dect
/
linux-2.6
Archived
13
0
Fork 0
This repository has been archived on 2022-02-17. You can view files and clone it, but cannot push or open issues or pull requests.
linux-2.6/arch/x86/crypto/serpent-avx-x86_64-asm_64.S

755 lines
22 KiB
ArmAsm
Raw Normal View History

crypto: serpent - add x86_64/avx assembler implementation This patch adds a x86_64/avx assembler implementation of the Serpent block cipher. The implementation is very similar to the sse2 implementation and processes eight blocks in parallel. Because of the new non-destructive three operand syntax all move-instructions can be removed and therefore a little performance increase is provided. Patch has been tested with tcrypt and automated filesystem tests. Tcrypt benchmark results: Intel Core i5-2500 CPU (fam:6, model:42, step:7) serpent-avx-x86_64 vs. serpent-sse2-x86_64 128bit key: (lrw:256bit) (xts:256bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.03x 1.01x 1.01x 1.01x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 0.99x 1.00x 1.01x 1.00x 1.00x 256B 1.05x 1.03x 1.00x 1.02x 1.05x 1.06x 1.05x 1.02x 1.05x 1.02x 1024B 1.05x 1.02x 1.00x 1.02x 1.05x 1.06x 1.05x 1.03x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.03x 1.04x 1.02x 256bit key: (lrw:384bit) (xts:512bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.01x 1.00x 1.01x 1.01x 1.00x 1.00x 0.99x 1.03x 1.01x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 1.00x 1.02x 256B 1.05x 1.02x 1.00x 1.02x 1.05x 1.02x 1.04x 1.05x 1.05x 1.02x 1024B 1.06x 1.02x 1.00x 1.02x 1.07x 1.06x 1.05x 1.04x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.05x 1.05x 1.02x serpent-avx-x86_64 vs aes-asm (8kB block): 128bit 256bit ecb-enc 1.26x 1.73x ecb-dec 1.20x 1.64x cbc-enc 0.33x 0.45x cbc-dec 1.24x 1.67x ctr-enc 1.32x 1.76x ctr-dec 1.32x 1.76x lrw-enc 1.20x 1.60x lrw-dec 1.15x 1.54x xts-enc 1.22x 1.64x xts-dec 1.17x 1.57x Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-06-12 08:47:43 +00:00
/*
* Serpent Cipher 8-way parallel algorithm (x86_64/AVX)
*
* Copyright (C) 2012 Johannes Goetzfried
* <Johannes.Goetzfried@informatik.stud.uni-erlangen.de>
*
* Based on arch/x86/crypto/serpent-sse2-x86_64-asm_64.S by
* Copyright (C) 2011 Jussi Kivilinna <jussi.kivilinna@mbnet.fi>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program 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 General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
* USA
*
*/
#include "glue_helper-asm-avx.S"
crypto: serpent - add x86_64/avx assembler implementation This patch adds a x86_64/avx assembler implementation of the Serpent block cipher. The implementation is very similar to the sse2 implementation and processes eight blocks in parallel. Because of the new non-destructive three operand syntax all move-instructions can be removed and therefore a little performance increase is provided. Patch has been tested with tcrypt and automated filesystem tests. Tcrypt benchmark results: Intel Core i5-2500 CPU (fam:6, model:42, step:7) serpent-avx-x86_64 vs. serpent-sse2-x86_64 128bit key: (lrw:256bit) (xts:256bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.03x 1.01x 1.01x 1.01x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 0.99x 1.00x 1.01x 1.00x 1.00x 256B 1.05x 1.03x 1.00x 1.02x 1.05x 1.06x 1.05x 1.02x 1.05x 1.02x 1024B 1.05x 1.02x 1.00x 1.02x 1.05x 1.06x 1.05x 1.03x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.03x 1.04x 1.02x 256bit key: (lrw:384bit) (xts:512bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.01x 1.00x 1.01x 1.01x 1.00x 1.00x 0.99x 1.03x 1.01x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 1.00x 1.02x 256B 1.05x 1.02x 1.00x 1.02x 1.05x 1.02x 1.04x 1.05x 1.05x 1.02x 1024B 1.06x 1.02x 1.00x 1.02x 1.07x 1.06x 1.05x 1.04x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.05x 1.05x 1.02x serpent-avx-x86_64 vs aes-asm (8kB block): 128bit 256bit ecb-enc 1.26x 1.73x ecb-dec 1.20x 1.64x cbc-enc 0.33x 0.45x cbc-dec 1.24x 1.67x ctr-enc 1.32x 1.76x ctr-dec 1.32x 1.76x lrw-enc 1.20x 1.60x lrw-dec 1.15x 1.54x xts-enc 1.22x 1.64x xts-dec 1.17x 1.57x Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-06-12 08:47:43 +00:00
.file "serpent-avx-x86_64-asm_64.S"
.data
.align 16
.Lbswap128_mask:
.byte 15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5, 4, 3, 2, 1, 0
crypto: serpent - add x86_64/avx assembler implementation This patch adds a x86_64/avx assembler implementation of the Serpent block cipher. The implementation is very similar to the sse2 implementation and processes eight blocks in parallel. Because of the new non-destructive three operand syntax all move-instructions can be removed and therefore a little performance increase is provided. Patch has been tested with tcrypt and automated filesystem tests. Tcrypt benchmark results: Intel Core i5-2500 CPU (fam:6, model:42, step:7) serpent-avx-x86_64 vs. serpent-sse2-x86_64 128bit key: (lrw:256bit) (xts:256bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.03x 1.01x 1.01x 1.01x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 0.99x 1.00x 1.01x 1.00x 1.00x 256B 1.05x 1.03x 1.00x 1.02x 1.05x 1.06x 1.05x 1.02x 1.05x 1.02x 1024B 1.05x 1.02x 1.00x 1.02x 1.05x 1.06x 1.05x 1.03x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.03x 1.04x 1.02x 256bit key: (lrw:384bit) (xts:512bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.01x 1.00x 1.01x 1.01x 1.00x 1.00x 0.99x 1.03x 1.01x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 1.00x 1.02x 256B 1.05x 1.02x 1.00x 1.02x 1.05x 1.02x 1.04x 1.05x 1.05x 1.02x 1024B 1.06x 1.02x 1.00x 1.02x 1.07x 1.06x 1.05x 1.04x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.05x 1.05x 1.02x serpent-avx-x86_64 vs aes-asm (8kB block): 128bit 256bit ecb-enc 1.26x 1.73x ecb-dec 1.20x 1.64x cbc-enc 0.33x 0.45x cbc-dec 1.24x 1.67x ctr-enc 1.32x 1.76x ctr-dec 1.32x 1.76x lrw-enc 1.20x 1.60x lrw-dec 1.15x 1.54x xts-enc 1.22x 1.64x xts-dec 1.17x 1.57x Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-06-12 08:47:43 +00:00
.text
#define CTX %rdi
/**********************************************************************
8-way AVX serpent
**********************************************************************/
#define RA1 %xmm0
#define RB1 %xmm1
#define RC1 %xmm2
#define RD1 %xmm3
#define RE1 %xmm4
#define tp %xmm5
#define RA2 %xmm6
#define RB2 %xmm7
#define RC2 %xmm8
#define RD2 %xmm9
#define RE2 %xmm10
#define RNOT %xmm11
#define RK0 %xmm12
#define RK1 %xmm13
#define RK2 %xmm14
#define RK3 %xmm15
#define S0_1(x0, x1, x2, x3, x4) \
vpor x0, x3, tp; \
vpxor x3, x0, x0; \
vpxor x2, x3, x4; \
vpxor RNOT, x4, x4; \
vpxor x1, tp, x3; \
vpand x0, x1, x1; \
vpxor x4, x1, x1; \
vpxor x0, x2, x2;
#define S0_2(x0, x1, x2, x3, x4) \
vpxor x3, x0, x0; \
vpor x0, x4, x4; \
vpxor x2, x0, x0; \
vpand x1, x2, x2; \
vpxor x2, x3, x3; \
vpxor RNOT, x1, x1; \
vpxor x4, x2, x2; \
vpxor x2, x1, x1;
#define S1_1(x0, x1, x2, x3, x4) \
vpxor x0, x1, tp; \
vpxor x3, x0, x0; \
vpxor RNOT, x3, x3; \
vpand tp, x1, x4; \
vpor tp, x0, x0; \
vpxor x2, x3, x3; \
vpxor x3, x0, x0; \
vpxor x3, tp, x1;
#define S1_2(x0, x1, x2, x3, x4) \
vpxor x4, x3, x3; \
vpor x4, x1, x1; \
vpxor x2, x4, x4; \
vpand x0, x2, x2; \
vpxor x1, x2, x2; \
vpor x0, x1, x1; \
vpxor RNOT, x0, x0; \
vpxor x2, x0, x0; \
vpxor x1, x4, x4;
#define S2_1(x0, x1, x2, x3, x4) \
vpxor RNOT, x3, x3; \
vpxor x0, x1, x1; \
vpand x2, x0, tp; \
vpxor x3, tp, tp; \
vpor x0, x3, x3; \
vpxor x1, x2, x2; \
vpxor x1, x3, x3; \
vpand tp, x1, x1;
#define S2_2(x0, x1, x2, x3, x4) \
vpxor x2, tp, tp; \
vpand x3, x2, x2; \
vpor x1, x3, x3; \
vpxor RNOT, tp, tp; \
vpxor tp, x3, x3; \
vpxor tp, x0, x4; \
vpxor x2, tp, x0; \
vpor x2, x1, x1;
#define S3_1(x0, x1, x2, x3, x4) \
vpxor x3, x1, tp; \
vpor x0, x3, x3; \
vpand x0, x1, x4; \
vpxor x2, x0, x0; \
vpxor tp, x2, x2; \
vpand x3, tp, x1; \
vpxor x3, x2, x2; \
vpor x4, x0, x0; \
vpxor x3, x4, x4;
#define S3_2(x0, x1, x2, x3, x4) \
vpxor x0, x1, x1; \
vpand x3, x0, x0; \
vpand x4, x3, x3; \
vpxor x2, x3, x3; \
vpor x1, x4, x4; \
vpand x1, x2, x2; \
vpxor x3, x4, x4; \
vpxor x3, x0, x0; \
vpxor x2, x3, x3;
#define S4_1(x0, x1, x2, x3, x4) \
vpand x0, x3, tp; \
vpxor x3, x0, x0; \
vpxor x2, tp, tp; \
vpor x3, x2, x2; \
vpxor x1, x0, x0; \
vpxor tp, x3, x4; \
vpor x0, x2, x2; \
vpxor x1, x2, x2;
#define S4_2(x0, x1, x2, x3, x4) \
vpand x0, x1, x1; \
vpxor x4, x1, x1; \
vpand x2, x4, x4; \
vpxor tp, x2, x2; \
vpxor x0, x4, x4; \
vpor x1, tp, x3; \
vpxor RNOT, x1, x1; \
vpxor x0, x3, x3;
#define S5_1(x0, x1, x2, x3, x4) \
vpor x0, x1, tp; \
vpxor tp, x2, x2; \
vpxor RNOT, x3, x3; \
vpxor x0, x1, x4; \
vpxor x2, x0, x0; \
vpand x4, tp, x1; \
vpor x3, x4, x4; \
vpxor x0, x4, x4;
#define S5_2(x0, x1, x2, x3, x4) \
vpand x3, x0, x0; \
vpxor x3, x1, x1; \
vpxor x2, x3, x3; \
vpxor x1, x0, x0; \
vpand x4, x2, x2; \
vpxor x2, x1, x1; \
vpand x0, x2, x2; \
vpxor x2, x3, x3;
#define S6_1(x0, x1, x2, x3, x4) \
vpxor x0, x3, x3; \
vpxor x2, x1, tp; \
vpxor x0, x2, x2; \
vpand x3, x0, x0; \
vpor x3, tp, tp; \
vpxor RNOT, x1, x4; \
vpxor tp, x0, x0; \
vpxor x2, tp, x1;
#define S6_2(x0, x1, x2, x3, x4) \
vpxor x4, x3, x3; \
vpxor x0, x4, x4; \
vpand x0, x2, x2; \
vpxor x1, x4, x4; \
vpxor x3, x2, x2; \
vpand x1, x3, x3; \
vpxor x0, x3, x3; \
vpxor x2, x1, x1;
#define S7_1(x0, x1, x2, x3, x4) \
vpxor RNOT, x1, tp; \
vpxor RNOT, x0, x0; \
vpand x2, tp, x1; \
vpxor x3, x1, x1; \
vpor tp, x3, x3; \
vpxor x2, tp, x4; \
vpxor x3, x2, x2; \
vpxor x0, x3, x3; \
vpor x1, x0, x0;
#define S7_2(x0, x1, x2, x3, x4) \
vpand x0, x2, x2; \
vpxor x4, x0, x0; \
vpxor x3, x4, x4; \
vpand x0, x3, x3; \
vpxor x1, x4, x4; \
vpxor x4, x2, x2; \
vpxor x1, x3, x3; \
vpor x0, x4, x4; \
vpxor x1, x4, x4;
#define SI0_1(x0, x1, x2, x3, x4) \
vpxor x0, x1, x1; \
vpor x1, x3, tp; \
vpxor x1, x3, x4; \
vpxor RNOT, x0, x0; \
vpxor tp, x2, x2; \
vpxor x0, tp, x3; \
vpand x1, x0, x0; \
vpxor x2, x0, x0;
#define SI0_2(x0, x1, x2, x3, x4) \
vpand x3, x2, x2; \
vpxor x4, x3, x3; \
vpxor x3, x2, x2; \
vpxor x3, x1, x1; \
vpand x0, x3, x3; \
vpxor x0, x1, x1; \
vpxor x2, x0, x0; \
vpxor x3, x4, x4;
#define SI1_1(x0, x1, x2, x3, x4) \
vpxor x3, x1, x1; \
vpxor x2, x0, tp; \
vpxor RNOT, x2, x2; \
vpor x1, x0, x4; \
vpxor x3, x4, x4; \
vpand x1, x3, x3; \
vpxor x2, x1, x1; \
vpand x4, x2, x2;
#define SI1_2(x0, x1, x2, x3, x4) \
vpxor x1, x4, x4; \
vpor x3, x1, x1; \
vpxor tp, x3, x3; \
vpxor tp, x2, x2; \
vpor x4, tp, x0; \
vpxor x4, x2, x2; \
vpxor x0, x1, x1; \
vpxor x1, x4, x4;
#define SI2_1(x0, x1, x2, x3, x4) \
vpxor x1, x2, x2; \
vpxor RNOT, x3, tp; \
vpor x2, tp, tp; \
vpxor x3, x2, x2; \
vpxor x0, x3, x4; \
vpxor x1, tp, x3; \
vpor x2, x1, x1; \
vpxor x0, x2, x2;
#define SI2_2(x0, x1, x2, x3, x4) \
vpxor x4, x1, x1; \
vpor x3, x4, x4; \
vpxor x3, x2, x2; \
vpxor x2, x4, x4; \
vpand x1, x2, x2; \
vpxor x3, x2, x2; \
vpxor x4, x3, x3; \
vpxor x0, x4, x4;
#define SI3_1(x0, x1, x2, x3, x4) \
vpxor x1, x2, x2; \
vpand x2, x1, tp; \
vpxor x0, tp, tp; \
vpor x1, x0, x0; \
vpxor x3, x1, x4; \
vpxor x3, x0, x0; \
vpor tp, x3, x3; \
vpxor x2, tp, x1;
#define SI3_2(x0, x1, x2, x3, x4) \
vpxor x3, x1, x1; \
vpxor x2, x0, x0; \
vpxor x3, x2, x2; \
vpand x1, x3, x3; \
vpxor x0, x1, x1; \
vpand x2, x0, x0; \
vpxor x3, x4, x4; \
vpxor x0, x3, x3; \
vpxor x1, x0, x0;
#define SI4_1(x0, x1, x2, x3, x4) \
vpxor x3, x2, x2; \
vpand x1, x0, tp; \
vpxor x2, tp, tp; \
vpor x3, x2, x2; \
vpxor RNOT, x0, x4; \
vpxor tp, x1, x1; \
vpxor x2, tp, x0; \
vpand x4, x2, x2;
#define SI4_2(x0, x1, x2, x3, x4) \
vpxor x0, x2, x2; \
vpor x4, x0, x0; \
vpxor x3, x0, x0; \
vpand x2, x3, x3; \
vpxor x3, x4, x4; \
vpxor x1, x3, x3; \
vpand x0, x1, x1; \
vpxor x1, x4, x4; \
vpxor x3, x0, x0;
#define SI5_1(x0, x1, x2, x3, x4) \
vpor x2, x1, tp; \
vpxor x1, x2, x2; \
vpxor x3, tp, tp; \
vpand x1, x3, x3; \
vpxor x3, x2, x2; \
vpor x0, x3, x3; \
vpxor RNOT, x0, x0; \
vpxor x2, x3, x3; \
vpor x0, x2, x2;
#define SI5_2(x0, x1, x2, x3, x4) \
vpxor tp, x1, x4; \
vpxor x4, x2, x2; \
vpand x0, x4, x4; \
vpxor tp, x0, x0; \
vpxor x3, tp, x1; \
vpand x2, x0, x0; \
vpxor x3, x2, x2; \
vpxor x2, x0, x0; \
vpxor x4, x2, x2; \
vpxor x3, x4, x4;
#define SI6_1(x0, x1, x2, x3, x4) \
vpxor x2, x0, x0; \
vpand x3, x0, tp; \
vpxor x3, x2, x2; \
vpxor x2, tp, tp; \
vpxor x1, x3, x3; \
vpor x0, x2, x2; \
vpxor x3, x2, x2; \
vpand tp, x3, x3;
#define SI6_2(x0, x1, x2, x3, x4) \
vpxor RNOT, tp, tp; \
vpxor x1, x3, x3; \
vpand x2, x1, x1; \
vpxor tp, x0, x4; \
vpxor x4, x3, x3; \
vpxor x2, x4, x4; \
vpxor x1, tp, x0; \
vpxor x0, x2, x2;
#define SI7_1(x0, x1, x2, x3, x4) \
vpand x0, x3, tp; \
vpxor x2, x0, x0; \
vpor x3, x2, x2; \
vpxor x1, x3, x4; \
vpxor RNOT, x0, x0; \
vpor tp, x1, x1; \
vpxor x0, x4, x4; \
vpand x2, x0, x0; \
vpxor x1, x0, x0;
#define SI7_2(x0, x1, x2, x3, x4) \
vpand x2, x1, x1; \
vpxor x2, tp, x3; \
vpxor x3, x4, x4; \
vpand x3, x2, x2; \
vpor x0, x3, x3; \
vpxor x4, x1, x1; \
vpxor x4, x3, x3; \
vpand x0, x4, x4; \
vpxor x2, x4, x4;
#define get_key(i, j, t) \
vbroadcastss (4*(i)+(j))*4(CTX), t;
#define K2(x0, x1, x2, x3, x4, i) \
get_key(i, 0, RK0); \
get_key(i, 1, RK1); \
get_key(i, 2, RK2); \
get_key(i, 3, RK3); \
vpxor RK0, x0 ## 1, x0 ## 1; \
vpxor RK1, x1 ## 1, x1 ## 1; \
vpxor RK2, x2 ## 1, x2 ## 1; \
vpxor RK3, x3 ## 1, x3 ## 1; \
vpxor RK0, x0 ## 2, x0 ## 2; \
vpxor RK1, x1 ## 2, x1 ## 2; \
vpxor RK2, x2 ## 2, x2 ## 2; \
vpxor RK3, x3 ## 2, x3 ## 2;
#define LK2(x0, x1, x2, x3, x4, i) \
vpslld $13, x0 ## 1, x4 ## 1; \
vpsrld $(32 - 13), x0 ## 1, x0 ## 1; \
vpor x4 ## 1, x0 ## 1, x0 ## 1; \
vpxor x0 ## 1, x1 ## 1, x1 ## 1; \
vpslld $3, x2 ## 1, x4 ## 1; \
vpsrld $(32 - 3), x2 ## 1, x2 ## 1; \
vpor x4 ## 1, x2 ## 1, x2 ## 1; \
vpxor x2 ## 1, x1 ## 1, x1 ## 1; \
vpslld $13, x0 ## 2, x4 ## 2; \
vpsrld $(32 - 13), x0 ## 2, x0 ## 2; \
vpor x4 ## 2, x0 ## 2, x0 ## 2; \
vpxor x0 ## 2, x1 ## 2, x1 ## 2; \
vpslld $3, x2 ## 2, x4 ## 2; \
vpsrld $(32 - 3), x2 ## 2, x2 ## 2; \
vpor x4 ## 2, x2 ## 2, x2 ## 2; \
vpxor x2 ## 2, x1 ## 2, x1 ## 2; \
vpslld $1, x1 ## 1, x4 ## 1; \
vpsrld $(32 - 1), x1 ## 1, x1 ## 1; \
vpor x4 ## 1, x1 ## 1, x1 ## 1; \
vpslld $3, x0 ## 1, x4 ## 1; \
vpxor x2 ## 1, x3 ## 1, x3 ## 1; \
vpxor x4 ## 1, x3 ## 1, x3 ## 1; \
get_key(i, 1, RK1); \
vpslld $1, x1 ## 2, x4 ## 2; \
vpsrld $(32 - 1), x1 ## 2, x1 ## 2; \
vpor x4 ## 2, x1 ## 2, x1 ## 2; \
vpslld $3, x0 ## 2, x4 ## 2; \
vpxor x2 ## 2, x3 ## 2, x3 ## 2; \
vpxor x4 ## 2, x3 ## 2, x3 ## 2; \
get_key(i, 3, RK3); \
vpslld $7, x3 ## 1, x4 ## 1; \
vpsrld $(32 - 7), x3 ## 1, x3 ## 1; \
vpor x4 ## 1, x3 ## 1, x3 ## 1; \
vpslld $7, x1 ## 1, x4 ## 1; \
vpxor x1 ## 1, x0 ## 1, x0 ## 1; \
vpxor x3 ## 1, x0 ## 1, x0 ## 1; \
vpxor x3 ## 1, x2 ## 1, x2 ## 1; \
vpxor x4 ## 1, x2 ## 1, x2 ## 1; \
get_key(i, 0, RK0); \
vpslld $7, x3 ## 2, x4 ## 2; \
vpsrld $(32 - 7), x3 ## 2, x3 ## 2; \
vpor x4 ## 2, x3 ## 2, x3 ## 2; \
vpslld $7, x1 ## 2, x4 ## 2; \
vpxor x1 ## 2, x0 ## 2, x0 ## 2; \
vpxor x3 ## 2, x0 ## 2, x0 ## 2; \
vpxor x3 ## 2, x2 ## 2, x2 ## 2; \
vpxor x4 ## 2, x2 ## 2, x2 ## 2; \
get_key(i, 2, RK2); \
vpxor RK1, x1 ## 1, x1 ## 1; \
vpxor RK3, x3 ## 1, x3 ## 1; \
vpslld $5, x0 ## 1, x4 ## 1; \
vpsrld $(32 - 5), x0 ## 1, x0 ## 1; \
vpor x4 ## 1, x0 ## 1, x0 ## 1; \
vpslld $22, x2 ## 1, x4 ## 1; \
vpsrld $(32 - 22), x2 ## 1, x2 ## 1; \
vpor x4 ## 1, x2 ## 1, x2 ## 1; \
vpxor RK0, x0 ## 1, x0 ## 1; \
vpxor RK2, x2 ## 1, x2 ## 1; \
vpxor RK1, x1 ## 2, x1 ## 2; \
vpxor RK3, x3 ## 2, x3 ## 2; \
vpslld $5, x0 ## 2, x4 ## 2; \
vpsrld $(32 - 5), x0 ## 2, x0 ## 2; \
vpor x4 ## 2, x0 ## 2, x0 ## 2; \
vpslld $22, x2 ## 2, x4 ## 2; \
vpsrld $(32 - 22), x2 ## 2, x2 ## 2; \
vpor x4 ## 2, x2 ## 2, x2 ## 2; \
vpxor RK0, x0 ## 2, x0 ## 2; \
vpxor RK2, x2 ## 2, x2 ## 2;
#define KL2(x0, x1, x2, x3, x4, i) \
vpxor RK0, x0 ## 1, x0 ## 1; \
vpxor RK2, x2 ## 1, x2 ## 1; \
vpsrld $5, x0 ## 1, x4 ## 1; \
vpslld $(32 - 5), x0 ## 1, x0 ## 1; \
vpor x4 ## 1, x0 ## 1, x0 ## 1; \
vpxor RK3, x3 ## 1, x3 ## 1; \
vpxor RK1, x1 ## 1, x1 ## 1; \
vpsrld $22, x2 ## 1, x4 ## 1; \
vpslld $(32 - 22), x2 ## 1, x2 ## 1; \
vpor x4 ## 1, x2 ## 1, x2 ## 1; \
vpxor x3 ## 1, x2 ## 1, x2 ## 1; \
vpxor RK0, x0 ## 2, x0 ## 2; \
vpxor RK2, x2 ## 2, x2 ## 2; \
vpsrld $5, x0 ## 2, x4 ## 2; \
vpslld $(32 - 5), x0 ## 2, x0 ## 2; \
vpor x4 ## 2, x0 ## 2, x0 ## 2; \
vpxor RK3, x3 ## 2, x3 ## 2; \
vpxor RK1, x1 ## 2, x1 ## 2; \
vpsrld $22, x2 ## 2, x4 ## 2; \
vpslld $(32 - 22), x2 ## 2, x2 ## 2; \
vpor x4 ## 2, x2 ## 2, x2 ## 2; \
vpxor x3 ## 2, x2 ## 2, x2 ## 2; \
vpxor x3 ## 1, x0 ## 1, x0 ## 1; \
vpslld $7, x1 ## 1, x4 ## 1; \
vpxor x1 ## 1, x0 ## 1, x0 ## 1; \
vpxor x4 ## 1, x2 ## 1, x2 ## 1; \
vpsrld $1, x1 ## 1, x4 ## 1; \
vpslld $(32 - 1), x1 ## 1, x1 ## 1; \
vpor x4 ## 1, x1 ## 1, x1 ## 1; \
vpxor x3 ## 2, x0 ## 2, x0 ## 2; \
vpslld $7, x1 ## 2, x4 ## 2; \
vpxor x1 ## 2, x0 ## 2, x0 ## 2; \
vpxor x4 ## 2, x2 ## 2, x2 ## 2; \
vpsrld $1, x1 ## 2, x4 ## 2; \
vpslld $(32 - 1), x1 ## 2, x1 ## 2; \
vpor x4 ## 2, x1 ## 2, x1 ## 2; \
vpsrld $7, x3 ## 1, x4 ## 1; \
vpslld $(32 - 7), x3 ## 1, x3 ## 1; \
vpor x4 ## 1, x3 ## 1, x3 ## 1; \
vpxor x0 ## 1, x1 ## 1, x1 ## 1; \
vpslld $3, x0 ## 1, x4 ## 1; \
vpxor x4 ## 1, x3 ## 1, x3 ## 1; \
vpsrld $7, x3 ## 2, x4 ## 2; \
vpslld $(32 - 7), x3 ## 2, x3 ## 2; \
vpor x4 ## 2, x3 ## 2, x3 ## 2; \
vpxor x0 ## 2, x1 ## 2, x1 ## 2; \
vpslld $3, x0 ## 2, x4 ## 2; \
vpxor x4 ## 2, x3 ## 2, x3 ## 2; \
vpsrld $13, x0 ## 1, x4 ## 1; \
vpslld $(32 - 13), x0 ## 1, x0 ## 1; \
vpor x4 ## 1, x0 ## 1, x0 ## 1; \
vpxor x2 ## 1, x1 ## 1, x1 ## 1; \
vpxor x2 ## 1, x3 ## 1, x3 ## 1; \
vpsrld $3, x2 ## 1, x4 ## 1; \
vpslld $(32 - 3), x2 ## 1, x2 ## 1; \
vpor x4 ## 1, x2 ## 1, x2 ## 1; \
vpsrld $13, x0 ## 2, x4 ## 2; \
vpslld $(32 - 13), x0 ## 2, x0 ## 2; \
vpor x4 ## 2, x0 ## 2, x0 ## 2; \
vpxor x2 ## 2, x1 ## 2, x1 ## 2; \
vpxor x2 ## 2, x3 ## 2, x3 ## 2; \
vpsrld $3, x2 ## 2, x4 ## 2; \
vpslld $(32 - 3), x2 ## 2, x2 ## 2; \
vpor x4 ## 2, x2 ## 2, x2 ## 2;
#define S(SBOX, x0, x1, x2, x3, x4) \
SBOX ## _1(x0 ## 1, x1 ## 1, x2 ## 1, x3 ## 1, x4 ## 1); \
SBOX ## _2(x0 ## 1, x1 ## 1, x2 ## 1, x3 ## 1, x4 ## 1); \
SBOX ## _1(x0 ## 2, x1 ## 2, x2 ## 2, x3 ## 2, x4 ## 2); \
SBOX ## _2(x0 ## 2, x1 ## 2, x2 ## 2, x3 ## 2, x4 ## 2);
#define SP(SBOX, x0, x1, x2, x3, x4, i) \
get_key(i, 0, RK0); \
SBOX ## _1(x0 ## 1, x1 ## 1, x2 ## 1, x3 ## 1, x4 ## 1); \
get_key(i, 2, RK2); \
SBOX ## _2(x0 ## 1, x1 ## 1, x2 ## 1, x3 ## 1, x4 ## 1); \
get_key(i, 3, RK3); \
SBOX ## _1(x0 ## 2, x1 ## 2, x2 ## 2, x3 ## 2, x4 ## 2); \
get_key(i, 1, RK1); \
SBOX ## _2(x0 ## 2, x1 ## 2, x2 ## 2, x3 ## 2, x4 ## 2); \
#define transpose_4x4(x0, x1, x2, x3, t0, t1, t2) \
vpunpckldq x1, x0, t0; \
vpunpckhdq x1, x0, t2; \
vpunpckldq x3, x2, t1; \
vpunpckhdq x3, x2, x3; \
\
vpunpcklqdq t1, t0, x0; \
vpunpckhqdq t1, t0, x1; \
vpunpcklqdq x3, t2, x2; \
vpunpckhqdq x3, t2, x3;
#define read_blocks(x0, x1, x2, x3, t0, t1, t2) \
crypto: serpent - add x86_64/avx assembler implementation This patch adds a x86_64/avx assembler implementation of the Serpent block cipher. The implementation is very similar to the sse2 implementation and processes eight blocks in parallel. Because of the new non-destructive three operand syntax all move-instructions can be removed and therefore a little performance increase is provided. Patch has been tested with tcrypt and automated filesystem tests. Tcrypt benchmark results: Intel Core i5-2500 CPU (fam:6, model:42, step:7) serpent-avx-x86_64 vs. serpent-sse2-x86_64 128bit key: (lrw:256bit) (xts:256bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.03x 1.01x 1.01x 1.01x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 0.99x 1.00x 1.01x 1.00x 1.00x 256B 1.05x 1.03x 1.00x 1.02x 1.05x 1.06x 1.05x 1.02x 1.05x 1.02x 1024B 1.05x 1.02x 1.00x 1.02x 1.05x 1.06x 1.05x 1.03x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.03x 1.04x 1.02x 256bit key: (lrw:384bit) (xts:512bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.01x 1.00x 1.01x 1.01x 1.00x 1.00x 0.99x 1.03x 1.01x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 1.00x 1.02x 256B 1.05x 1.02x 1.00x 1.02x 1.05x 1.02x 1.04x 1.05x 1.05x 1.02x 1024B 1.06x 1.02x 1.00x 1.02x 1.07x 1.06x 1.05x 1.04x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.05x 1.05x 1.02x serpent-avx-x86_64 vs aes-asm (8kB block): 128bit 256bit ecb-enc 1.26x 1.73x ecb-dec 1.20x 1.64x cbc-enc 0.33x 0.45x cbc-dec 1.24x 1.67x ctr-enc 1.32x 1.76x ctr-dec 1.32x 1.76x lrw-enc 1.20x 1.60x lrw-dec 1.15x 1.54x xts-enc 1.22x 1.64x xts-dec 1.17x 1.57x Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-06-12 08:47:43 +00:00
transpose_4x4(x0, x1, x2, x3, t0, t1, t2)
#define write_blocks(x0, x1, x2, x3, t0, t1, t2) \
transpose_4x4(x0, x1, x2, x3, t0, t1, t2)
crypto: serpent - add x86_64/avx assembler implementation This patch adds a x86_64/avx assembler implementation of the Serpent block cipher. The implementation is very similar to the sse2 implementation and processes eight blocks in parallel. Because of the new non-destructive three operand syntax all move-instructions can be removed and therefore a little performance increase is provided. Patch has been tested with tcrypt and automated filesystem tests. Tcrypt benchmark results: Intel Core i5-2500 CPU (fam:6, model:42, step:7) serpent-avx-x86_64 vs. serpent-sse2-x86_64 128bit key: (lrw:256bit) (xts:256bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.03x 1.01x 1.01x 1.01x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 0.99x 1.00x 1.01x 1.00x 1.00x 256B 1.05x 1.03x 1.00x 1.02x 1.05x 1.06x 1.05x 1.02x 1.05x 1.02x 1024B 1.05x 1.02x 1.00x 1.02x 1.05x 1.06x 1.05x 1.03x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.03x 1.04x 1.02x 256bit key: (lrw:384bit) (xts:512bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.01x 1.00x 1.01x 1.01x 1.00x 1.00x 0.99x 1.03x 1.01x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 1.00x 1.02x 256B 1.05x 1.02x 1.00x 1.02x 1.05x 1.02x 1.04x 1.05x 1.05x 1.02x 1024B 1.06x 1.02x 1.00x 1.02x 1.07x 1.06x 1.05x 1.04x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.05x 1.05x 1.02x serpent-avx-x86_64 vs aes-asm (8kB block): 128bit 256bit ecb-enc 1.26x 1.73x ecb-dec 1.20x 1.64x cbc-enc 0.33x 0.45x cbc-dec 1.24x 1.67x ctr-enc 1.32x 1.76x ctr-dec 1.32x 1.76x lrw-enc 1.20x 1.60x lrw-dec 1.15x 1.54x xts-enc 1.22x 1.64x xts-dec 1.17x 1.57x Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-06-12 08:47:43 +00:00
.align 8
.type __serpent_enc_blk8_avx,@function;
crypto: serpent - add x86_64/avx assembler implementation This patch adds a x86_64/avx assembler implementation of the Serpent block cipher. The implementation is very similar to the sse2 implementation and processes eight blocks in parallel. Because of the new non-destructive three operand syntax all move-instructions can be removed and therefore a little performance increase is provided. Patch has been tested with tcrypt and automated filesystem tests. Tcrypt benchmark results: Intel Core i5-2500 CPU (fam:6, model:42, step:7) serpent-avx-x86_64 vs. serpent-sse2-x86_64 128bit key: (lrw:256bit) (xts:256bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.03x 1.01x 1.01x 1.01x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 0.99x 1.00x 1.01x 1.00x 1.00x 256B 1.05x 1.03x 1.00x 1.02x 1.05x 1.06x 1.05x 1.02x 1.05x 1.02x 1024B 1.05x 1.02x 1.00x 1.02x 1.05x 1.06x 1.05x 1.03x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.03x 1.04x 1.02x 256bit key: (lrw:384bit) (xts:512bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.01x 1.00x 1.01x 1.01x 1.00x 1.00x 0.99x 1.03x 1.01x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 1.00x 1.02x 256B 1.05x 1.02x 1.00x 1.02x 1.05x 1.02x 1.04x 1.05x 1.05x 1.02x 1024B 1.06x 1.02x 1.00x 1.02x 1.07x 1.06x 1.05x 1.04x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.05x 1.05x 1.02x serpent-avx-x86_64 vs aes-asm (8kB block): 128bit 256bit ecb-enc 1.26x 1.73x ecb-dec 1.20x 1.64x cbc-enc 0.33x 0.45x cbc-dec 1.24x 1.67x ctr-enc 1.32x 1.76x ctr-dec 1.32x 1.76x lrw-enc 1.20x 1.60x lrw-dec 1.15x 1.54x xts-enc 1.22x 1.64x xts-dec 1.17x 1.57x Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-06-12 08:47:43 +00:00
__serpent_enc_blk8_avx:
crypto: serpent - add x86_64/avx assembler implementation This patch adds a x86_64/avx assembler implementation of the Serpent block cipher. The implementation is very similar to the sse2 implementation and processes eight blocks in parallel. Because of the new non-destructive three operand syntax all move-instructions can be removed and therefore a little performance increase is provided. Patch has been tested with tcrypt and automated filesystem tests. Tcrypt benchmark results: Intel Core i5-2500 CPU (fam:6, model:42, step:7) serpent-avx-x86_64 vs. serpent-sse2-x86_64 128bit key: (lrw:256bit) (xts:256bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.03x 1.01x 1.01x 1.01x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 0.99x 1.00x 1.01x 1.00x 1.00x 256B 1.05x 1.03x 1.00x 1.02x 1.05x 1.06x 1.05x 1.02x 1.05x 1.02x 1024B 1.05x 1.02x 1.00x 1.02x 1.05x 1.06x 1.05x 1.03x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.03x 1.04x 1.02x 256bit key: (lrw:384bit) (xts:512bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.01x 1.00x 1.01x 1.01x 1.00x 1.00x 0.99x 1.03x 1.01x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 1.00x 1.02x 256B 1.05x 1.02x 1.00x 1.02x 1.05x 1.02x 1.04x 1.05x 1.05x 1.02x 1024B 1.06x 1.02x 1.00x 1.02x 1.07x 1.06x 1.05x 1.04x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.05x 1.05x 1.02x serpent-avx-x86_64 vs aes-asm (8kB block): 128bit 256bit ecb-enc 1.26x 1.73x ecb-dec 1.20x 1.64x cbc-enc 0.33x 0.45x cbc-dec 1.24x 1.67x ctr-enc 1.32x 1.76x ctr-dec 1.32x 1.76x lrw-enc 1.20x 1.60x lrw-dec 1.15x 1.54x xts-enc 1.22x 1.64x xts-dec 1.17x 1.57x Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-06-12 08:47:43 +00:00
/* input:
* %rdi: ctx, CTX
* RA1, RB1, RC1, RD1, RA2, RB2, RC2, RD2: blocks
* output:
* RA1, RB1, RC1, RD1, RA2, RB2, RC2, RD2: encrypted blocks
crypto: serpent - add x86_64/avx assembler implementation This patch adds a x86_64/avx assembler implementation of the Serpent block cipher. The implementation is very similar to the sse2 implementation and processes eight blocks in parallel. Because of the new non-destructive three operand syntax all move-instructions can be removed and therefore a little performance increase is provided. Patch has been tested with tcrypt and automated filesystem tests. Tcrypt benchmark results: Intel Core i5-2500 CPU (fam:6, model:42, step:7) serpent-avx-x86_64 vs. serpent-sse2-x86_64 128bit key: (lrw:256bit) (xts:256bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.03x 1.01x 1.01x 1.01x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 0.99x 1.00x 1.01x 1.00x 1.00x 256B 1.05x 1.03x 1.00x 1.02x 1.05x 1.06x 1.05x 1.02x 1.05x 1.02x 1024B 1.05x 1.02x 1.00x 1.02x 1.05x 1.06x 1.05x 1.03x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.03x 1.04x 1.02x 256bit key: (lrw:384bit) (xts:512bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.01x 1.00x 1.01x 1.01x 1.00x 1.00x 0.99x 1.03x 1.01x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 1.00x 1.02x 256B 1.05x 1.02x 1.00x 1.02x 1.05x 1.02x 1.04x 1.05x 1.05x 1.02x 1024B 1.06x 1.02x 1.00x 1.02x 1.07x 1.06x 1.05x 1.04x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.05x 1.05x 1.02x serpent-avx-x86_64 vs aes-asm (8kB block): 128bit 256bit ecb-enc 1.26x 1.73x ecb-dec 1.20x 1.64x cbc-enc 0.33x 0.45x cbc-dec 1.24x 1.67x ctr-enc 1.32x 1.76x ctr-dec 1.32x 1.76x lrw-enc 1.20x 1.60x lrw-dec 1.15x 1.54x xts-enc 1.22x 1.64x xts-dec 1.17x 1.57x Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-06-12 08:47:43 +00:00
*/
vpcmpeqd RNOT, RNOT, RNOT;
read_blocks(RA1, RB1, RC1, RD1, RK0, RK1, RK2);
read_blocks(RA2, RB2, RC2, RD2, RK0, RK1, RK2);
crypto: serpent - add x86_64/avx assembler implementation This patch adds a x86_64/avx assembler implementation of the Serpent block cipher. The implementation is very similar to the sse2 implementation and processes eight blocks in parallel. Because of the new non-destructive three operand syntax all move-instructions can be removed and therefore a little performance increase is provided. Patch has been tested with tcrypt and automated filesystem tests. Tcrypt benchmark results: Intel Core i5-2500 CPU (fam:6, model:42, step:7) serpent-avx-x86_64 vs. serpent-sse2-x86_64 128bit key: (lrw:256bit) (xts:256bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.03x 1.01x 1.01x 1.01x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 0.99x 1.00x 1.01x 1.00x 1.00x 256B 1.05x 1.03x 1.00x 1.02x 1.05x 1.06x 1.05x 1.02x 1.05x 1.02x 1024B 1.05x 1.02x 1.00x 1.02x 1.05x 1.06x 1.05x 1.03x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.03x 1.04x 1.02x 256bit key: (lrw:384bit) (xts:512bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.01x 1.00x 1.01x 1.01x 1.00x 1.00x 0.99x 1.03x 1.01x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 1.00x 1.02x 256B 1.05x 1.02x 1.00x 1.02x 1.05x 1.02x 1.04x 1.05x 1.05x 1.02x 1024B 1.06x 1.02x 1.00x 1.02x 1.07x 1.06x 1.05x 1.04x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.05x 1.05x 1.02x serpent-avx-x86_64 vs aes-asm (8kB block): 128bit 256bit ecb-enc 1.26x 1.73x ecb-dec 1.20x 1.64x cbc-enc 0.33x 0.45x cbc-dec 1.24x 1.67x ctr-enc 1.32x 1.76x ctr-dec 1.32x 1.76x lrw-enc 1.20x 1.60x lrw-dec 1.15x 1.54x xts-enc 1.22x 1.64x xts-dec 1.17x 1.57x Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-06-12 08:47:43 +00:00
K2(RA, RB, RC, RD, RE, 0);
S(S0, RA, RB, RC, RD, RE); LK2(RC, RB, RD, RA, RE, 1);
S(S1, RC, RB, RD, RA, RE); LK2(RE, RD, RA, RC, RB, 2);
S(S2, RE, RD, RA, RC, RB); LK2(RB, RD, RE, RC, RA, 3);
S(S3, RB, RD, RE, RC, RA); LK2(RC, RA, RD, RB, RE, 4);
S(S4, RC, RA, RD, RB, RE); LK2(RA, RD, RB, RE, RC, 5);
S(S5, RA, RD, RB, RE, RC); LK2(RC, RA, RD, RE, RB, 6);
S(S6, RC, RA, RD, RE, RB); LK2(RD, RB, RA, RE, RC, 7);
S(S7, RD, RB, RA, RE, RC); LK2(RC, RA, RE, RD, RB, 8);
S(S0, RC, RA, RE, RD, RB); LK2(RE, RA, RD, RC, RB, 9);
S(S1, RE, RA, RD, RC, RB); LK2(RB, RD, RC, RE, RA, 10);
S(S2, RB, RD, RC, RE, RA); LK2(RA, RD, RB, RE, RC, 11);
S(S3, RA, RD, RB, RE, RC); LK2(RE, RC, RD, RA, RB, 12);
S(S4, RE, RC, RD, RA, RB); LK2(RC, RD, RA, RB, RE, 13);
S(S5, RC, RD, RA, RB, RE); LK2(RE, RC, RD, RB, RA, 14);
S(S6, RE, RC, RD, RB, RA); LK2(RD, RA, RC, RB, RE, 15);
S(S7, RD, RA, RC, RB, RE); LK2(RE, RC, RB, RD, RA, 16);
S(S0, RE, RC, RB, RD, RA); LK2(RB, RC, RD, RE, RA, 17);
S(S1, RB, RC, RD, RE, RA); LK2(RA, RD, RE, RB, RC, 18);
S(S2, RA, RD, RE, RB, RC); LK2(RC, RD, RA, RB, RE, 19);
S(S3, RC, RD, RA, RB, RE); LK2(RB, RE, RD, RC, RA, 20);
S(S4, RB, RE, RD, RC, RA); LK2(RE, RD, RC, RA, RB, 21);
S(S5, RE, RD, RC, RA, RB); LK2(RB, RE, RD, RA, RC, 22);
S(S6, RB, RE, RD, RA, RC); LK2(RD, RC, RE, RA, RB, 23);
S(S7, RD, RC, RE, RA, RB); LK2(RB, RE, RA, RD, RC, 24);
S(S0, RB, RE, RA, RD, RC); LK2(RA, RE, RD, RB, RC, 25);
S(S1, RA, RE, RD, RB, RC); LK2(RC, RD, RB, RA, RE, 26);
S(S2, RC, RD, RB, RA, RE); LK2(RE, RD, RC, RA, RB, 27);
S(S3, RE, RD, RC, RA, RB); LK2(RA, RB, RD, RE, RC, 28);
S(S4, RA, RB, RD, RE, RC); LK2(RB, RD, RE, RC, RA, 29);
S(S5, RB, RD, RE, RC, RA); LK2(RA, RB, RD, RC, RE, 30);
S(S6, RA, RB, RD, RC, RE); LK2(RD, RE, RB, RC, RA, 31);
S(S7, RD, RE, RB, RC, RA); K2(RA, RB, RC, RD, RE, 32);
write_blocks(RA1, RB1, RC1, RD1, RK0, RK1, RK2);
write_blocks(RA2, RB2, RC2, RD2, RK0, RK1, RK2);
crypto: serpent - add x86_64/avx assembler implementation This patch adds a x86_64/avx assembler implementation of the Serpent block cipher. The implementation is very similar to the sse2 implementation and processes eight blocks in parallel. Because of the new non-destructive three operand syntax all move-instructions can be removed and therefore a little performance increase is provided. Patch has been tested with tcrypt and automated filesystem tests. Tcrypt benchmark results: Intel Core i5-2500 CPU (fam:6, model:42, step:7) serpent-avx-x86_64 vs. serpent-sse2-x86_64 128bit key: (lrw:256bit) (xts:256bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.03x 1.01x 1.01x 1.01x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 0.99x 1.00x 1.01x 1.00x 1.00x 256B 1.05x 1.03x 1.00x 1.02x 1.05x 1.06x 1.05x 1.02x 1.05x 1.02x 1024B 1.05x 1.02x 1.00x 1.02x 1.05x 1.06x 1.05x 1.03x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.03x 1.04x 1.02x 256bit key: (lrw:384bit) (xts:512bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.01x 1.00x 1.01x 1.01x 1.00x 1.00x 0.99x 1.03x 1.01x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 1.00x 1.02x 256B 1.05x 1.02x 1.00x 1.02x 1.05x 1.02x 1.04x 1.05x 1.05x 1.02x 1024B 1.06x 1.02x 1.00x 1.02x 1.07x 1.06x 1.05x 1.04x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.05x 1.05x 1.02x serpent-avx-x86_64 vs aes-asm (8kB block): 128bit 256bit ecb-enc 1.26x 1.73x ecb-dec 1.20x 1.64x cbc-enc 0.33x 0.45x cbc-dec 1.24x 1.67x ctr-enc 1.32x 1.76x ctr-dec 1.32x 1.76x lrw-enc 1.20x 1.60x lrw-dec 1.15x 1.54x xts-enc 1.22x 1.64x xts-dec 1.17x 1.57x Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-06-12 08:47:43 +00:00
ret;
.align 8
.type __serpent_dec_blk8_avx,@function;
crypto: serpent - add x86_64/avx assembler implementation This patch adds a x86_64/avx assembler implementation of the Serpent block cipher. The implementation is very similar to the sse2 implementation and processes eight blocks in parallel. Because of the new non-destructive three operand syntax all move-instructions can be removed and therefore a little performance increase is provided. Patch has been tested with tcrypt and automated filesystem tests. Tcrypt benchmark results: Intel Core i5-2500 CPU (fam:6, model:42, step:7) serpent-avx-x86_64 vs. serpent-sse2-x86_64 128bit key: (lrw:256bit) (xts:256bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.03x 1.01x 1.01x 1.01x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 0.99x 1.00x 1.01x 1.00x 1.00x 256B 1.05x 1.03x 1.00x 1.02x 1.05x 1.06x 1.05x 1.02x 1.05x 1.02x 1024B 1.05x 1.02x 1.00x 1.02x 1.05x 1.06x 1.05x 1.03x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.03x 1.04x 1.02x 256bit key: (lrw:384bit) (xts:512bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.01x 1.00x 1.01x 1.01x 1.00x 1.00x 0.99x 1.03x 1.01x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 1.00x 1.02x 256B 1.05x 1.02x 1.00x 1.02x 1.05x 1.02x 1.04x 1.05x 1.05x 1.02x 1024B 1.06x 1.02x 1.00x 1.02x 1.07x 1.06x 1.05x 1.04x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.05x 1.05x 1.02x serpent-avx-x86_64 vs aes-asm (8kB block): 128bit 256bit ecb-enc 1.26x 1.73x ecb-dec 1.20x 1.64x cbc-enc 0.33x 0.45x cbc-dec 1.24x 1.67x ctr-enc 1.32x 1.76x ctr-dec 1.32x 1.76x lrw-enc 1.20x 1.60x lrw-dec 1.15x 1.54x xts-enc 1.22x 1.64x xts-dec 1.17x 1.57x Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-06-12 08:47:43 +00:00
__serpent_dec_blk8_avx:
crypto: serpent - add x86_64/avx assembler implementation This patch adds a x86_64/avx assembler implementation of the Serpent block cipher. The implementation is very similar to the sse2 implementation and processes eight blocks in parallel. Because of the new non-destructive three operand syntax all move-instructions can be removed and therefore a little performance increase is provided. Patch has been tested with tcrypt and automated filesystem tests. Tcrypt benchmark results: Intel Core i5-2500 CPU (fam:6, model:42, step:7) serpent-avx-x86_64 vs. serpent-sse2-x86_64 128bit key: (lrw:256bit) (xts:256bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.03x 1.01x 1.01x 1.01x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 0.99x 1.00x 1.01x 1.00x 1.00x 256B 1.05x 1.03x 1.00x 1.02x 1.05x 1.06x 1.05x 1.02x 1.05x 1.02x 1024B 1.05x 1.02x 1.00x 1.02x 1.05x 1.06x 1.05x 1.03x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.03x 1.04x 1.02x 256bit key: (lrw:384bit) (xts:512bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.01x 1.00x 1.01x 1.01x 1.00x 1.00x 0.99x 1.03x 1.01x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 1.00x 1.02x 256B 1.05x 1.02x 1.00x 1.02x 1.05x 1.02x 1.04x 1.05x 1.05x 1.02x 1024B 1.06x 1.02x 1.00x 1.02x 1.07x 1.06x 1.05x 1.04x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.05x 1.05x 1.02x serpent-avx-x86_64 vs aes-asm (8kB block): 128bit 256bit ecb-enc 1.26x 1.73x ecb-dec 1.20x 1.64x cbc-enc 0.33x 0.45x cbc-dec 1.24x 1.67x ctr-enc 1.32x 1.76x ctr-dec 1.32x 1.76x lrw-enc 1.20x 1.60x lrw-dec 1.15x 1.54x xts-enc 1.22x 1.64x xts-dec 1.17x 1.57x Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-06-12 08:47:43 +00:00
/* input:
* %rdi: ctx, CTX
* RA1, RB1, RC1, RD1, RA2, RB2, RC2, RD2: encrypted blocks
* output:
* RC1, RD1, RB1, RE1, RC2, RD2, RB2, RE2: decrypted blocks
crypto: serpent - add x86_64/avx assembler implementation This patch adds a x86_64/avx assembler implementation of the Serpent block cipher. The implementation is very similar to the sse2 implementation and processes eight blocks in parallel. Because of the new non-destructive three operand syntax all move-instructions can be removed and therefore a little performance increase is provided. Patch has been tested with tcrypt and automated filesystem tests. Tcrypt benchmark results: Intel Core i5-2500 CPU (fam:6, model:42, step:7) serpent-avx-x86_64 vs. serpent-sse2-x86_64 128bit key: (lrw:256bit) (xts:256bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.03x 1.01x 1.01x 1.01x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 0.99x 1.00x 1.01x 1.00x 1.00x 256B 1.05x 1.03x 1.00x 1.02x 1.05x 1.06x 1.05x 1.02x 1.05x 1.02x 1024B 1.05x 1.02x 1.00x 1.02x 1.05x 1.06x 1.05x 1.03x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.03x 1.04x 1.02x 256bit key: (lrw:384bit) (xts:512bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.01x 1.00x 1.01x 1.01x 1.00x 1.00x 0.99x 1.03x 1.01x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 1.00x 1.02x 256B 1.05x 1.02x 1.00x 1.02x 1.05x 1.02x 1.04x 1.05x 1.05x 1.02x 1024B 1.06x 1.02x 1.00x 1.02x 1.07x 1.06x 1.05x 1.04x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.05x 1.05x 1.02x serpent-avx-x86_64 vs aes-asm (8kB block): 128bit 256bit ecb-enc 1.26x 1.73x ecb-dec 1.20x 1.64x cbc-enc 0.33x 0.45x cbc-dec 1.24x 1.67x ctr-enc 1.32x 1.76x ctr-dec 1.32x 1.76x lrw-enc 1.20x 1.60x lrw-dec 1.15x 1.54x xts-enc 1.22x 1.64x xts-dec 1.17x 1.57x Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-06-12 08:47:43 +00:00
*/
vpcmpeqd RNOT, RNOT, RNOT;
read_blocks(RA1, RB1, RC1, RD1, RK0, RK1, RK2);
read_blocks(RA2, RB2, RC2, RD2, RK0, RK1, RK2);
crypto: serpent - add x86_64/avx assembler implementation This patch adds a x86_64/avx assembler implementation of the Serpent block cipher. The implementation is very similar to the sse2 implementation and processes eight blocks in parallel. Because of the new non-destructive three operand syntax all move-instructions can be removed and therefore a little performance increase is provided. Patch has been tested with tcrypt and automated filesystem tests. Tcrypt benchmark results: Intel Core i5-2500 CPU (fam:6, model:42, step:7) serpent-avx-x86_64 vs. serpent-sse2-x86_64 128bit key: (lrw:256bit) (xts:256bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.03x 1.01x 1.01x 1.01x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 0.99x 1.00x 1.01x 1.00x 1.00x 256B 1.05x 1.03x 1.00x 1.02x 1.05x 1.06x 1.05x 1.02x 1.05x 1.02x 1024B 1.05x 1.02x 1.00x 1.02x 1.05x 1.06x 1.05x 1.03x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.03x 1.04x 1.02x 256bit key: (lrw:384bit) (xts:512bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.01x 1.00x 1.01x 1.01x 1.00x 1.00x 0.99x 1.03x 1.01x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 1.00x 1.02x 256B 1.05x 1.02x 1.00x 1.02x 1.05x 1.02x 1.04x 1.05x 1.05x 1.02x 1024B 1.06x 1.02x 1.00x 1.02x 1.07x 1.06x 1.05x 1.04x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.05x 1.05x 1.02x serpent-avx-x86_64 vs aes-asm (8kB block): 128bit 256bit ecb-enc 1.26x 1.73x ecb-dec 1.20x 1.64x cbc-enc 0.33x 0.45x cbc-dec 1.24x 1.67x ctr-enc 1.32x 1.76x ctr-dec 1.32x 1.76x lrw-enc 1.20x 1.60x lrw-dec 1.15x 1.54x xts-enc 1.22x 1.64x xts-dec 1.17x 1.57x Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-06-12 08:47:43 +00:00
K2(RA, RB, RC, RD, RE, 32);
SP(SI7, RA, RB, RC, RD, RE, 31); KL2(RB, RD, RA, RE, RC, 31);
SP(SI6, RB, RD, RA, RE, RC, 30); KL2(RA, RC, RE, RB, RD, 30);
SP(SI5, RA, RC, RE, RB, RD, 29); KL2(RC, RD, RA, RE, RB, 29);
SP(SI4, RC, RD, RA, RE, RB, 28); KL2(RC, RA, RB, RE, RD, 28);
SP(SI3, RC, RA, RB, RE, RD, 27); KL2(RB, RC, RD, RE, RA, 27);
SP(SI2, RB, RC, RD, RE, RA, 26); KL2(RC, RA, RE, RD, RB, 26);
SP(SI1, RC, RA, RE, RD, RB, 25); KL2(RB, RA, RE, RD, RC, 25);
SP(SI0, RB, RA, RE, RD, RC, 24); KL2(RE, RC, RA, RB, RD, 24);
SP(SI7, RE, RC, RA, RB, RD, 23); KL2(RC, RB, RE, RD, RA, 23);
SP(SI6, RC, RB, RE, RD, RA, 22); KL2(RE, RA, RD, RC, RB, 22);
SP(SI5, RE, RA, RD, RC, RB, 21); KL2(RA, RB, RE, RD, RC, 21);
SP(SI4, RA, RB, RE, RD, RC, 20); KL2(RA, RE, RC, RD, RB, 20);
SP(SI3, RA, RE, RC, RD, RB, 19); KL2(RC, RA, RB, RD, RE, 19);
SP(SI2, RC, RA, RB, RD, RE, 18); KL2(RA, RE, RD, RB, RC, 18);
SP(SI1, RA, RE, RD, RB, RC, 17); KL2(RC, RE, RD, RB, RA, 17);
SP(SI0, RC, RE, RD, RB, RA, 16); KL2(RD, RA, RE, RC, RB, 16);
SP(SI7, RD, RA, RE, RC, RB, 15); KL2(RA, RC, RD, RB, RE, 15);
SP(SI6, RA, RC, RD, RB, RE, 14); KL2(RD, RE, RB, RA, RC, 14);
SP(SI5, RD, RE, RB, RA, RC, 13); KL2(RE, RC, RD, RB, RA, 13);
SP(SI4, RE, RC, RD, RB, RA, 12); KL2(RE, RD, RA, RB, RC, 12);
SP(SI3, RE, RD, RA, RB, RC, 11); KL2(RA, RE, RC, RB, RD, 11);
SP(SI2, RA, RE, RC, RB, RD, 10); KL2(RE, RD, RB, RC, RA, 10);
SP(SI1, RE, RD, RB, RC, RA, 9); KL2(RA, RD, RB, RC, RE, 9);
SP(SI0, RA, RD, RB, RC, RE, 8); KL2(RB, RE, RD, RA, RC, 8);
SP(SI7, RB, RE, RD, RA, RC, 7); KL2(RE, RA, RB, RC, RD, 7);
SP(SI6, RE, RA, RB, RC, RD, 6); KL2(RB, RD, RC, RE, RA, 6);
SP(SI5, RB, RD, RC, RE, RA, 5); KL2(RD, RA, RB, RC, RE, 5);
SP(SI4, RD, RA, RB, RC, RE, 4); KL2(RD, RB, RE, RC, RA, 4);
SP(SI3, RD, RB, RE, RC, RA, 3); KL2(RE, RD, RA, RC, RB, 3);
SP(SI2, RE, RD, RA, RC, RB, 2); KL2(RD, RB, RC, RA, RE, 2);
SP(SI1, RD, RB, RC, RA, RE, 1); KL2(RE, RB, RC, RA, RD, 1);
S(SI0, RE, RB, RC, RA, RD); K2(RC, RD, RB, RE, RA, 0);
write_blocks(RC1, RD1, RB1, RE1, RK0, RK1, RK2);
write_blocks(RC2, RD2, RB2, RE2, RK0, RK1, RK2);
ret;
.align 8
.global serpent_ecb_enc_8way_avx
.type serpent_ecb_enc_8way_avx,@function;
serpent_ecb_enc_8way_avx:
/* input:
* %rdi: ctx, CTX
* %rsi: dst
* %rdx: src
*/
load_8way(%rdx, RA1, RB1, RC1, RD1, RA2, RB2, RC2, RD2);
call __serpent_enc_blk8_avx;
store_8way(%rsi, RA1, RB1, RC1, RD1, RA2, RB2, RC2, RD2);
ret;
.align 8
.global serpent_ecb_dec_8way_avx
.type serpent_ecb_dec_8way_avx,@function;
serpent_ecb_dec_8way_avx:
/* input:
* %rdi: ctx, CTX
* %rsi: dst
* %rdx: src
*/
load_8way(%rdx, RA1, RB1, RC1, RD1, RA2, RB2, RC2, RD2);
call __serpent_dec_blk8_avx;
store_8way(%rsi, RC1, RD1, RB1, RE1, RC2, RD2, RB2, RE2);
ret;
.align 8
.global serpent_cbc_dec_8way_avx
.type serpent_cbc_dec_8way_avx,@function;
serpent_cbc_dec_8way_avx:
/* input:
* %rdi: ctx, CTX
* %rsi: dst
* %rdx: src
*/
load_8way(%rdx, RA1, RB1, RC1, RD1, RA2, RB2, RC2, RD2);
call __serpent_dec_blk8_avx;
store_cbc_8way(%rdx, %rsi, RC1, RD1, RB1, RE1, RC2, RD2, RB2, RE2);
ret;
.align 8
.global serpent_ctr_8way_avx
.type serpent_ctr_8way_avx,@function;
serpent_ctr_8way_avx:
/* input:
* %rdi: ctx, CTX
* %rsi: dst
* %rdx: src
* %rcx: iv (little endian, 128bit)
*/
load_ctr_8way(%rcx, .Lbswap128_mask, RA1, RB1, RC1, RD1, RA2, RB2, RC2,
RD2, RK0, RK1, RK2);
call __serpent_enc_blk8_avx;
store_ctr_8way(%rdx, %rsi, RA1, RB1, RC1, RD1, RA2, RB2, RC2, RD2);
crypto: serpent - add x86_64/avx assembler implementation This patch adds a x86_64/avx assembler implementation of the Serpent block cipher. The implementation is very similar to the sse2 implementation and processes eight blocks in parallel. Because of the new non-destructive three operand syntax all move-instructions can be removed and therefore a little performance increase is provided. Patch has been tested with tcrypt and automated filesystem tests. Tcrypt benchmark results: Intel Core i5-2500 CPU (fam:6, model:42, step:7) serpent-avx-x86_64 vs. serpent-sse2-x86_64 128bit key: (lrw:256bit) (xts:256bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.03x 1.01x 1.01x 1.01x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 0.99x 1.00x 1.01x 1.00x 1.00x 256B 1.05x 1.03x 1.00x 1.02x 1.05x 1.06x 1.05x 1.02x 1.05x 1.02x 1024B 1.05x 1.02x 1.00x 1.02x 1.05x 1.06x 1.05x 1.03x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.03x 1.04x 1.02x 256bit key: (lrw:384bit) (xts:512bit) size ecb-enc ecb-dec cbc-enc cbc-dec ctr-enc ctr-dec lrw-enc lrw-dec xts-enc xts-dec 16B 1.01x 1.00x 1.01x 1.01x 1.00x 1.00x 0.99x 1.03x 1.01x 1.01x 64B 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.00x 1.01x 1.00x 1.02x 256B 1.05x 1.02x 1.00x 1.02x 1.05x 1.02x 1.04x 1.05x 1.05x 1.02x 1024B 1.06x 1.02x 1.00x 1.02x 1.07x 1.06x 1.05x 1.04x 1.05x 1.02x 8192B 1.05x 1.02x 1.00x 1.02x 1.06x 1.06x 1.04x 1.05x 1.05x 1.02x serpent-avx-x86_64 vs aes-asm (8kB block): 128bit 256bit ecb-enc 1.26x 1.73x ecb-dec 1.20x 1.64x cbc-enc 0.33x 0.45x cbc-dec 1.24x 1.67x ctr-enc 1.32x 1.76x ctr-dec 1.32x 1.76x lrw-enc 1.20x 1.60x lrw-dec 1.15x 1.54x xts-enc 1.22x 1.64x xts-dec 1.17x 1.57x Signed-off-by: Johannes Goetzfried <Johannes.Goetzfried@informatik.stud.uni-erlangen.de> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2012-06-12 08:47:43 +00:00
ret;