2 * Serpent Cipher 4-way parallel algorithm (i586/SSE2)
4 * Copyright (C) 2011 Jussi Kivilinna <jussi.kivilinna@mbnet.fi>
6 * Based on crypto/serpent.c by
7 * Copyright (C) 2002 Dag Arne Osvik <osvik@ii.uib.no>
8 * 2003 Herbert Valerio Riedel <hvr@gnu.org>
10 * This program is free software; you can redistribute it and/or modify
11 * it under the terms of the GNU General Public License as published by
12 * the Free Software Foundation; either version 2 of the License, or
13 * (at your option) any later version.
15 * This program is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 * GNU General Public License for more details.
20 * You should have received a copy of the GNU General Public License
21 * along with this program; if not, write to the Free Software
22 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307
27 #include <linux/linkage.h>
29 .file "serpent-sse2-i586-asm_32.S"
37 /**********************************************************************
39 **********************************************************************/
53 #define get_key(i, j, t) \
54 movd (4*(i)+(j))*4(CTX), t; \
57 #define K(x0, x1, x2, x3, x4, i) \
67 #define LK(x0, x1, x2, x3, x4, i) \
70 psrld $(32 - 13), x4; \
75 psrld $(32 - 3), x4; \
80 psrld $(32 - 1), x4; \
88 psrld $(32 - 7), x4; \
102 psrld $(32 - 5), x4; \
106 psrld $(32 - 22), x4; \
108 get_key(i, 0, RT0); \
110 get_key(i, 2, RT0); \
113 #define KL(x0, x1, x2, x3, x4, i) \
114 K(x0, x1, x2, x3, x4, i); \
117 pslld $(32 - 5), x4; \
121 pslld $(32 - 22), x4; \
131 pslld $(32 - 1), x4; \
135 pslld $(32 - 7), x4; \
143 pslld $(32 - 13), x4; \
149 pslld $(32 - 3), x4; \
152 #define S0(x0, x1, x2, x3, x4) \
171 #define S1(x0, x1, x2, x3, x4) \
191 #define S2(x0, x1, x2, x3, x4) \
210 #define S3(x0, x1, x2, x3, x4) \
231 #define S4(x0, x1, x2, x3, x4) \
250 #define S5(x0, x1, x2, x3, x4) \
269 #define S6(x0, x1, x2, x3, x4) \
288 #define S7(x0, x1, x2, x3, x4) \
309 #define SI0(x0, x1, x2, x3, x4) \
328 #define SI1(x0, x1, x2, x3, x4) \
347 #define SI2(x0, x1, x2, x3, x4) \
366 #define SI3(x0, x1, x2, x3, x4) \
386 #define SI4(x0, x1, x2, x3, x4) \
406 #define SI5(x0, x1, x2, x3, x4) \
428 #define SI6(x0, x1, x2, x3, x4) \
447 #define SI7(x0, x1, x2, x3, x4) \
468 #define transpose_4x4(x0, x1, x2, x3, t0, t1, t2) \
483 #define read_blocks(in, x0, x1, x2, x3, t0, t1, t2) \
484 movdqu (0*4*4)(in), x0; \
485 movdqu (1*4*4)(in), x1; \
486 movdqu (2*4*4)(in), x2; \
487 movdqu (3*4*4)(in), x3; \
489 transpose_4x4(x0, x1, x2, x3, t0, t1, t2)
491 #define write_blocks(out, x0, x1, x2, x3, t0, t1, t2) \
492 transpose_4x4(x0, x1, x2, x3, t0, t1, t2) \
494 movdqu x0, (0*4*4)(out); \
495 movdqu x1, (1*4*4)(out); \
496 movdqu x2, (2*4*4)(out); \
497 movdqu x3, (3*4*4)(out);
499 #define xor_blocks(out, x0, x1, x2, x3, t0, t1, t2) \
500 transpose_4x4(x0, x1, x2, x3, t0, t1, t2) \
502 movdqu (0*4*4)(out), t0; \
504 movdqu x0, (0*4*4)(out); \
505 movdqu (1*4*4)(out), t0; \
507 movdqu x1, (1*4*4)(out); \
508 movdqu (2*4*4)(out), t0; \
510 movdqu x2, (2*4*4)(out); \
511 movdqu (3*4*4)(out), t0; \
513 movdqu x3, (3*4*4)(out);
515 ENTRY(__serpent_enc_blk_4way)
517 * arg_ctx(%esp): ctx, CTX
520 * arg_xor(%esp): bool, if true: xor output
525 movl arg_ctx(%esp), CTX;
527 movl arg_src(%esp), %eax;
528 read_blocks(%eax, RA, RB, RC, RD, RT0, RT1, RE);
530 K(RA, RB, RC, RD, RE, 0);
531 S0(RA, RB, RC, RD, RE); LK(RC, RB, RD, RA, RE, 1);
532 S1(RC, RB, RD, RA, RE); LK(RE, RD, RA, RC, RB, 2);
533 S2(RE, RD, RA, RC, RB); LK(RB, RD, RE, RC, RA, 3);
534 S3(RB, RD, RE, RC, RA); LK(RC, RA, RD, RB, RE, 4);
535 S4(RC, RA, RD, RB, RE); LK(RA, RD, RB, RE, RC, 5);
536 S5(RA, RD, RB, RE, RC); LK(RC, RA, RD, RE, RB, 6);
537 S6(RC, RA, RD, RE, RB); LK(RD, RB, RA, RE, RC, 7);
538 S7(RD, RB, RA, RE, RC); LK(RC, RA, RE, RD, RB, 8);
539 S0(RC, RA, RE, RD, RB); LK(RE, RA, RD, RC, RB, 9);
540 S1(RE, RA, RD, RC, RB); LK(RB, RD, RC, RE, RA, 10);
541 S2(RB, RD, RC, RE, RA); LK(RA, RD, RB, RE, RC, 11);
542 S3(RA, RD, RB, RE, RC); LK(RE, RC, RD, RA, RB, 12);
543 S4(RE, RC, RD, RA, RB); LK(RC, RD, RA, RB, RE, 13);
544 S5(RC, RD, RA, RB, RE); LK(RE, RC, RD, RB, RA, 14);
545 S6(RE, RC, RD, RB, RA); LK(RD, RA, RC, RB, RE, 15);
546 S7(RD, RA, RC, RB, RE); LK(RE, RC, RB, RD, RA, 16);
547 S0(RE, RC, RB, RD, RA); LK(RB, RC, RD, RE, RA, 17);
548 S1(RB, RC, RD, RE, RA); LK(RA, RD, RE, RB, RC, 18);
549 S2(RA, RD, RE, RB, RC); LK(RC, RD, RA, RB, RE, 19);
550 S3(RC, RD, RA, RB, RE); LK(RB, RE, RD, RC, RA, 20);
551 S4(RB, RE, RD, RC, RA); LK(RE, RD, RC, RA, RB, 21);
552 S5(RE, RD, RC, RA, RB); LK(RB, RE, RD, RA, RC, 22);
553 S6(RB, RE, RD, RA, RC); LK(RD, RC, RE, RA, RB, 23);
554 S7(RD, RC, RE, RA, RB); LK(RB, RE, RA, RD, RC, 24);
555 S0(RB, RE, RA, RD, RC); LK(RA, RE, RD, RB, RC, 25);
556 S1(RA, RE, RD, RB, RC); LK(RC, RD, RB, RA, RE, 26);
557 S2(RC, RD, RB, RA, RE); LK(RE, RD, RC, RA, RB, 27);
558 S3(RE, RD, RC, RA, RB); LK(RA, RB, RD, RE, RC, 28);
559 S4(RA, RB, RD, RE, RC); LK(RB, RD, RE, RC, RA, 29);
560 S5(RB, RD, RE, RC, RA); LK(RA, RB, RD, RC, RE, 30);
561 S6(RA, RB, RD, RC, RE); LK(RD, RE, RB, RC, RA, 31);
562 S7(RD, RE, RB, RC, RA); K(RA, RB, RC, RD, RE, 32);
564 movl arg_dst(%esp), %eax;
566 cmpb $0, arg_xor(%esp);
569 write_blocks(%eax, RA, RB, RC, RD, RT0, RT1, RE);
574 xor_blocks(%eax, RA, RB, RC, RD, RT0, RT1, RE);
577 ENDPROC(__serpent_enc_blk_4way)
579 ENTRY(serpent_dec_blk_4way)
581 * arg_ctx(%esp): ctx, CTX
588 movl arg_ctx(%esp), CTX;
590 movl arg_src(%esp), %eax;
591 read_blocks(%eax, RA, RB, RC, RD, RT0, RT1, RE);
593 K(RA, RB, RC, RD, RE, 32);
594 SI7(RA, RB, RC, RD, RE); KL(RB, RD, RA, RE, RC, 31);
595 SI6(RB, RD, RA, RE, RC); KL(RA, RC, RE, RB, RD, 30);
596 SI5(RA, RC, RE, RB, RD); KL(RC, RD, RA, RE, RB, 29);
597 SI4(RC, RD, RA, RE, RB); KL(RC, RA, RB, RE, RD, 28);
598 SI3(RC, RA, RB, RE, RD); KL(RB, RC, RD, RE, RA, 27);
599 SI2(RB, RC, RD, RE, RA); KL(RC, RA, RE, RD, RB, 26);
600 SI1(RC, RA, RE, RD, RB); KL(RB, RA, RE, RD, RC, 25);
601 SI0(RB, RA, RE, RD, RC); KL(RE, RC, RA, RB, RD, 24);
602 SI7(RE, RC, RA, RB, RD); KL(RC, RB, RE, RD, RA, 23);
603 SI6(RC, RB, RE, RD, RA); KL(RE, RA, RD, RC, RB, 22);
604 SI5(RE, RA, RD, RC, RB); KL(RA, RB, RE, RD, RC, 21);
605 SI4(RA, RB, RE, RD, RC); KL(RA, RE, RC, RD, RB, 20);
606 SI3(RA, RE, RC, RD, RB); KL(RC, RA, RB, RD, RE, 19);
607 SI2(RC, RA, RB, RD, RE); KL(RA, RE, RD, RB, RC, 18);
608 SI1(RA, RE, RD, RB, RC); KL(RC, RE, RD, RB, RA, 17);
609 SI0(RC, RE, RD, RB, RA); KL(RD, RA, RE, RC, RB, 16);
610 SI7(RD, RA, RE, RC, RB); KL(RA, RC, RD, RB, RE, 15);
611 SI6(RA, RC, RD, RB, RE); KL(RD, RE, RB, RA, RC, 14);
612 SI5(RD, RE, RB, RA, RC); KL(RE, RC, RD, RB, RA, 13);
613 SI4(RE, RC, RD, RB, RA); KL(RE, RD, RA, RB, RC, 12);
614 SI3(RE, RD, RA, RB, RC); KL(RA, RE, RC, RB, RD, 11);
615 SI2(RA, RE, RC, RB, RD); KL(RE, RD, RB, RC, RA, 10);
616 SI1(RE, RD, RB, RC, RA); KL(RA, RD, RB, RC, RE, 9);
617 SI0(RA, RD, RB, RC, RE); KL(RB, RE, RD, RA, RC, 8);
618 SI7(RB, RE, RD, RA, RC); KL(RE, RA, RB, RC, RD, 7);
619 SI6(RE, RA, RB, RC, RD); KL(RB, RD, RC, RE, RA, 6);
620 SI5(RB, RD, RC, RE, RA); KL(RD, RA, RB, RC, RE, 5);
621 SI4(RD, RA, RB, RC, RE); KL(RD, RB, RE, RC, RA, 4);
622 SI3(RD, RB, RE, RC, RA); KL(RE, RD, RA, RC, RB, 3);
623 SI2(RE, RD, RA, RC, RB); KL(RD, RB, RC, RA, RE, 2);
624 SI1(RD, RB, RC, RA, RE); KL(RE, RB, RC, RA, RD, 1);
625 SI0(RE, RB, RC, RA, RD); K(RC, RD, RB, RE, RA, 0);
627 movl arg_dst(%esp), %eax;
628 write_blocks(%eax, RC, RD, RB, RE, RT0, RT1, RA);
631 ENDPROC(serpent_dec_blk_4way)