aesni-avx2-amd64.lo"
fi
fi
fi
LIST_MEMBER(idea, $enabled_ciphers)
if test "$found" = "1" ; then
GCRYPT_CIPHERS="$GCRYPT_CIPHERS idea.lo"
AC_DEFINE(USE_IDEA, 1, [Defined if this module should be included])
fi
LIST_MEMBER(salsa20, $enabled_ciphers)
if test "$found" = "1" ; then
GCRYPT_CIPHERS="$GCRYPT_CIPHERS salsa20.lo"
AC_DEFINE(USE_SALSA20, 1, [Defined if this module should be included])
case "${host}" in
x86_64-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS salsa20-amd64.lo"
;;
esac
if test x"$neonsupport" = xyes ; then
# Build with the NEON implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS salsa20-armv7-neon.lo"
fi
fi
LIST_MEMBER(gost28147, $enabled_ciphers)
if test "$found" = "1" ; then
GCRYPT_CIPHERS="$GCRYPT_CIPHERS gost28147.lo"
AC_DEFINE(USE_GOST28147, 1, [Defined if this module should be included])
fi
LIST_MEMBER(chacha20, $enabled_ciphers)
if test "$found" = "1" ; then
GCRYPT_CIPHERS="$GCRYPT_CIPHERS chacha20.lo"
AC_DEFINE(USE_CHACHA20, 1, [Defined if this module should be included])
case "${host}" in
x86_64-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS chacha20-amd64-ssse3.lo"
GCRYPT_CIPHERS="$GCRYPT_CIPHERS chacha20-amd64-avx2.lo"
;;
aarch64-*-*)
# Build with the assembly implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS chacha20-aarch64.lo"
;;
esac
if test x"$neonsupport" = xyes ; then
# Build with the NEON implementation
GCRYPT_CIPHERS="$GCRYPT_CIPHERS chacha20-armv7-neon.lo"
fi
fi
LIST_MEMBER(dsa, $enabled_pubkey_ciphers)
if test "$found" = "1" ; then
GCRYPT_PUBKEY_CIPHERS="$GCRYPT_PUBKEY_CIPHERS dsa.lo"
AC_DEFINE(USE_DSA, 1, [Defined if this module should be included])
fi
LIST_MEMBER(rsa, $enabled_pubkey_ciphers)
if test "$found" = "1" ; then
GCRYPT_PUBKEY_CIPHERS="$GCRYPT_PUBKEY_CIPHERS rsa.lo"
AC_DEFINE(USE_RSA, 1, [Defined if this module should be included])
fi
LIST_MEMBER(elgamal, $enabled_pubkey_ciphers)
if test "$found" = "1" ; then
GCRYPT_PUBKEY_CIPHERS="$GCRYPT_PUBKEY_CIPHERS elgamal.lo"
AC_DEFINE(USE_ELGAMAL, 1, [Defined if this module should be included])
fi
LIST_MEMBER(ecc, $enabled_pubkey_ciphers)
if test "$found" = "1" ; then
GCRYPT_PUBKEY_CIPHERS="$GCRYPT_PUBKEY_CIPHERS \
ecc.lo ecc-curves.lo ecc-misc.lo \
ecc-ecdsa.lo ecc-eddsa.lo ecc-gost.lo"
AC_DEFINE(USE_ECC, 1, [Defined if this module should be included])
fi
LIST_MEMBER(crc, $enabled_digests)
if test "$found" = "1" ; then
GCRYPT_DIGESTS="$GCRYPT_DIGESTS crc.lo"
AC_DEFINE(USE_CRC, 1, [Defined if this module should be included])
case "${host}" in
i?86-*-* | x86_64-*-*)
# Build with the assembly implementation
GCRYPT_DIGESTS="$GCRYPT_DIGESTS crc-intel-pclmul.lo"
;;
esac
fi
LIST_MEMBER(gostr3411-94, $enabled_digests)
if test "$found" = "1" ; then
# GOST R 34.11-94 internally uses GOST 28147-89
LIST_MEMBER(gost28147, $enabled_ciphers)
if test "$found" = "1" ; then
GCRYPT_DIGESTS="$GCRYPT_DIGESTS gostr3411-94.lo"
AC_DEFINE(USE_GOST_R_3411_94, 1, [Defined if this module should be included])
fi
fi
LIST_MEMBER(stribog, $enabled_digests)
if test "$found" = "1" ; then
GCRYPT_DIGESTS="$GCRYPT_DIGESTS stribog.lo"
AC_DEFINE(USE_GOST_R_3411_12, 1, [Defined if this module should be included])
fi
LIST_MEMBER(md2, $enabled_digests)
if test "$found" = "1" ; then
GCRYPT_DIGESTS="$GCRYPT_DIGESTS md2.lo"
AC_DEFINE(USE_MD2, 1, [Defined if this module should be included])
fi
LIST_MEMBER(md4, $enabled_digests)
if test "$found" = "1" ; then
GCRYPT_DIGESTS="$GCRYPT_DIGESTS md4.lo"
AC_DEFINE(USE_MD4, 1, [Defined if this module should be included])
fi
LIST_MEMBER(md5, $enabled_digests)
if test "$found" = "1" ; then
GCRYPT_DIGESTS="$GCRYPT_DIGESTS md5.lo"
AC_DEFINE(USE_MD5, 1, [Defined if this module should be included])
fi
LIST_MEMBER(rmd160, $enabled_digests)
if test "$found" = "1" ; then
GCRYPT_DIGESTS="$GCRYPT_DIGESTS rmd160.lo"
AC_DEFINE(USE_RMD160, 1, [Defined if this module should be included])
fi
LIST_MEMBER(sha256, $enabled_digests)
if test "$found" = "1" ; then
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha256.lo"
AC_DEFINE(USE_SHA256, 1, [Defined if this module should be included])
case "${host}" in
x86_64-*-*)
# Build with the assembly implementation
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha256-ssse3-amd64.lo"
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha256-avx-amd64.lo"
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha256-avx2-bmi2-amd64.lo"
;;
arm*-*-*)
# Build with the assembly implementation
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha256-armv8-aarch32-ce.lo"
;;
aarch64-*-*)
# Build with the assembly implementation
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha256-armv8-aarch64-ce.lo"
;;
esac
case "$mpi_cpu_arch" in
x86)
# Build with the SHAEXT implementation
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha256-intel-shaext.lo"
;;
esac
fi
LIST_MEMBER(sha512, $enabled_digests)
if test "$found" = "1" ; then
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha512.lo"
AC_DEFINE(USE_SHA512, 1, [Defined if this module should be included])
case "${host}" in
x86_64-*-*)
# Build with the assembly implementation
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha512-ssse3-amd64.lo"
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha512-avx-amd64.lo"
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha512-avx2-bmi2-amd64.lo"
;;
arm*-*-*)
# Build with the assembly implementation
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha512-arm.lo"
;;
esac
if test x"$neonsupport" = xyes ; then
# Build with the NEON implementation
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha512-armv7-neon.lo"
fi
fi
LIST_MEMBER(sha3, $enabled_digests)
if test "$found" = "1" ; then
GCRYPT_DIGESTS="$GCRYPT_DIGESTS keccak.lo"
AC_DEFINE(USE_SHA3, 1, [Defined if this module should be included])
case "${host}" in
x86_64-*-*)
# Build with the assembly implementation
:
;;
esac
if test x"$neonsupport" = xyes ; then
# Build with the NEON implementation
GCRYPT_DIGESTS="$GCRYPT_DIGESTS keccak-armv7-neon.lo"
fi
fi
LIST_MEMBER(tiger, $enabled_digests)
if test "$found" = "1" ; then
GCRYPT_DIGESTS="$GCRYPT_DIGESTS tiger.lo"
AC_DEFINE(USE_TIGER, 1, [Defined if this module should be included])
fi
LIST_MEMBER(whirlpool, $enabled_digests)
if test "$found" = "1" ; then
GCRYPT_DIGESTS="$GCRYPT_DIGESTS whirlpool.lo"
AC_DEFINE(USE_WHIRLPOOL, 1, [Defined if this module should be included])
case "${host}" in
x86_64-*-*)
# Build with the assembly implementation
GCRYPT_DIGESTS="$GCRYPT_DIGESTS whirlpool-sse2-amd64.lo"
;;
esac
fi
LIST_MEMBER(blake2, $enabled_digests)
if test "$found" = "1" ; then
GCRYPT_DIGESTS="$GCRYPT_DIGESTS blake2.lo"
AC_DEFINE(USE_BLAKE2, 1, [Defined if this module should be included])
case "${host}" in
x86_64-*-*)
# Build with the assembly implementation
GCRYPT_DIGESTS="$GCRYPT_DIGESTS blake2b-amd64-avx2.lo"
GCRYPT_DIGESTS="$GCRYPT_DIGESTS blake2s-amd64-avx.lo"
;;
esac
fi
# SHA-1 needs to be included always for example because it is used by
# random-csprng.c.
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha1.lo"
AC_DEFINE(USE_SHA1, 1, [Defined if this module should be included])
case "${host}" in
x86_64-*-*)
# Build with the assembly implementation
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha1-ssse3-amd64.lo"
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha1-avx-amd64.lo"
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha1-avx-bmi2-amd64.lo"
+ GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha1-avx2-bmi2-amd64.lo"
;;
arm*-*-*)
# Build with the assembly implementation
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha1-armv7-neon.lo"
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha1-armv8-aarch32-ce.lo"
;;
aarch64-*-*)
# Build with the assembly implementation
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha1-armv8-aarch64-ce.lo"
;;
esac
case "$mpi_cpu_arch" in
x86)
# Build with the SHAEXT implementation
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sha1-intel-shaext.lo"
;;
esac
LIST_MEMBER(sm3, $enabled_digests)
if test "$found" = "1" ; then
GCRYPT_DIGESTS="$GCRYPT_DIGESTS sm3.lo"
AC_DEFINE(USE_SM3, 1, [Defined if this module should be included])
fi
LIST_MEMBER(scrypt, $enabled_kdfs)
if test "$found" = "1" ; then
GCRYPT_KDFS="$GCRYPT_KDFS scrypt.lo"
AC_DEFINE(USE_SCRYPT, 1, [Defined if this module should be included])
fi
LIST_MEMBER(linux, $random_modules)
if test "$found" = "1" ; then
GCRYPT_RANDOM="$GCRYPT_RANDOM rndlinux.lo"
AC_DEFINE(USE_RNDLINUX, 1, [Defined if the /dev/random RNG should be used.])
fi
LIST_MEMBER(unix, $random_modules)
if test "$found" = "1" ; then
GCRYPT_RANDOM="$GCRYPT_RANDOM rndunix.lo"
AC_DEFINE(USE_RNDUNIX, 1, [Defined if the default Unix RNG should be used.])
fi
LIST_MEMBER(egd, $random_modules)
if test "$found" = "1" ; then
GCRYPT_RANDOM="$GCRYPT_RANDOM rndegd.lo"
AC_DEFINE(USE_RNDEGD, 1, [Defined if the EGD based RNG should be used.])
fi
LIST_MEMBER(w32, $random_modules)
if test "$found" = "1" ; then
GCRYPT_RANDOM="$GCRYPT_RANDOM rndw32.lo"
AC_DEFINE(USE_RNDW32, 1,
[Defined if the Windows specific RNG should be used.])
fi
LIST_MEMBER(w32ce, $random_modules)
if test "$found" = "1" ; then
GCRYPT_RANDOM="$GCRYPT_RANDOM rndw32ce.lo"
AC_DEFINE(USE_RNDW32CE, 1,
[Defined if the WindowsCE specific RNG should be used.])
fi
AC_SUBST([GCRYPT_CIPHERS])
AC_SUBST([GCRYPT_PUBKEY_CIPHERS])
AC_SUBST([GCRYPT_DIGESTS])
AC_SUBST([GCRYPT_KDFS])
AC_SUBST([GCRYPT_RANDOM])
AC_SUBST(LIBGCRYPT_CIPHERS, $enabled_ciphers)
AC_SUBST(LIBGCRYPT_PUBKEY_CIPHERS, $enabled_pubkey_ciphers)
AC_SUBST(LIBGCRYPT_DIGESTS, $enabled_digests)
# For printing the configuration we need a colon separated list of
# algorithm names.
tmp=`echo "$enabled_ciphers" | tr ' ' : `
AC_DEFINE_UNQUOTED(LIBGCRYPT_CIPHERS, "$tmp",
[List of available cipher algorithms])
tmp=`echo "$enabled_pubkey_ciphers" | tr ' ' : `
AC_DEFINE_UNQUOTED(LIBGCRYPT_PUBKEY_CIPHERS, "$tmp",
[List of available public key cipher algorithms])
tmp=`echo "$enabled_digests" | tr ' ' : `
AC_DEFINE_UNQUOTED(LIBGCRYPT_DIGESTS, "$tmp",
[List of available digest algorithms])
tmp=`echo "$enabled_kdfs" | tr ' ' : `
AC_DEFINE_UNQUOTED(LIBGCRYPT_KDFS, "$tmp",
[List of available KDF algorithms])
#
# Define conditional sources depending on the used hardware platform.
# Note that all possible modules must also be listed in
# src/Makefile.am (EXTRA_libgcrypt_la_SOURCES).
#
GCRYPT_HWF_MODULES=
case "$mpi_cpu_arch" in
x86)
AC_DEFINE(HAVE_CPU_ARCH_X86, 1, [Defined for the x86 platforms])
GCRYPT_HWF_MODULES="libgcrypt_la-hwf-x86.lo"
;;
alpha)
AC_DEFINE(HAVE_CPU_ARCH_ALPHA, 1, [Defined for Alpha platforms])
;;
sparc)
AC_DEFINE(HAVE_CPU_ARCH_SPARC, 1, [Defined for SPARC platforms])
;;
mips)
AC_DEFINE(HAVE_CPU_ARCH_MIPS, 1, [Defined for MIPS platforms])
;;
m68k)
AC_DEFINE(HAVE_CPU_ARCH_M68K, 1, [Defined for M68k platforms])
;;
ppc)
AC_DEFINE(HAVE_CPU_ARCH_PPC, 1, [Defined for PPC platforms])
;;
arm)
AC_DEFINE(HAVE_CPU_ARCH_ARM, 1, [Defined for ARM platforms])
GCRYPT_HWF_MODULES="libgcrypt_la-hwf-arm.lo"
;;
aarch64)
AC_DEFINE(HAVE_CPU_ARCH_ARM, 1, [Defined for ARM AArch64 platforms])
GCRYPT_HWF_MODULES="libgcrypt_la-hwf-arm.lo"
;;
esac
AC_SUBST([GCRYPT_HWF_MODULES])
#
# Option to disable building of doc file
#
build_doc=yes
AC_ARG_ENABLE([doc], AC_HELP_STRING([--disable-doc],
[do not build the documentation]),
build_doc=$enableval, build_doc=yes)
AM_CONDITIONAL([BUILD_DOC], [test "x$build_doc" != xno])
#
# Provide information about the build.
#
BUILD_REVISION="mym4_revision"
AC_SUBST(BUILD_REVISION)
AC_DEFINE_UNQUOTED(BUILD_REVISION, "$BUILD_REVISION",
[GIT commit id revision used to build this package])
changequote(,)dnl
BUILD_VERSION=`echo "$PACKAGE_VERSION" | sed 's/\([0-9.]*\).*/\1./'`
changequote([,])dnl
BUILD_VERSION="${BUILD_VERSION}mym4_revision_dec"
BUILD_FILEVERSION=`echo "${BUILD_VERSION}" | tr . ,`
AC_SUBST(BUILD_VERSION)
AC_SUBST(BUILD_FILEVERSION)
AC_ARG_ENABLE([build-timestamp],
AC_HELP_STRING([--enable-build-timestamp],
[set an explicit build timestamp for reproducibility.
(default is the current time in ISO-8601 format)]),
[if test "$enableval" = "yes"; then
BUILD_TIMESTAMP=`date -u +%Y-%m-%dT%H:%M+0000 2>/dev/null || date`
else
BUILD_TIMESTAMP="$enableval"
fi],
[BUILD_TIMESTAMP=""])
AC_SUBST(BUILD_TIMESTAMP)
AC_DEFINE_UNQUOTED(BUILD_TIMESTAMP, "$BUILD_TIMESTAMP",
[The time this package was configured for a build])
# And create the files.
AC_CONFIG_FILES([
Makefile
m4/Makefile
compat/Makefile
mpi/Makefile
cipher/Makefile
random/Makefile
doc/Makefile
src/Makefile
src/gcrypt.h
src/libgcrypt-config
src/libgcrypt.pc
src/versioninfo.rc
tests/Makefile
])
AC_CONFIG_FILES([tests/hashtest-256g], [chmod +x tests/hashtest-256g])
AC_CONFIG_FILES([tests/basic-disable-all-hwf], [chmod +x tests/basic-disable-all-hwf])
AC_OUTPUT
detection_module="${GCRYPT_HWF_MODULES%.lo}"
test -n "$detection_module" || detection_module="none"
# Give some feedback
GCRY_MSG_SHOW([],[])
GCRY_MSG_SHOW([Libgcrypt],[v${VERSION} has been configured as follows:])
GCRY_MSG_SHOW([],[])
GCRY_MSG_SHOW([Platform: ],[$PRINTABLE_OS_NAME ($host)])
GCRY_MSG_SHOW([Hardware detection module:],[$detection_module])
GCRY_MSG_WRAP([Enabled cipher algorithms:],[$enabled_ciphers])
GCRY_MSG_WRAP([Enabled digest algorithms:],[$enabled_digests])
GCRY_MSG_WRAP([Enabled kdf algorithms: ],[$enabled_kdfs])
GCRY_MSG_WRAP([Enabled pubkey algorithms:],[$enabled_pubkey_ciphers])
GCRY_MSG_SHOW([Random number generator: ],[$random])
GCRY_MSG_SHOW([Try using jitter entropy: ],[$jentsupport])
GCRY_MSG_SHOW([Using linux capabilities: ],[$use_capabilities])
GCRY_MSG_SHOW([Try using Padlock crypto: ],[$padlocksupport])
GCRY_MSG_SHOW([Try using AES-NI crypto: ],[$aesnisupport])
GCRY_MSG_SHOW([Try using Intel SHAEXT: ],[$shaextsupport])
GCRY_MSG_SHOW([Try using Intel PCLMUL: ],[$pclmulsupport])
GCRY_MSG_SHOW([Try using Intel SSE4.1: ],[$sse41support])
GCRY_MSG_SHOW([Try using DRNG (RDRAND): ],[$drngsupport])
GCRY_MSG_SHOW([Try using Intel AVX: ],[$avxsupport])
GCRY_MSG_SHOW([Try using Intel AVX2: ],[$avx2support])
GCRY_MSG_SHOW([Try using ARM NEON: ],[$neonsupport])
GCRY_MSG_SHOW([Try using ARMv8 crypto: ],[$armcryptosupport])
GCRY_MSG_SHOW([],[])
if test "x${gpg_config_script_warn}" != x; then
cat <