/*******************************************************************************
* Copyright (c) 2013, Intel Corporation
*
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are
* met:
*
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
*
* * Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the
* distribution.
*
* * Neither the name of the Intel Corporation nor the names of its
* contributors may be used to endorse or promote products derived from
* this software without specific prior written permission.
*
*
* THIS SOFTWARE IS PROVIDED BY INTEL CORPORATION ""AS IS"" AND ANY
* EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL INTEL CORPORATION OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
* PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
* LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
* NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
* SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
********************************************************************************
*
* Intel SHA Extensions optimized implementation of a SHA-256 update function
*
* The function takes a pointer to the current hash values, a pointer to the
* input data, and a number of 64 byte blocks to process. Once all blocks have
* been processed, the digest pointer is updated with the resulting hash value.
* The function only processes complete blocks, there is no functionality to
* store partial blocks. All message padding and hash value initialization must
* be done outside the update function.
*
* The indented lines in the loop are instructions related to rounds processing.
* The non-indented lines are instructions related to the message schedule.
*
* Author: Sean Gulley <sean.m.gulley@intel.com>
* Date: July 2013
*
********************************************************************************
*
* Example complier command line:
* icc intel_sha_extensions_sha256_intrinsic.c
* gcc -msha -msse4 intel_sha_extensions_sha256_intrinsic.c
*
*******************************************************************************/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/types.h>
#include <crypto/aesni/aesni_os.h>
#include <crypto/aesni/sha_sse.h>
#include <immintrin.h>
void intel_sha256_step(uint32_t *digest, const char *data, uint32_t num_blks) {
__m128i state0, state1;
__m128i msg;
__m128i msgtmp0, msgtmp1, msgtmp2, msgtmp3;
__m128i tmp;
__m128i shuf_mask;
__m128i abef_save, cdgh_save;
// Load initial hash values
// Need to reorder these appropriately
// DCBA, HGFE -> ABEF, CDGH
tmp = _mm_loadu_si128((__m128i*) digest);
state1 = _mm_loadu_si128((__m128i*) (digest+4));
tmp = _mm_shuffle_epi32(tmp, 0xB1); // CDAB
state1 = _mm_shuffle_epi32(state1, 0x1B); // EFGH
state0 = _mm_alignr_epi8(tmp, state1, 8); // ABEF
state1 = _mm_blend_epi16(state1, tmp, 0xF0); // CDGH
shuf_mask = _mm_set_epi64x(0x0c0d0e0f08090a0bull, 0x0405060700010203ull);
while (num_blks > 0) {
// Save hash values for addition after rounds
abef_save = state0;
cdgh_save = state1;
// Rounds 0-3
msg = _mm_loadu_si128((const __m128i*) data);
msgtmp0 = _mm_shuffle_epi8(msg, shuf_mask);
msg = _mm_add_epi32(msgtmp0,
_mm_set_epi64x(0xE9B5DBA5B5C0FBCFull, 0x71374491428A2F98ull));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
// Rounds 4-7
msgtmp1 = _mm_loadu_si128((const __m128i*) (data+16));
msgtmp1 = _mm_shuffle_epi8(msgtmp1, shuf_mask);
msg = _mm_add_epi32(msgtmp1,
_mm_set_epi64x(0xAB1C5ED5923F82A4ull, 0x59F111F13956C25Bull));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
msgtmp0 = _mm_sha256msg1_epu32(msgtmp0, msgtmp1);
// Rounds 8-11
msgtmp2 = _mm_loadu_si128((const __m128i*) (data+32));
msgtmp2 = _mm_shuffle_epi8(msgtmp2, shuf_mask);
msg = _mm_add_epi32(msgtmp2,
_mm_set_epi64x(0x550C7DC3243185BEull, 0x12835B01D807AA98ull));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
msgtmp1 = _mm_sha256msg1_epu32(msgtmp1, msgtmp2);
// Rounds 12-15
msgtmp3 = _mm_loadu_si128((const __m128i*) (data+48));
msgtmp3 = _mm_shuffle_epi8(msgtmp3, shuf_mask);
msg = _mm_add_epi32(msgtmp3,
_mm_set_epi64x(0xC19BF1749BDC06A7ull, 0x80DEB1FE72BE5D74ull));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
tmp = _mm_alignr_epi8(msgtmp3, msgtmp2, 4);
msgtmp0 = _mm_add_epi32(msgtmp0, tmp);
msgtmp0 = _mm_sha256msg2_epu32(msgtmp0, msgtmp3);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
msgtmp2 = _mm_sha256msg1_epu32(msgtmp2, msgtmp3);
// Rounds 16-19
msg = _mm_add_epi32(msgtmp0,
_mm_set_epi64x(0x240CA1CC0FC19DC6ull, 0xEFBE4786E49B69C1ull));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
tmp = _mm_alignr_epi8(msgtmp0, msgtmp3, 4);
msgtmp1 = _mm_add_epi32(msgtmp1, tmp);
msgtmp1 = _mm_sha256msg2_epu32(msgtmp1, msgtmp0);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
msgtmp3 = _mm_sha256msg1_epu32(msgtmp3, msgtmp0);
// Rounds 20-23
msg = _mm_add_epi32(msgtmp1,
_mm_set_epi64x(0x76F988DA5CB0A9DCull, 0x4A7484AA2DE92C6Full));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
tmp = _mm_alignr_epi8(msgtmp1, msgtmp0, 4);
msgtmp2 = _mm_add_epi32(msgtmp2, tmp);
msgtmp2 = _mm_sha256msg2_epu32(msgtmp2, msgtmp1);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
msgtmp0 = _mm_sha256msg1_epu32(msgtmp0, msgtmp1);
// Rounds 24-27
msg = _mm_add_epi32(msgtmp2,
_mm_set_epi64x(0xBF597FC7B00327C8ull, 0xA831C66D983E5152ull));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
tmp = _mm_alignr_epi8(msgtmp2, msgtmp1, 4);
msgtmp3 = _mm_add_epi32(msgtmp3, tmp);
msgtmp3 = _mm_sha256msg2_epu32(msgtmp3, msgtmp2);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
msgtmp1 = _mm_sha256msg1_epu32(msgtmp1, msgtmp2);
// Rounds 28-31
msg = _mm_add_epi32(msgtmp3,
_mm_set_epi64x(0x1429296706CA6351ull, 0xD5A79147C6E00BF3ull));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
tmp = _mm_alignr_epi8(msgtmp3, msgtmp2, 4);
msgtmp0 = _mm_add_epi32(msgtmp0, tmp);
msgtmp0 = _mm_sha256msg2_epu32(msgtmp0, msgtmp3);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
msgtmp2 = _mm_sha256msg1_epu32(msgtmp2, msgtmp3);
// Rounds 32-35
msg = _mm_add_epi32(msgtmp0,
_mm_set_epi64x(0x53380D134D2C6DFCull, 0x2E1B213827B70A85ull));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
tmp = _mm_alignr_epi8(msgtmp0, msgtmp3, 4);
msgtmp1 = _mm_add_epi32(msgtmp1, tmp);
msgtmp1 = _mm_sha256msg2_epu32(msgtmp1, msgtmp0);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
msgtmp3 = _mm_sha256msg1_epu32(msgtmp3, msgtmp0);
// Rounds 36-39
msg = _mm_add_epi32(msgtmp1,
_mm_set_epi64x(0x92722C8581C2C92Eull, 0x766A0ABB650A7354ull));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
tmp = _mm_alignr_epi8(msgtmp1, msgtmp0, 4);
msgtmp2 = _mm_add_epi32(msgtmp2, tmp);
msgtmp2 = _mm_sha256msg2_epu32(msgtmp2, msgtmp1);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
msgtmp0 = _mm_sha256msg1_epu32(msgtmp0, msgtmp1);
// Rounds 40-43
msg = _mm_add_epi32(msgtmp2,
_mm_set_epi64x(0xC76C51A3C24B8B70ull, 0xA81A664BA2BFE8A1ull));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
tmp = _mm_alignr_epi8(msgtmp2, msgtmp1, 4);
msgtmp3 = _mm_add_epi32(msgtmp3, tmp);
msgtmp3 = _mm_sha256msg2_epu32(msgtmp3, msgtmp2);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
msgtmp1 = _mm_sha256msg1_epu32(msgtmp1, msgtmp2);
// Rounds 44-47
msg = _mm_add_epi32(msgtmp3,
_mm_set_epi64x(0x106AA070F40E3585ull, 0xD6990624D192E819ull));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
tmp = _mm_alignr_epi8(msgtmp3, msgtmp2, 4);
msgtmp0 = _mm_add_epi32(msgtmp0, tmp);
msgtmp0 = _mm_sha256msg2_epu32(msgtmp0, msgtmp3);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
msgtmp2 = _mm_sha256msg1_epu32(msgtmp2, msgtmp3);
// Rounds 48-51
msg = _mm_add_epi32(msgtmp0,
_mm_set_epi64x(0x34B0BCB52748774Cull, 0x1E376C0819A4C116ull));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
tmp = _mm_alignr_epi8(msgtmp0, msgtmp3, 4);
msgtmp1 = _mm_add_epi32(msgtmp1, tmp);
msgtmp1 = _mm_sha256msg2_epu32(msgtmp1, msgtmp0);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
msgtmp3 = _mm_sha256msg1_epu32(msgtmp3, msgtmp0);
// Rounds 52-55
msg = _mm_add_epi32(msgtmp1,
_mm_set_epi64x(0x682E6FF35B9CCA4Full, 0x4ED8AA4A391C0CB3ull));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
tmp = _mm_alignr_epi8(msgtmp1, msgtmp0, 4);
msgtmp2 = _mm_add_epi32(msgtmp2, tmp);
msgtmp2 = _mm_sha256msg2_epu32(msgtmp2, msgtmp1);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
// Rounds 56-59
msg = _mm_add_epi32(msgtmp2,
_mm_set_epi64x(0x8CC7020884C87814ull, 0x78A5636F748F82EEull));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
tmp = _mm_alignr_epi8(msgtmp2, msgtmp1, 4);
msgtmp3 = _mm_add_epi32(msgtmp3, tmp);
msgtmp3 = _mm_sha256msg2_epu32(msgtmp3, msgtmp2);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
// Rounds 60-63
msg = _mm_add_epi32(msgtmp3,
_mm_set_epi64x(0xC67178F2BEF9A3F7ull, 0xA4506CEB90BEFFFAull));
state1 = _mm_sha256rnds2_epu32(state1, state0, msg);
msg = _mm_shuffle_epi32(msg, 0x0E);
state0 = _mm_sha256rnds2_epu32(state0, state1, msg);
// Add current hash values with previously saved
state0 = _mm_add_epi32(state0, abef_save);
state1 = _mm_add_epi32(state1, cdgh_save);
data += 64;
num_blks--;
}
// Write hash values back in the correct order
tmp = _mm_shuffle_epi32(state0, 0x1B); // FEBA
state1 = _mm_shuffle_epi32(state1, 0xB1); // DCHG
state0 = _mm_blend_epi16(tmp, state1, 0xF0); // DCBA
state1 = _mm_alignr_epi8(state1, tmp, 8); // ABEF
_mm_store_si128((__m128i*) digest, state0);
_mm_store_si128((__m128i*) (digest+4), state1);
}