1/* 2 * Buffer submit code for multi buffer SHA256 algorithm 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) 2016 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 * Megha Dey <megha.dey@linux.intel.com> 22 * 23 * BSD LICENSE 24 * 25 * Copyright(c) 2016 Intel Corporation. 26 * 27 * Redistribution and use in source and binary forms, with or without 28 * modification, are permitted provided that the following conditions 29 * are met: 30 * 31 * * Redistributions of source code must retain the above copyright 32 * notice, this list of conditions and the following disclaimer. 33 * * Redistributions in binary form must reproduce the above copyright 34 * notice, this list of conditions and the following disclaimer in 35 * the documentation and/or other materials provided with the 36 * distribution. 37 * * Neither the name of Intel Corporation nor the names of its 38 * contributors may be used to endorse or promote products derived 39 * from this software without specific prior written permission. 40 * 41 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS 42 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT 43 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR 44 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT 45 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, 46 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT 47 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 48 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 49 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 50 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE 51 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 52 */ 53 54#include <linux/linkage.h> 55#include <asm/frame.h> 56#include "sha256_mb_mgr_datastruct.S" 57 58.extern sha256_x8_avx2 59 60# LINUX register definitions 61arg1 = %rdi 62arg2 = %rsi 63size_offset = %rcx 64tmp2 = %rcx 65extra_blocks = %rdx 66 67# Common definitions 68#define state arg1 69#define job %rsi 70#define len2 arg2 71#define p2 arg2 72 73# idx must be a register not clobberred by sha1_x8_avx2 74idx = %r8 75DWORD_idx = %r8d 76last_len = %r8 77 78p = %r11 79start_offset = %r11 80 81unused_lanes = %rbx 82BYTE_unused_lanes = %bl 83 84job_rax = %rax 85len = %rax 86DWORD_len = %eax 87 88lane = %r12 89tmp3 = %r12 90 91tmp = %r9 92DWORD_tmp = %r9d 93 94lane_data = %r10 95 96# JOB* sha256_mb_mgr_submit_avx2(MB_MGR *state, JOB_SHA256 *job) 97# arg 1 : rcx : state 98# arg 2 : rdx : job 99ENTRY(sha256_mb_mgr_submit_avx2) 100 FRAME_BEGIN 101 push %rbx 102 push %r12 103 104 mov _unused_lanes(state), unused_lanes 105 mov unused_lanes, lane 106 and $0xF, lane 107 shr $4, unused_lanes 108 imul $_LANE_DATA_size, lane, lane_data 109 movl $STS_BEING_PROCESSED, _status(job) 110 lea _ldata(state, lane_data), lane_data 111 mov unused_lanes, _unused_lanes(state) 112 movl _len(job), DWORD_len 113 114 mov job, _job_in_lane(lane_data) 115 shl $4, len 116 or lane, len 117 118 movl DWORD_len, _lens(state , lane, 4) 119 120 # Load digest words from result_digest 121 vmovdqu _result_digest(job), %xmm0 122 vmovdqu _result_digest+1*16(job), %xmm1 123 vmovd %xmm0, _args_digest(state, lane, 4) 124 vpextrd $1, %xmm0, _args_digest+1*32(state , lane, 4) 125 vpextrd $2, %xmm0, _args_digest+2*32(state , lane, 4) 126 vpextrd $3, %xmm0, _args_digest+3*32(state , lane, 4) 127 vmovd %xmm1, _args_digest+4*32(state , lane, 4) 128 129 vpextrd $1, %xmm1, _args_digest+5*32(state , lane, 4) 130 vpextrd $2, %xmm1, _args_digest+6*32(state , lane, 4) 131 vpextrd $3, %xmm1, _args_digest+7*32(state , lane, 4) 132 133 mov _buffer(job), p 134 mov p, _args_data_ptr(state, lane, 8) 135 136 cmp $0xF, unused_lanes 137 jne return_null 138 139start_loop: 140 # Find min length 141 vmovdqa _lens(state), %xmm0 142 vmovdqa _lens+1*16(state), %xmm1 143 144 vpminud %xmm1, %xmm0, %xmm2 # xmm2 has {D,C,B,A} 145 vpalignr $8, %xmm2, %xmm3, %xmm3 # xmm3 has {x,x,D,C} 146 vpminud %xmm3, %xmm2, %xmm2 # xmm2 has {x,x,E,F} 147 vpalignr $4, %xmm2, %xmm3, %xmm3 # xmm3 has {x,x,x,E} 148 vpminud %xmm3, %xmm2, %xmm2 # xmm2 has min val in low dword 149 150 vmovd %xmm2, DWORD_idx 151 mov idx, len2 152 and $0xF, idx 153 shr $4, len2 154 jz len_is_0 155 156 vpand clear_low_nibble(%rip), %xmm2, %xmm2 157 vpshufd $0, %xmm2, %xmm2 158 159 vpsubd %xmm2, %xmm0, %xmm0 160 vpsubd %xmm2, %xmm1, %xmm1 161 162 vmovdqa %xmm0, _lens + 0*16(state) 163 vmovdqa %xmm1, _lens + 1*16(state) 164 165 # "state" and "args" are the same address, arg1 166 # len is arg2 167 call sha256_x8_avx2 168 169 # state and idx are intact 170 171len_is_0: 172 # process completed job "idx" 173 imul $_LANE_DATA_size, idx, lane_data 174 lea _ldata(state, lane_data), lane_data 175 176 mov _job_in_lane(lane_data), job_rax 177 mov _unused_lanes(state), unused_lanes 178 movq $0, _job_in_lane(lane_data) 179 movl $STS_COMPLETED, _status(job_rax) 180 shl $4, unused_lanes 181 or idx, unused_lanes 182 mov unused_lanes, _unused_lanes(state) 183 184 movl $0xFFFFFFFF, _lens(state,idx,4) 185 186 vmovd _args_digest(state, idx, 4), %xmm0 187 vpinsrd $1, _args_digest+1*32(state , idx, 4), %xmm0, %xmm0 188 vpinsrd $2, _args_digest+2*32(state , idx, 4), %xmm0, %xmm0 189 vpinsrd $3, _args_digest+3*32(state , idx, 4), %xmm0, %xmm0 190 vmovd _args_digest+4*32(state, idx, 4), %xmm1 191 192 vpinsrd $1, _args_digest+5*32(state , idx, 4), %xmm1, %xmm1 193 vpinsrd $2, _args_digest+6*32(state , idx, 4), %xmm1, %xmm1 194 vpinsrd $3, _args_digest+7*32(state , idx, 4), %xmm1, %xmm1 195 196 vmovdqu %xmm0, _result_digest(job_rax) 197 vmovdqu %xmm1, _result_digest+1*16(job_rax) 198 199return: 200 pop %r12 201 pop %rbx 202 FRAME_END 203 ret 204 205return_null: 206 xor job_rax, job_rax 207 jmp return 208 209ENDPROC(sha256_mb_mgr_submit_avx2) 210 211.section .rodata.cst16.clear_low_nibble, "aM", @progbits, 16 212.align 16 213clear_low_nibble: 214 .octa 0x000000000000000000000000FFFFFFF0 215