1 /*
   2  * Intel SHA Extensions optimized implementation of a SHA-1 update function
   3  *
   4  * This file is provided under a dual BSD/GPLv2 license.  When using or
   5  * redistributing this file, you may do so under either license.
   6  *
   7  * GPL LICENSE SUMMARY
   8  *
   9  * Copyright(c) 2015 Intel Corporation.
  10  *
  11  * This program is free software; you can redistribute it and/or modify
  12  * it under the terms of version 2 of the GNU General Public License as
  13  * published by the Free Software Foundation.
  14  *
  15  * This program is distributed in the hope that it will be useful, but
  16  * WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  18  * General Public License for more details.
  19  *
  20  * Contact Information:
  21  *      Sean Gulley <sean.m.gulley@intel.com>
  22  *      Tim Chen <tim.c.chen@linux.intel.com>
  23  *
  24  * BSD LICENSE
  25  *
  26  * Copyright(c) 2015 Intel Corporation.
  27  * Copyright (c) 2018, Joyent, Inc.
  28  *
  29  * Redistribution and use in source and binary forms, with or without
  30  * modification, are permitted provided that the following conditions
  31  * are met:
  32  *
  33  *      * Redistributions of source code must retain the above copyright
  34  *        notice, this list of conditions and the following disclaimer.
  35  *      * Redistributions in binary form must reproduce the above copyright
  36  *        notice, this list of conditions and the following disclaimer in
  37  *        the documentation and/or other materials provided with the
  38  *        distribution.
  39  *      * Neither the name of Intel Corporation nor the names of its
  40  *        contributors may be used to endorse or promote products derived
  41  *        from this software without specific prior written permission.
  42  *
  43  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
  44  * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
  45  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
  46  * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
  47  * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
  48  * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
  49  * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
  50  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
  51  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
  52  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
  53  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  54  *
  55  */
  56 
  57 /*
  58  * illumos uses this file under the terms of the BSD license.
  59  */
  60 
  61 #include <sys/asm_linkage.h>
  62 
  63 #define DIGEST_PTR      %rdi    /* 1st arg */
  64 #define DATA_PTR        %rsi    /* 2nd arg */
  65 #define NUM_BLKS        %rdx    /* 3rd arg */
  66 
  67 #define RSPSAVE         %rax
  68 
  69 /* gcc conversion */
  70 #define FRAME_SIZE      32      /* space for 2x16 bytes */
  71 
  72 #define ABCD            %xmm0
  73 #define E0              %xmm1   /* Need two E's b/c they ping pong */
  74 #define E1              %xmm2
  75 #define MSG0            %xmm3
  76 #define MSG1            %xmm4
  77 #define MSG2            %xmm5
  78 #define MSG3            %xmm6
  79 #define SHUF_MASK       %xmm7
  80 
  81 
  82 /*
  83  * Intel SHA Extensions optimized implementation of a SHA-1 update function
  84  *
  85  * The function takes a pointer to the current hash values, a pointer to the
  86  * input data, and a number of 64 byte blocks to process.  Once all blocks have
  87  * been processed, the digest pointer is  updated with the resulting hash value.
  88  * The function only processes complete blocks, there is no functionality to
  89  * store partial blocks. All message padding and hash value initialization must
  90  * be done outside the update function.
  91  *
  92  * The indented lines in the loop are instructions related to rounds processing.
  93  * The non-indented lines are instructions related to the message schedule.
  94  *
  95  * void sha1_block_data_order(uint32_t *digest, const void *data,
  96                 uint32_t numBlocks)
  97  * digest : pointer to digest
  98  * data: pointer to input data
  99  * numBlocks: Number of blocks to process
 100  */
 101 .text
 102 .align 32
 103 ENTRY_NP(sha1_block_data_order)
 104         mov             %rsp, RSPSAVE
 105         sub             $FRAME_SIZE, %rsp
 106         and             $~0xF, %rsp
 107 
 108         shl             $6, NUM_BLKS            /* convert to bytes */
 109         jz              .Ldone_hash
 110         add             DATA_PTR, NUM_BLKS      /* pointer to end of data */
 111 
 112         /* load initial hash values */
 113         pinsrd          $3, 1*16(DIGEST_PTR), E0
 114         movdqu          0*16(DIGEST_PTR), ABCD
 115         pand            UPPER_WORD_MASK(%rip), E0
 116         pshufd          $0x1B, ABCD, ABCD
 117 
 118         movdqa          PSHUFFLE_BYTE_FLIP_MASK(%rip), SHUF_MASK
 119 
 120 .Lloop0:
 121         /* Save hash values for addition after rounds */
 122         movdqa          E0, (0*16)(%rsp)
 123         movdqa          ABCD, (1*16)(%rsp)
 124 
 125         /* Rounds 0-3 */
 126         movdqu          0*16(DATA_PTR), MSG0
 127         pshufb          SHUF_MASK, MSG0
 128                 paddd           MSG0, E0
 129                 movdqa          ABCD, E1
 130                 sha1rnds4       $0, E0, ABCD
 131 
 132         /* Rounds 4-7 */
 133         movdqu          1*16(DATA_PTR), MSG1
 134         pshufb          SHUF_MASK, MSG1
 135                 sha1nexte       MSG1, E1
 136                 movdqa          ABCD, E0
 137                 sha1rnds4       $0, E1, ABCD
 138         sha1msg1        MSG1, MSG0
 139 
 140         /* Rounds 8-11 */
 141         movdqu          2*16(DATA_PTR), MSG2
 142         pshufb          SHUF_MASK, MSG2
 143                 sha1nexte       MSG2, E0
 144                 movdqa          ABCD, E1
 145                 sha1rnds4       $0, E0, ABCD
 146         sha1msg1        MSG2, MSG1
 147         pxor            MSG2, MSG0
 148 
 149         /* Rounds 12-15 */
 150         movdqu          3*16(DATA_PTR), MSG3
 151         pshufb          SHUF_MASK, MSG3
 152                 sha1nexte       MSG3, E1
 153                 movdqa          ABCD, E0
 154         sha1msg2        MSG3, MSG0
 155                 sha1rnds4       $0, E1, ABCD
 156         sha1msg1        MSG3, MSG2
 157         pxor            MSG3, MSG1
 158 
 159         /* Rounds 16-19 */
 160                 sha1nexte       MSG0, E0
 161                 movdqa          ABCD, E1
 162         sha1msg2        MSG0, MSG1
 163                 sha1rnds4       $0, E0, ABCD
 164         sha1msg1        MSG0, MSG3
 165         pxor            MSG0, MSG2
 166 
 167         /* Rounds 20-23 */
 168                 sha1nexte       MSG1, E1
 169                 movdqa          ABCD, E0
 170         sha1msg2        MSG1, MSG2
 171                 sha1rnds4       $1, E1, ABCD
 172         sha1msg1        MSG1, MSG0
 173         pxor            MSG1, MSG3
 174 
 175         /* Rounds 24-27 */
 176                 sha1nexte       MSG2, E0
 177                 movdqa          ABCD, E1
 178         sha1msg2        MSG2, MSG3
 179                 sha1rnds4       $1, E0, ABCD
 180         sha1msg1        MSG2, MSG1
 181         pxor            MSG2, MSG0
 182 
 183         /* Rounds 28-31 */
 184                 sha1nexte       MSG3, E1
 185                 movdqa          ABCD, E0
 186         sha1msg2        MSG3, MSG0
 187                 sha1rnds4       $1, E1, ABCD
 188         sha1msg1        MSG3, MSG2
 189         pxor            MSG3, MSG1
 190 
 191         /* Rounds 32-35 */
 192                 sha1nexte       MSG0, E0
 193                 movdqa          ABCD, E1
 194         sha1msg2        MSG0, MSG1
 195                 sha1rnds4       $1, E0, ABCD
 196         sha1msg1        MSG0, MSG3
 197         pxor            MSG0, MSG2
 198 
 199         /* Rounds 36-39 */
 200                 sha1nexte       MSG1, E1
 201                 movdqa          ABCD, E0
 202         sha1msg2        MSG1, MSG2
 203                 sha1rnds4       $1, E1, ABCD
 204         sha1msg1        MSG1, MSG0
 205         pxor            MSG1, MSG3
 206 
 207         /* Rounds 40-43 */
 208                 sha1nexte       MSG2, E0
 209                 movdqa          ABCD, E1
 210         sha1msg2        MSG2, MSG3
 211                 sha1rnds4       $2, E0, ABCD
 212         sha1msg1        MSG2, MSG1
 213         pxor            MSG2, MSG0
 214 
 215         /* Rounds 44-47 */
 216                 sha1nexte       MSG3, E1
 217                 movdqa          ABCD, E0
 218         sha1msg2        MSG3, MSG0
 219                 sha1rnds4       $2, E1, ABCD
 220         sha1msg1        MSG3, MSG2
 221         pxor            MSG3, MSG1
 222 
 223         /* Rounds 48-51 */
 224                 sha1nexte       MSG0, E0
 225                 movdqa          ABCD, E1
 226         sha1msg2        MSG0, MSG1
 227                 sha1rnds4       $2, E0, ABCD
 228         sha1msg1        MSG0, MSG3
 229         pxor            MSG0, MSG2
 230 
 231         /* Rounds 52-55 */
 232                 sha1nexte       MSG1, E1
 233                 movdqa          ABCD, E0
 234         sha1msg2        MSG1, MSG2
 235                 sha1rnds4       $2, E1, ABCD
 236         sha1msg1        MSG1, MSG0
 237         pxor            MSG1, MSG3
 238 
 239         /* Rounds 56-59 */
 240                 sha1nexte       MSG2, E0
 241                 movdqa          ABCD, E1
 242         sha1msg2        MSG2, MSG3
 243                 sha1rnds4       $2, E0, ABCD
 244         sha1msg1        MSG2, MSG1
 245         pxor            MSG2, MSG0
 246 
 247         /* Rounds 60-63 */
 248                 sha1nexte       MSG3, E1
 249                 movdqa          ABCD, E0
 250         sha1msg2        MSG3, MSG0
 251                 sha1rnds4       $3, E1, ABCD
 252         sha1msg1        MSG3, MSG2
 253         pxor            MSG3, MSG1
 254 
 255         /* Rounds 64-67 */
 256                 sha1nexte       MSG0, E0
 257                 movdqa          ABCD, E1
 258         sha1msg2        MSG0, MSG1
 259                 sha1rnds4       $3, E0, ABCD
 260         sha1msg1        MSG0, MSG3
 261         pxor            MSG0, MSG2
 262 
 263         /* Rounds 68-71 */
 264                 sha1nexte       MSG1, E1
 265                 movdqa          ABCD, E0
 266         sha1msg2        MSG1, MSG2
 267                 sha1rnds4       $3, E1, ABCD
 268         pxor            MSG1, MSG3
 269 
 270         /* Rounds 72-75 */
 271                 sha1nexte       MSG2, E0
 272                 movdqa          ABCD, E1
 273         sha1msg2        MSG2, MSG3
 274                 sha1rnds4       $3, E0, ABCD
 275 
 276         /* Rounds 76-79 */
 277                 sha1nexte       MSG3, E1
 278                 movdqa          ABCD, E0
 279                 sha1rnds4       $3, E1, ABCD
 280 
 281         /* Add current hash values with previously saved */
 282         sha1nexte       (0*16)(%rsp), E0
 283         paddd           (1*16)(%rsp), ABCD
 284 
 285         /* Increment data pointer and loop if more to process */
 286         add             $64, DATA_PTR
 287         cmp             NUM_BLKS, DATA_PTR
 288         jne             .Lloop0
 289 
 290         /* Write hash values back in the correct order */
 291         pshufd          $0x1B, ABCD, ABCD
 292         movdqu          ABCD, 0*16(DIGEST_PTR)
 293         pextrd          $3, E0, 1*16(DIGEST_PTR)
 294 
 295 .Ldone_hash:
 296         mov             RSPSAVE, %rsp
 297 
 298         ret
 299 SET_SIZE(sha1_block_data_order)
 300 
 301 .section        .rodata.cst16.PSHUFFLE_BYTE_FLIP_MASK, "aM", @progbits, 16
 302 .align 16
 303 PSHUFFLE_BYTE_FLIP_MASK:
 304         .octa 0x000102030405060708090a0b0c0d0e0f
 305 
 306 .section        .rodata.cst16.UPPER_WORD_MASK, "aM", @progbits, 16
 307 .align 16
 308 UPPER_WORD_MASK:
 309         .octa 0xFFFFFFFF000000000000000000000000