+++ /dev/null
-/*
- * Copyright 2012 pooler@litecoinpool.org
- *
- * 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. See COPYING for more details.
- */
-
-#if defined(__arm__) && defined(__APCS_32__)
-
-.macro sha256_k
- .align 2
- .long 0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5
- .long 0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5
- .long 0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3
- .long 0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174
- .long 0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc
- .long 0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da
- .long 0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7
- .long 0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967
- .long 0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13
- .long 0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85
- .long 0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3
- .long 0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070
- .long 0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5
- .long 0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3
- .long 0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208
- .long 0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2
-.endm
-
-.macro sha256_extend_doubleround_core i, rw, ra, rb, ry, rz
- mov r12, \ry, ror #17
- add r11, r11, \ra
- eor r12, r12, \ry, ror #19
- mov \ra, lr, ror #7
- eor r12, r12, \ry, lsr #10
- eor \ra, \ra, lr, ror #18
- add r12, r12, r11
- ldr r11, [\rw, #(\i+2)*4]
- eor \ra, \ra, lr, lsr #3
- add \ra, \ra, r12
-
- mov r12, \rz, ror #17
- str \ra, [\rw, #(\i+16)*4]
- add lr, lr, \rb
- eor r12, r12, \rz, ror #19
- mov \rb, r11, ror #7
- eor r12, r12, \rz, lsr #10
- eor \rb, \rb, r11, ror #18
- add lr, lr, r12
- eor \rb, \rb, r11, lsr #3
- add \rb, \rb, lr
-.endm
-
-.macro sha256_extend_doubleround_head i, rw, ra, rb, ry, rz
- ldr lr, [\rw, #(\i+1)*4]
- sha256_extend_doubleround_core \i, \rw, \ra, \rb, \ry, \rz
- ldr lr, [\rw, #(\i+3)*4]
-.endm
-
-.macro sha256_extend_doubleround_body i, rw, ra, rb, ry, rz
- str \rz, [\rw, #(\i+15)*4]
- sha256_extend_doubleround_core \i, \rw, \ra, \rb, \ry, \rz
- ldr lr, [\rw, #(\i+3)*4]
-.endm
-
-.macro sha256_extend_doubleround_foot i, rw, ra, rb, ry, rz
- str \rz, [\rw, #(\i+15)*4]
- sha256_extend_doubleround_core \i, \rw, \ra, \rb, \ry, \rz
- str \rb, [\rw, #(\i+17)*4]
-.endm
-
-.macro sha256_main_round i, ka, rw, ra, rb, rc, rd, re, rf, rg, rh
- ldr r12, [\rw, #(\i)*4]
- and r3, \rf, \re
- bic lr, \rg, \re
- orr lr, lr, r3
- ldr r3, \ka + (\i)*4
- add \rh, \rh, lr
- eor lr, \re, \re, ror #5
- add \rh, \rh, r12
- eor lr, lr, \re, ror #19
- add \rh, \rh, r3
- eor r3, \ra, \rb
- add \rh, \rh, lr, ror #6
-
- and r3, r3, \rc
- eor r12, \ra, \ra, ror #11
- and lr, \ra, \rb
- eor r12, r12, \ra, ror #20
- eor lr, lr, r3
- add r3, \rh, lr
- add \rh, \rh, \rd
- add \rd, r3, r12, ror #2
-.endm
-
-.macro sha256_main_quadround i, ka, rw
- sha256_main_round \i+0, \ka, \rw, r4, r5, r6, r7, r8, r9, r10, r11
- sha256_main_round \i+1, \ka, \rw, r7, r4, r5, r6, r11, r8, r9, r10
- sha256_main_round \i+2, \ka, \rw, r6, r7, r4, r5, r10, r11, r8, r9
- sha256_main_round \i+3, \ka, \rw, r5, r6, r7, r4, r9, r10, r11, r8
-.endm
-
-
- .text
- .code 32
- .align 2
- .globl sha256_transform
- .globl _sha256_transform
-#ifdef __ELF__
- .type sha256_transform, %function
-#endif
-sha256_transform:
-_sha256_transform:
- stmfd sp!, {r4-r11, lr}
- cmp r2, #0
- sub sp, sp, #64*4
- bne sha256_transform_swap
-
- ldmia r1!, {r4-r11}
- stmia sp, {r4-r11}
- add r3, sp, #8*4
- ldmia r1, {r4-r11}
- stmia r3, {r4-r11}
- b sha256_transform_extend
-
-.macro bswap rd, rn
- eor r12, \rn, \rn, ror #16
- bic r12, r12, #0x00ff0000
- mov \rd, \rn, ror #8
- eor \rd, \rd, r12, lsr #8
-.endm
-
-sha256_transform_swap:
- ldmia r1!, {r4-r11}
- bswap r4, r4
- bswap r5, r5
- bswap r6, r6
- bswap r7, r7
- bswap r8, r8
- bswap r9, r9
- bswap r10, r10
- bswap r11, r11
- stmia sp, {r4-r11}
- add r3, sp, #8*4
- ldmia r1, {r4-r11}
- bswap r4, r4
- bswap r5, r5
- bswap r6, r6
- bswap r7, r7
- bswap r8, r8
- bswap r9, r9
- bswap r10, r10
- bswap r11, r11
- stmia r3, {r4-r11}
-
-sha256_transform_extend:
- add r12, sp, #9*4
- ldr r11, [sp, #0*4]
- ldmia r12, {r4-r10}
- sha256_extend_doubleround_head 0, sp, r4, r5, r9, r10
- sha256_extend_doubleround_body 2, sp, r6, r7, r4, r5
- sha256_extend_doubleround_body 4, sp, r8, r9, r6, r7
- sha256_extend_doubleround_body 6, sp, r10, r4, r8, r9
- sha256_extend_doubleround_body 8, sp, r5, r6, r10, r4
- sha256_extend_doubleround_body 10, sp, r7, r8, r5, r6
- sha256_extend_doubleround_body 12, sp, r9, r10, r7, r8
- sha256_extend_doubleround_body 14, sp, r4, r5, r9, r10
- sha256_extend_doubleround_body 16, sp, r6, r7, r4, r5
- sha256_extend_doubleround_body 18, sp, r8, r9, r6, r7
- sha256_extend_doubleround_body 20, sp, r10, r4, r8, r9
- sha256_extend_doubleround_body 22, sp, r5, r6, r10, r4
- sha256_extend_doubleround_body 24, sp, r7, r8, r5, r6
- sha256_extend_doubleround_body 26, sp, r9, r10, r7, r8
- sha256_extend_doubleround_body 28, sp, r4, r5, r9, r10
- sha256_extend_doubleround_body 30, sp, r6, r7, r4, r5
- sha256_extend_doubleround_body 32, sp, r8, r9, r6, r7
- sha256_extend_doubleround_body 34, sp, r10, r4, r8, r9
- sha256_extend_doubleround_body 36, sp, r5, r6, r10, r4
- sha256_extend_doubleround_body 38, sp, r7, r8, r5, r6
- sha256_extend_doubleround_body 40, sp, r9, r10, r7, r8
- sha256_extend_doubleround_body 42, sp, r4, r5, r9, r10
- sha256_extend_doubleround_body 44, sp, r6, r7, r4, r5
- sha256_extend_doubleround_foot 46, sp, r8, r9, r6, r7
-
- ldmia r0, {r4-r11}
- sha256_main_quadround 0, sha256_transform_k, sp
- sha256_main_quadround 4, sha256_transform_k, sp
- sha256_main_quadround 8, sha256_transform_k, sp
- sha256_main_quadround 12, sha256_transform_k, sp
- sha256_main_quadround 16, sha256_transform_k, sp
- sha256_main_quadround 20, sha256_transform_k, sp
- sha256_main_quadround 24, sha256_transform_k, sp
- sha256_main_quadround 28, sha256_transform_k, sp
- b sha256_transform_k_over
-sha256_transform_k:
- sha256_k
-sha256_transform_k_over:
- sha256_main_quadround 32, sha256_transform_k, sp
- sha256_main_quadround 36, sha256_transform_k, sp
- sha256_main_quadround 40, sha256_transform_k, sp
- sha256_main_quadround 44, sha256_transform_k, sp
- sha256_main_quadround 48, sha256_transform_k, sp
- sha256_main_quadround 52, sha256_transform_k, sp
- sha256_main_quadround 56, sha256_transform_k, sp
- sha256_main_quadround 60, sha256_transform_k, sp
-
- ldmia r0, {r1, r2, r3, r12}
- add r4, r4, r1
- add r5, r5, r2
- add r6, r6, r3
- add r7, r7, r12
- stmia r0!, {r4-r7}
- ldmia r0, {r1, r2, r3, r12}
- add r8, r8, r1
- add r9, r9, r2
- add r10, r10, r3
- add r11, r11, r12
- stmia r0, {r8-r11}
-
- add sp, sp, #64*4
-#ifdef __thumb__
- ldmfd sp!, {r4-r11, lr}
- bx lr
-#else
- ldmfd sp!, {r4-r11, pc}
-#endif
-
-.macro sha256_main_round_red i, ka, rw, rd, re, rf, rg, rh
- ldr r12, [\rw, #(\i)*4]
- and r3, \rf, \re
- bic lr, \rg, \re
- add \rh, \rh, \rd
- orr lr, lr, r3
- ldr r3, \ka + (\i)*4
- add \rh, \rh, lr
- eor lr, \re, \re, ror #5
- add \rh, \rh, r12
- eor lr, lr, \re, ror #19
- add \rh, \rh, r3
- add \rh, \rh, lr, ror #6
-.endm
-
- .text
- .code 32
- .align 2
- .globl sha256_init_4way
- .globl _sha256_init_4way
-#ifdef __ELF__
- .type sha256_init_4way, %function
-#endif
-sha256_init_4way:
-_sha256_init_4way:
- adr r12, sha256_4h
- vldmia r12, {q8-q15}
- vstmia r0, {q8-q15}
- bx lr
- .align 4
-sha256_4h:
- .long 0x6a09e667, 0x6a09e667, 0x6a09e667, 0x6a09e667
- .long 0xbb67ae85, 0xbb67ae85, 0xbb67ae85, 0xbb67ae85
- .long 0x3c6ef372, 0x3c6ef372, 0x3c6ef372, 0x3c6ef372
- .long 0xa54ff53a, 0xa54ff53a, 0xa54ff53a, 0xa54ff53a
- .long 0x510e527f, 0x510e527f, 0x510e527f, 0x510e527f
- .long 0x9b05688c, 0x9b05688c, 0x9b05688c, 0x9b05688c
- .long 0x1f83d9ab, 0x1f83d9ab, 0x1f83d9ab, 0x1f83d9ab
- .long 0x5be0cd19, 0x5be0cd19, 0x5be0cd19, 0x5be0cd19
-
-.macro sha256_4k
- .long 0x428a2f98, 0x428a2f98, 0x428a2f98, 0x428a2f98
- .long 0x71374491, 0x71374491, 0x71374491, 0x71374491
- .long 0xb5c0fbcf, 0xb5c0fbcf, 0xb5c0fbcf, 0xb5c0fbcf
- .long 0xe9b5dba5, 0xe9b5dba5, 0xe9b5dba5, 0xe9b5dba5
- .long 0x3956c25b, 0x3956c25b, 0x3956c25b, 0x3956c25b
- .long 0x59f111f1, 0x59f111f1, 0x59f111f1, 0x59f111f1
- .long 0x923f82a4, 0x923f82a4, 0x923f82a4, 0x923f82a4
- .long 0xab1c5ed5, 0xab1c5ed5, 0xab1c5ed5, 0xab1c5ed5
- .long 0xd807aa98, 0xd807aa98, 0xd807aa98, 0xd807aa98
- .long 0x12835b01, 0x12835b01, 0x12835b01, 0x12835b01
- .long 0x243185be, 0x243185be, 0x243185be, 0x243185be
- .long 0x550c7dc3, 0x550c7dc3, 0x550c7dc3, 0x550c7dc3
- .long 0x72be5d74, 0x72be5d74, 0x72be5d74, 0x72be5d74
- .long 0x80deb1fe, 0x80deb1fe, 0x80deb1fe, 0x80deb1fe
- .long 0x9bdc06a7, 0x9bdc06a7, 0x9bdc06a7, 0x9bdc06a7
- .long 0xc19bf174, 0xc19bf174, 0xc19bf174, 0xc19bf174
- .long 0xe49b69c1, 0xe49b69c1, 0xe49b69c1, 0xe49b69c1
- .long 0xefbe4786, 0xefbe4786, 0xefbe4786, 0xefbe4786
- .long 0x0fc19dc6, 0x0fc19dc6, 0x0fc19dc6, 0x0fc19dc6
- .long 0x240ca1cc, 0x240ca1cc, 0x240ca1cc, 0x240ca1cc
- .long 0x2de92c6f, 0x2de92c6f, 0x2de92c6f, 0x2de92c6f
- .long 0x4a7484aa, 0x4a7484aa, 0x4a7484aa, 0x4a7484aa
- .long 0x5cb0a9dc, 0x5cb0a9dc, 0x5cb0a9dc, 0x5cb0a9dc
- .long 0x76f988da, 0x76f988da, 0x76f988da, 0x76f988da
- .long 0x983e5152, 0x983e5152, 0x983e5152, 0x983e5152
- .long 0xa831c66d, 0xa831c66d, 0xa831c66d, 0xa831c66d
- .long 0xb00327c8, 0xb00327c8, 0xb00327c8, 0xb00327c8
- .long 0xbf597fc7, 0xbf597fc7, 0xbf597fc7, 0xbf597fc7
- .long 0xc6e00bf3, 0xc6e00bf3, 0xc6e00bf3, 0xc6e00bf3
- .long 0xd5a79147, 0xd5a79147, 0xd5a79147, 0xd5a79147
- .long 0x06ca6351, 0x06ca6351, 0x06ca6351, 0x06ca6351
- .long 0x14292967, 0x14292967, 0x14292967, 0x14292967
- .long 0x27b70a85, 0x27b70a85, 0x27b70a85, 0x27b70a85
- .long 0x2e1b2138, 0x2e1b2138, 0x2e1b2138, 0x2e1b2138
- .long 0x4d2c6dfc, 0x4d2c6dfc, 0x4d2c6dfc, 0x4d2c6dfc
- .long 0x53380d13, 0x53380d13, 0x53380d13, 0x53380d13
- .long 0x650a7354, 0x650a7354, 0x650a7354, 0x650a7354
- .long 0x766a0abb, 0x766a0abb, 0x766a0abb, 0x766a0abb
- .long 0x81c2c92e, 0x81c2c92e, 0x81c2c92e, 0x81c2c92e
- .long 0x92722c85, 0x92722c85, 0x92722c85, 0x92722c85
- .long 0xa2bfe8a1, 0xa2bfe8a1, 0xa2bfe8a1, 0xa2bfe8a1
- .long 0xa81a664b, 0xa81a664b, 0xa81a664b, 0xa81a664b
- .long 0xc24b8b70, 0xc24b8b70, 0xc24b8b70, 0xc24b8b70
- .long 0xc76c51a3, 0xc76c51a3, 0xc76c51a3, 0xc76c51a3
- .long 0xd192e819, 0xd192e819, 0xd192e819, 0xd192e819
- .long 0xd6990624, 0xd6990624, 0xd6990624, 0xd6990624
- .long 0xf40e3585, 0xf40e3585, 0xf40e3585, 0xf40e3585
- .long 0x106aa070, 0x106aa070, 0x106aa070, 0x106aa070
- .long 0x19a4c116, 0x19a4c116, 0x19a4c116, 0x19a4c116
- .long 0x1e376c08, 0x1e376c08, 0x1e376c08, 0x1e376c08
- .long 0x2748774c, 0x2748774c, 0x2748774c, 0x2748774c
- .long 0x34b0bcb5, 0x34b0bcb5, 0x34b0bcb5, 0x34b0bcb5
- .long 0x391c0cb3, 0x391c0cb3, 0x391c0cb3, 0x391c0cb3
- .long 0x4ed8aa4a, 0x4ed8aa4a, 0x4ed8aa4a, 0x4ed8aa4a
- .long 0x5b9cca4f, 0x5b9cca4f, 0x5b9cca4f, 0x5b9cca4f
- .long 0x682e6ff3, 0x682e6ff3, 0x682e6ff3, 0x682e6ff3
- .long 0x748f82ee, 0x748f82ee, 0x748f82ee, 0x748f82ee
- .long 0x78a5636f, 0x78a5636f, 0x78a5636f, 0x78a5636f
- .long 0x84c87814, 0x84c87814, 0x84c87814, 0x84c87814
- .long 0x8cc70208, 0x8cc70208, 0x8cc70208, 0x8cc70208
- .long 0x90befffa, 0x90befffa, 0x90befffa, 0x90befffa
- .long 0xa4506ceb, 0xa4506ceb, 0xa4506ceb, 0xa4506ceb
- .long 0xbef9a3f7, 0xbef9a3f7, 0xbef9a3f7, 0xbef9a3f7
- .long 0xc67178f2, 0xc67178f2, 0xc67178f2, 0xc67178f2
-.endm
-
-.macro sha256_4way_extend_doubleround_core i, rr, rw, ra, rb, ry, rz
- vadd.u32 q5, q5, \ra
- veor.u32 q4, q4, q0
- vshr.u32 q0, \ry, #19
- vshl.u32 q1, \ry, #32-19
- veor.u32 q4, q4, q0
- vshr.u32 \ra, q6, #7
- vshl.u32 q0, q6, #32-7
- veor.u32 q4, q4, q1
- veor.u32 \ra, \ra, q0
- vshr.u32 q1, \ry, #10
- vshr.u32 q0, q6, #18
- veor.u32 q4, q4, q1
- veor.u32 \ra, \ra, q0
- vshl.u32 q1, q6, #32-18
- vshr.u32 q0, q6, #3
- veor.u32 \ra, \ra, q1
- vadd.u32 q4, q4, q5
- veor.u32 \ra, \ra, q0
- vld1.u32 {q5}, [\rr]!
- vadd.u32 \ra, \ra, q4
-
- vshr.u32 q4, \rz, #17
- vshl.u32 q0, \rz, #32-17
- vadd.u32 q6, q6, \rb
- vst1.u32 {\ra}, [\rw]!
- veor.u32 q4, q4, q0
- vshr.u32 q0, \rz, #19
- vshl.u32 q1, \rz, #32-19
- veor.u32 q4, q4, q0
- vshr.u32 \rb, q5, #7
- veor.u32 q4, q4, q1
- vshl.u32 q0, q5, #32-7
- vshr.u32 q1, \rz, #10
- veor.u32 \rb, \rb, q0
- vshr.u32 q0, q5, #18
- veor.u32 q4, q4, q1
- veor.u32 \rb, \rb, q0
- vshl.u32 q1, q5, #32-18
- vshr.u32 q0, q5, #3
- veor.u32 \rb, \rb, q1
- vadd.u32 q1, q6, q4
- veor.u32 \rb, \rb, q0
-.endm
-
-.macro sha256_4way_extend_doubleround_head i, rr, rw, ra, rb, ry, rz
- vld1.u32 {q6}, [\rr]!
- vshr.u32 q4, \ry, #17
- vshl.u32 q0, \ry, #32-17
- sha256_4way_extend_doubleround_core \i, \rr, \rw, \ra, \rb, \ry, \rz
- vld1.u32 {q6}, [\rr]!
- vadd.u32 \rb, \rb, q1
-.endm
-
-.macro sha256_4way_extend_doubleround_body i, rr, rw, ra, rb, ry, rz
- vshr.u32 q4, \ry, #17
- vshl.u32 q0, \ry, #32-17
- vst1.u32 {\rz}, [\rw]!
- sha256_4way_extend_doubleround_core \i, \rr, \rw, \ra, \rb, \ry, \rz
- vld1.u32 {q6}, [\rr]!
- vadd.u32 \rb, \rb, q1
-.endm
-
-.macro sha256_4way_extend_doubleround_foot i, rr, rw, ra, rb, ry, rz
- vshr.u32 q4, \ry, #17
- vshl.u32 q0, \ry, #32-17
- vst1.u32 {\rz}, [\rw]!
- sha256_4way_extend_doubleround_core \i, \rr, \rw, \ra, \rb, \ry, \rz
- vadd.u32 \rb, \rb, q1
- vst1.u32 {\rb}, [\rw]!
-.endm
-
-.macro sha256_4way_main_round i, rk, rw, ra, rb, rc, rd, re, rf, rg, rh
- vld1.u32 {q8}, [\rw]!
- vand.u32 q9, \rf, \re
- vbic.u32 q10, \rg, \re
- vshr.u32 q11, \re, #5
- vorr.u32 q10, q10, q9
- vld1.u32 {q9}, [\rk]!
- vadd.u32 \rh, \rh, q10
- vshl.u32 q12, \re, #32-5
- veor.u32 q10, \re, q11
- vshr.u32 q11, \re, #19
- veor.u32 q10, q10, q12
- vshl.u32 q12, \re, #32-19
- veor.u32 q10, q10, q11
- vadd.u32 \rh, \rh, q8
- veor.u32 q10, q10, q12
- vadd.u32 \rh, \rh, q9
- veor.u32 q9, \ra, \rb
- vshr.u32 q11, q10, #6
- vshl.u32 q13, q10, #32-6
- vadd.u32 \rh, \rh, q11
-
- vshr.u32 q11, \ra, #11
- vshl.u32 q12, \ra, #32-11
- veor.u32 q8, \ra, q11
- vand.u32 q10, \ra, \rb
- veor.u32 q8, q8, q12
- vshr.u32 q11, \ra, #20
- vshl.u32 q12, \ra, #32-20
- veor.u32 q8, q8, q11
- vand.u32 q9, q9, \rc
- veor.u32 q8, q8, q12
- vadd.u32 \rh, \rh, q13
- veor.u32 q10, q10, q9
- vshr.u32 q11, q8, #2
- vshl.u32 q12, q8, #32-2
- vadd.u32 q9, \rh, q10
- vadd.u32 q12, q12, q11
- vadd.u32 \rh, \rh, \rd
- vadd.u32 \rd, q9, q12
-.endm
-
-.macro sha256_4way_main_quadround i, rk, rw
- sha256_4way_main_round \i+0, \rk, \rw, q0, q1, q2, q3, q4, q5, q6, q7
- sha256_4way_main_round \i+1, \rk, \rw, q3, q0, q1, q2, q7, q4, q5, q6
- sha256_4way_main_round \i+2, \rk, \rw, q2, q3, q0, q1, q6, q7, q4, q5
- sha256_4way_main_round \i+3, \rk, \rw, q1, q2, q3, q0, q5, q6, q7, q4
-.endm
-
-
- .text
- .code 32
- .align 2
- .globl sha256_transform_4way
- .globl _sha256_transform_4way
-#ifdef __ELF__
- .type sha256_transform_4way, %function
-#endif
-sha256_transform_4way:
-_sha256_transform_4way:
- stmfd sp!, {r4, lr}
- vpush {q4-q7}
- mov r12, sp
- sub sp, sp, #64*16
- bic sp, sp, #63
- cmp r2, #0
- bne sha256_transform_4way_swap
-
- vldmia r1!, {q0-q7}
- vstmia sp, {q0-q7}
- add r3, sp, #8*16
- vldmia r1, {q8-q15}
- vstmia r3, {q8-q15}
- b sha256_transform_4way_extend
-
-sha256_transform_4way_swap:
- vldmia r1!, {q0-q7}
- vrev32.8 q0, q0
- vrev32.8 q1, q1
- vrev32.8 q2, q2
- vrev32.8 q3, q3
- vldmia r1, {q8-q15}
- vrev32.8 q4, q4
- vrev32.8 q5, q5
- vrev32.8 q6, q6
- vrev32.8 q7, q7
- vstmia sp, {q0-q7}
- vrev32.8 q8, q8
- vrev32.8 q9, q9
- vrev32.8 q10, q10
- vrev32.8 q11, q11
- vrev32.8 q12, q12
- vrev32.8 q13, q13
- vrev32.8 q14, q14
- vrev32.8 q15, q15
- add r3, sp, #8*16
- vstmia r3, {q8-q15}
-
-sha256_transform_4way_extend:
- add r1, sp, #1*16
- add r2, sp, #16*16
- vmov.u32 q5, q0
- sha256_4way_extend_doubleround_head 0, r1, r2, q9, q10, q14, q15
- sha256_4way_extend_doubleround_body 2, r1, r2, q11, q12, q9, q10
- sha256_4way_extend_doubleround_body 4, r1, r2, q13, q14, q11, q12
- sha256_4way_extend_doubleround_body 6, r1, r2, q15, q9, q13, q14
- sha256_4way_extend_doubleround_body 8, r1, r2, q10, q11, q15, q9
- sha256_4way_extend_doubleround_body 10, r1, r2, q12, q13, q10, q11
- sha256_4way_extend_doubleround_body 12, r1, r2, q14, q15, q12, q13
- sha256_4way_extend_doubleround_body 14, r1, r2, q9, q10, q14, q15
- sha256_4way_extend_doubleround_body 16, r1, r2, q11, q12, q9, q10
- sha256_4way_extend_doubleround_body 18, r1, r2, q13, q14, q11, q12
- sha256_4way_extend_doubleround_body 20, r1, r2, q15, q9, q13, q14
- sha256_4way_extend_doubleround_body 22, r1, r2, q10, q11, q15, q9
- sha256_4way_extend_doubleround_body 24, r1, r2, q12, q13, q10, q11
- sha256_4way_extend_doubleround_body 26, r1, r2, q14, q15, q12, q13
- sha256_4way_extend_doubleround_body 28, r1, r2, q9, q10, q14, q15
- sha256_4way_extend_doubleround_body 30, r1, r2, q11, q12, q9, q10
- sha256_4way_extend_doubleround_body 32, r1, r2, q13, q14, q11, q12
- sha256_4way_extend_doubleround_body 34, r1, r2, q15, q9, q13, q14
- sha256_4way_extend_doubleround_body 36, r1, r2, q10, q11, q15, q9
- sha256_4way_extend_doubleround_body 38, r1, r2, q12, q13, q10, q11
- sha256_4way_extend_doubleround_body 40, r1, r2, q14, q15, q12, q13
- sha256_4way_extend_doubleround_body 42, r1, r2, q9, q10, q14, q15
- sha256_4way_extend_doubleround_body 44, r1, r2, q11, q12, q9, q10
- sha256_4way_extend_doubleround_foot 46, r1, r2, q13, q14, q11, q12
-
- vldmia r0, {q0-q7}
- adr r4, sha256_transform_4way_4k
- b sha256_transform_4way_4k_over
- .align 4
-sha256_transform_4way_4k:
- sha256_4k
-sha256_transform_4way_4k_over:
- sha256_4way_main_quadround 0, r4, sp
- sha256_4way_main_quadround 4, r4, sp
- sha256_4way_main_quadround 8, r4, sp
- sha256_4way_main_quadround 12, r4, sp
- sha256_4way_main_quadround 16, r4, sp
- sha256_4way_main_quadround 20, r4, sp
- sha256_4way_main_quadround 24, r4, sp
- sha256_4way_main_quadround 28, r4, sp
- sha256_4way_main_quadround 32, r4, sp
- sha256_4way_main_quadround 36, r4, sp
- sha256_4way_main_quadround 40, r4, sp
- sha256_4way_main_quadround 44, r4, sp
- sha256_4way_main_quadround 48, r4, sp
- sha256_4way_main_quadround 52, r4, sp
- sha256_4way_main_quadround 56, r4, sp
- sha256_4way_main_quadround 60, r4, sp
-
- vldmia r0, {q8-q15}
- vadd.u32 q0, q0, q8
- vadd.u32 q1, q1, q9
- vadd.u32 q2, q2, q10
- vadd.u32 q3, q3, q11
- vadd.u32 q4, q4, q12
- vadd.u32 q5, q5, q13
- vadd.u32 q6, q6, q14
- vadd.u32 q7, q7, q15
- vstmia r0, {q0-q7}
-
- mov sp, r12
- vpop {q4-q7}
- ldmfd sp!, {r4, pc}
-
-
-.macro sha256_4way_main_round_red i, rk, rw, rd, re, rf, rg, rh
- vld1.u32 {q8}, [\rw]!
- vand.u32 q9, \rf, \re
- vbic.u32 q10, \rg, \re
- vshr.u32 q11, \re, #5
- vorr.u32 q10, q10, q9
- vshl.u32 q12, \re, #32-5
- vadd.u32 \rh, \rh, q10
- veor.u32 q10, \re, q11
- vshr.u32 q11, \re, #19
- veor.u32 q10, q10, q12
- vshl.u32 q12, \re, #32-19
- veor.u32 q10, q10, q11
- vadd.u32 \rh, \rh, q8
- veor.u32 q10, q10, q12
- vld1.u32 {q9}, [\rk]!
- vadd.u32 \rh, \rh, \rd
- vshr.u32 q11, q10, #6
- vadd.u32 \rh, \rh, q9
- vshl.u32 q13, q10, #32-6
- vadd.u32 \rh, \rh, q11
- vadd.u32 \rh, \rh, q13
-.endm
-
- .text
- .code 32
- .align 2
- .globl sha256_use_4way
- .globl _sha256_use_4way
-#ifdef __ELF__
- .type sha256_use_4way, %function
-#endif
-sha256_use_4way:
-_sha256_use_4way:
- mov r0, #1
- bx lr
-
-#endif
+++ /dev/null
-/*
- * Copyright 2012-2015 pooler@litecoinpool.org
- *
- * 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. See COPYING for more details.
- */
-
-#if defined(__linux__) && defined(__ELF__)
- .section .note.GNU-stack,"",%progbits
-#endif
-
-#if defined(__x86_64__)
- .data
- .p2align 4
-sha256_h:
- .long 0x6a09e667, 0xbb67ae85, 0x3c6ef372, 0xa54ff53a
- .long 0x510e527f, 0x9b05688c, 0x1f83d9ab, 0x5be0cd19
-
- .data
- .p2align 6
-sha256_k:
- .long 0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5
- .long 0x3956c25b, 0x59f111f1, 0x923f82a4, 0xab1c5ed5
- .long 0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3
- .long 0x72be5d74, 0x80deb1fe, 0x9bdc06a7, 0xc19bf174
- .long 0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc
- .long 0x2de92c6f, 0x4a7484aa, 0x5cb0a9dc, 0x76f988da
- .long 0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7
- .long 0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967
- .long 0x27b70a85, 0x2e1b2138, 0x4d2c6dfc, 0x53380d13
- .long 0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85
- .long 0xa2bfe8a1, 0xa81a664b, 0xc24b8b70, 0xc76c51a3
- .long 0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070
- .long 0x19a4c116, 0x1e376c08, 0x2748774c, 0x34b0bcb5
- .long 0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3
- .long 0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208
- .long 0x90befffa, 0xa4506ceb, 0xbef9a3f7, 0xc67178f2
-
-bswap_xmm_mask:
- .long 0x00010203, 0x04050607, 0x08090a0b, 0x0c0d0e0f
-
-
-.macro sha256_mixed_quadround ra, rb, rc, rd, re, rf, rg, rh, x0, x1, x2, x3
- movdqa \x3, %xmm4
- movl \re, %eax
- movdqa \x2, %xmm6
- rorl $(25-11), %eax
- movl \ra, %ebx
- pslldq $12, %xmm4
- rorl $(22-13), %ebx
- psrldq $4, %xmm6
- xorl \re, %eax
- movl \rf, %ecx
- rorl $(11-6), %eax
- pxor %xmm6, %xmm4
- movdqa \x1, %xmm5
- xorl \ra, %ebx
- xorl \rg, %ecx
- xorl \re, %eax
- paddd \x0, %xmm4
- movdqa \x0, %xmm7
- andl \re, %ecx
- rorl $(13-2), %ebx
- xorl \ra, %ebx
- pslldq $12, %xmm5
- psrldq $4, %xmm7
- rorl $6, %eax
- xorl \rg, %ecx
- pxor %xmm7, %xmm5
- rorl $2, %ebx
- addl %eax, %ecx
- addl (%rsp) , %ecx
- movdqa %xmm5, %xmm6
- movl \ra, %eax
- addl %ecx, \rh
- movl \ra, %ecx
- movdqa %xmm5, %xmm7
- orl \rc, %eax
- addl \rh, \rd
- andl \rc, %ecx
- pslld $(32-7), %xmm5
- psrld $7, %xmm6
- andl \rb, %eax
- addl %ebx, \rh
- orl %ecx, %eax
- por %xmm6, %xmm5
- addl %eax, \rh
-
- movl \rd, %eax
- movdqa %xmm7, %xmm6
- movl \rh, %ebx
- rorl $(25-11), %eax
- xorl \rd, %eax
- movdqa %xmm7, %xmm8
- movl \re, %ecx
- rorl $(22-13), %ebx
- xorl \rh, %ebx
- pslld $(32-18), %xmm7
- rorl $(11-6), %eax
- xorl \rf, %ecx
- rorl $(13-2), %ebx
- psrld $18, %xmm6
- xorl \rd, %eax
- andl \rd, %ecx
- rorl $6, %eax
- pxor %xmm7, %xmm5
- xorl \rh, %ebx
- xorl \rf, %ecx
- psrld $3, %xmm8
- addl %eax, %ecx
- addl 1*4(%rsp), %ecx
- rorl $2, %ebx
- pxor %xmm6, %xmm5
- movl \rh, %eax
- addl %ecx, \rg
- movl \rh, %ecx
- pxor %xmm8, %xmm5
- orl \rb, %eax
- addl \rg, \rc
- andl \rb, %ecx
- pshufd $0xfa, \x3, %xmm6
- andl \ra, %eax
- addl %ebx, \rg
- paddd %xmm5, %xmm4
- orl %ecx, %eax
- addl %eax, \rg
-
- movl \rc, %eax
- movdqa %xmm6, %xmm7
- movl \rg, %ebx
- rorl $(25-11), %eax
- xorl \rc, %eax
- movdqa %xmm6, %xmm8
- rorl $(22-13), %ebx
- movl \rd, %ecx
- xorl \rg, %ebx
- psrlq $17, %xmm6
- psrlq $19, %xmm7
- rorl $(11-6), %eax
- xorl \re, %ecx
- xorl \rc, %eax
- psrld $10, %xmm8
- pxor %xmm7, %xmm6
- andl \rc, %ecx
- rorl $(13-2), %ebx
- xorl \rg, %ebx
- pxor %xmm6, %xmm8
- xorl \re, %ecx
- rorl $6, %eax
- addl %eax, %ecx
- pshufd $0x8f, %xmm8, %xmm8
- rorl $2, %ebx
- addl 2*4(%rsp), %ecx
- movl \rg, %eax
- psrldq $8, %xmm8
- addl %ecx, \rf
- movl \rg, %ecx
- orl \ra, %eax
- paddd %xmm8, %xmm4
- addl \rf, \rb
- andl \ra, %ecx
- andl \rh, %eax
- pshufd $0x50, %xmm4, %xmm6
- addl %ebx, \rf
- orl %ecx, %eax
- addl %eax, \rf
-
- movdqa %xmm6, %xmm7
- movl \rb, %eax
- rorl $(25-11), %eax
- movl \rf, %ebx
- movdqa %xmm6, \x0
- rorl $(22-13), %ebx
- xorl \rb, %eax
- movl \rc, %ecx
- psrlq $17, %xmm6
- rorl $(11-6), %eax
- xorl \rf, %ebx
- xorl \rd, %ecx
- psrlq $19, %xmm7
- xorl \rb, %eax
- andl \rb, %ecx
- rorl $(13-2), %ebx
- psrld $10, \x0
- xorl \rf, %ebx
- rorl $6, %eax
- pxor %xmm7, %xmm6
- xorl \rd, %ecx
- rorl $2, %ebx
- addl %eax, %ecx
- pxor %xmm6, \x0
- addl 3*4(%rsp), %ecx
- movl \rf, %eax
- addl %ecx, \re
- pshufd $0xf8, \x0, \x0
- movl \rf, %ecx
- orl \rh, %eax
- addl \re, \ra
- pslldq $8, \x0
- andl \rh, %ecx
- andl \rg, %eax
- paddd %xmm4, \x0
- addl %ebx, \re
- orl %ecx, %eax
- addl %eax, \re
-.endm
-
-.macro sha256_main_round i, ra, rb, rc, rd, re, rf, rg, rh
- movl \re, %eax
- rorl $(25-11), %eax
- movl \ra, %ebx
- xorl \re, %eax
- rorl $(22-13), %ebx
- movl \rf, %ecx
- xorl \ra, %ebx
- rorl $(11-6), %eax
- xorl \rg, %ecx
- xorl \re, %eax
- rorl $(13-2), %ebx
- andl \re, %ecx
- xorl \ra, %ebx
- rorl $6, %eax
- xorl \rg, %ecx
- addl %eax, %ecx
- rorl $2, %ebx
- addl \i*4(%rsp), %ecx
- movl \ra, %eax
- addl %ecx, \rh
- movl \ra, %ecx
- orl \rc, %eax
- addl \rh, \rd
- andl \rc, %ecx
- andl \rb, %eax
- addl %ebx, \rh
- orl %ecx, %eax
- addl %eax, \rh
-.endm
-
-
- .text
- .p2align 6
-sha256_transform_sse2:
- pushq %rbx
- pushq %r12
- pushq %r13
- pushq %r14
- pushq %r15
-#if defined(_WIN64) || defined(__CYGWIN__)
- pushq %rdi
- pushq %rsi
- subq $5*16, %rsp
- movdqa %xmm6, 1*16(%rsp)
- movdqa %xmm7, 2*16(%rsp)
- movdqa %xmm8, 3*16(%rsp)
- movdqa %xmm9, 4*16(%rsp)
- movq %rcx, %rdi
- movq %rdx, %rsi
- movq %r8, %rdx
-#else
- subq $16, %rsp
-#endif
-
- movl 0*4(%rdi), %r8d
- movl 1*4(%rdi), %r9d
- movl 2*4(%rdi), %r10d
- movl 3*4(%rdi), %r11d
- movl 4*4(%rdi), %r12d
- movl 5*4(%rdi), %r13d
- movl 6*4(%rdi), %r14d
- movl 7*4(%rdi), %r15d
-
- testq %rdx, %rdx
- jnz sha256_transform_sse2_swap
-
- movdqu 0*16(%rsi), %xmm0
- movdqu 1*16(%rsi), %xmm1
- movdqu 2*16(%rsi), %xmm2
- movdqu 3*16(%rsi), %xmm3
- jmp sha256_transform_sse2_core
-
-sha256_transform_sse2_swap:
- movdqu 0*16(%rsi), %xmm0
- movdqu 1*16(%rsi), %xmm1
- movdqu 2*16(%rsi), %xmm2
- movdqu 3*16(%rsi), %xmm3
- pshuflw $0xb1, %xmm0, %xmm0
- pshuflw $0xb1, %xmm1, %xmm1
- pshuflw $0xb1, %xmm2, %xmm2
- pshuflw $0xb1, %xmm3, %xmm3
- pshufhw $0xb1, %xmm0, %xmm0
- pshufhw $0xb1, %xmm1, %xmm1
- pshufhw $0xb1, %xmm2, %xmm2
- pshufhw $0xb1, %xmm3, %xmm3
- movdqa %xmm0, %xmm4
- movdqa %xmm1, %xmm5
- movdqa %xmm2, %xmm6
- movdqa %xmm3, %xmm7
- psrlw $8, %xmm4
- psrlw $8, %xmm5
- psrlw $8, %xmm6
- psrlw $8, %xmm7
- psllw $8, %xmm0
- psllw $8, %xmm1
- psllw $8, %xmm2
- psllw $8, %xmm3
- pxor %xmm4, %xmm0
- pxor %xmm5, %xmm1
- pxor %xmm6, %xmm2
- pxor %xmm7, %xmm3
-
-sha256_transform_sse2_core:
- leaq sha256_k(%rip), %rdx
- movq $48, %rsi
- .p2align 4
-sha256_transform_sse2_loop:
- movdqa 0*16(%rdx), %xmm9
- paddd %xmm0, %xmm9
- movdqa %xmm9, (%rsp)
- sha256_mixed_quadround %r8d, %r9d, %r10d, %r11d, %r12d, %r13d, %r14d, %r15d, %xmm0, %xmm1, %xmm2, %xmm3
- movdqa 1*16(%rdx), %xmm9
- paddd %xmm1, %xmm9
- movdqa %xmm9, (%rsp)
- sha256_mixed_quadround %r12d, %r13d, %r14d, %r15d, %r8d, %r9d, %r10d, %r11d, %xmm1, %xmm2, %xmm3, %xmm0
- movdqa 2*16(%rdx), %xmm9
- paddd %xmm2, %xmm9
- movdqa %xmm9, (%rsp)
- sha256_mixed_quadround %r8d, %r9d, %r10d, %r11d, %r12d, %r13d, %r14d, %r15d, %xmm2, %xmm3, %xmm0, %xmm1
- movdqa 3*16(%rdx), %xmm9
- paddd %xmm3, %xmm9
- movdqa %xmm9, (%rsp)
- addq $4*16, %rdx
- sha256_mixed_quadround %r12d, %r13d, %r14d, %r15d, %r8d, %r9d, %r10d, %r11d, %xmm3, %xmm0, %xmm1, %xmm2
-
- subq $16, %rsi
- jne sha256_transform_sse2_loop
-
- paddd 0*16(%rdx), %xmm0
- movdqa %xmm0, (%rsp)
- sha256_main_round 0, %r8d, %r9d, %r10d, %r11d, %r12d, %r13d, %r14d, %r15d
- sha256_main_round 1, %r15d, %r8d, %r9d, %r10d, %r11d, %r12d, %r13d, %r14d
- sha256_main_round 2, %r14d, %r15d, %r8d, %r9d, %r10d, %r11d, %r12d, %r13d
- sha256_main_round 3, %r13d, %r14d, %r15d, %r8d, %r9d, %r10d, %r11d, %r12d
- paddd 1*16(%rdx), %xmm1
- movdqa %xmm1, (%rsp)
- sha256_main_round 0, %r12d, %r13d, %r14d, %r15d, %r8d, %r9d, %r10d, %r11d
- sha256_main_round 1, %r11d, %r12d, %r13d, %r14d, %r15d, %r8d, %r9d, %r10d
- sha256_main_round 2, %r10d, %r11d, %r12d, %r13d, %r14d, %r15d, %r8d, %r9d
- sha256_main_round 3, %r9d, %r10d, %r11d, %r12d, %r13d, %r14d, %r15d, %r8d
- paddd 2*16(%rdx), %xmm2
- movdqa %xmm2, (%rsp)
- sha256_main_round 0, %r8d, %r9d, %r10d, %r11d, %r12d, %r13d, %r14d, %r15d
- sha256_main_round 1, %r15d, %r8d, %r9d, %r10d, %r11d, %r12d, %r13d, %r14d
- sha256_main_round 2, %r14d, %r15d, %r8d, %r9d, %r10d, %r11d, %r12d, %r13d
- sha256_main_round 3, %r13d, %r14d, %r15d, %r8d, %r9d, %r10d, %r11d, %r12d
- paddd 3*16(%rdx), %xmm3
- movdqa %xmm3, (%rsp)
- sha256_main_round 0, %r12d, %r13d, %r14d, %r15d, %r8d, %r9d, %r10d, %r11d
- sha256_main_round 1, %r11d, %r12d, %r13d, %r14d, %r15d, %r8d, %r9d, %r10d
- sha256_main_round 2, %r10d, %r11d, %r12d, %r13d, %r14d, %r15d, %r8d, %r9d
- sha256_main_round 3, %r9d, %r10d, %r11d, %r12d, %r13d, %r14d, %r15d, %r8d
-
- addl %r8d, 0*4(%rdi)
- addl %r9d, 1*4(%rdi)
- addl %r10d, 2*4(%rdi)
- addl %r11d, 3*4(%rdi)
- addl %r12d, 4*4(%rdi)
- addl %r13d, 5*4(%rdi)
- addl %r14d, 6*4(%rdi)
- addl %r15d, 7*4(%rdi)
-
-#if defined(_WIN64) || defined(__CYGWIN__)
- movdqa 1*16(%rsp), %xmm6
- movdqa 2*16(%rsp), %xmm7
- movdqa 3*16(%rsp), %xmm8
- movdqa 4*16(%rsp), %xmm9
- addq $5*16, %rsp
- popq %rsi
- popq %rdi
-#else
- addq $16, %rsp
-#endif
- popq %r15
- popq %r14
- popq %r13
- popq %r12
- popq %rbx
- ret
-
-
- .text
- .p2align 6
-sha256_transform_phe:
-#if defined(_WIN64) || defined(__CYGWIN__)
- pushq %rdi
- pushq %rsi
- movq %rcx, %rdi
- movq %rdx, %rsi
- movq %r8, %rdx
-#endif
- movq %rsp, %r8
- subq $64, %rsp
- andq $-64, %rsp
-
- testq %rdx, %rdx
- jnz sha256_transform_phe_noswap
-
- movl 0*4(%rsi), %eax
- movl 1*4(%rsi), %ecx
- movl 2*4(%rsi), %edx
- movl 3*4(%rsi), %r9d
- bswapl %eax
- bswapl %ecx
- bswapl %edx
- bswapl %r9d
- movl %eax, 0*4(%rsp)
- movl %ecx, 1*4(%rsp)
- movl %edx, 2*4(%rsp)
- movl %r9d, 3*4(%rsp)
- movl 4*4(%rsi), %eax
- movl 5*4(%rsi), %ecx
- movl 6*4(%rsi), %edx
- movl 7*4(%rsi), %r9d
- bswapl %eax
- bswapl %ecx
- bswapl %edx
- bswapl %r9d
- movl %eax, 4*4(%rsp)
- movl %ecx, 5*4(%rsp)
- movl %edx, 6*4(%rsp)
- movl %r9d, 7*4(%rsp)
-
- movdqu 2*16(%rsi), %xmm0
- movdqu 3*16(%rsi), %xmm2
- pshuflw $0xb1, %xmm0, %xmm0
- pshuflw $0xb1, %xmm2, %xmm2
- pshufhw $0xb1, %xmm0, %xmm0
- pshufhw $0xb1, %xmm2, %xmm2
- movdqa %xmm0, %xmm1
- movdqa %xmm2, %xmm3
- psrlw $8, %xmm1
- psrlw $8, %xmm3
- psllw $8, %xmm0
- psllw $8, %xmm2
- pxor %xmm1, %xmm0
- pxor %xmm3, %xmm2
- movdqa %xmm0, 2*16(%rsp)
- movdqa %xmm2, 3*16(%rsp)
-
- jmp sha256_transform_phe_core
-
-sha256_transform_phe_noswap:
- movdqu 0*16(%rsi), %xmm0
- movdqu 1*16(%rsi), %xmm1
- movdqu 2*16(%rsi), %xmm2
- movdqu 3*16(%rsi), %xmm3
- movdqa %xmm0, 0*16(%rsp)
- movdqa %xmm1, 1*16(%rsp)
- movdqa %xmm2, 2*16(%rsp)
- movdqa %xmm3, 3*16(%rsp)
-
-sha256_transform_phe_core:
- movq %rsp, %rsi
- movq $-1, %rax
- movq $1, %rcx
- /* rep xsha256 */
- .byte 0xf3, 0x0f, 0xa6, 0xd0
-
- movq %r8, %rsp
-#if defined(_WIN64) || defined(__CYGWIN__)
- popq %rsi
- popq %rdi
-#endif
- ret
-
-
- .data
- .p2align 3
-sha256_transform_addr:
- .quad sha256_transform_sse2
-
- .text
- .p2align 3
- .globl sha256_transform
- .globl _sha256_transform
-sha256_transform:
-_sha256_transform:
- jmp *sha256_transform_addr(%rip)
-
-
- .data
- .p2align 7
-sha256_4h:
- .long 0x6a09e667, 0x6a09e667, 0x6a09e667, 0x6a09e667
- .long 0xbb67ae85, 0xbb67ae85, 0xbb67ae85, 0xbb67ae85
- .long 0x3c6ef372, 0x3c6ef372, 0x3c6ef372, 0x3c6ef372
- .long 0xa54ff53a, 0xa54ff53a, 0xa54ff53a, 0xa54ff53a
- .long 0x510e527f, 0x510e527f, 0x510e527f, 0x510e527f
- .long 0x9b05688c, 0x9b05688c, 0x9b05688c, 0x9b05688c
- .long 0x1f83d9ab, 0x1f83d9ab, 0x1f83d9ab, 0x1f83d9ab
- .long 0x5be0cd19, 0x5be0cd19, 0x5be0cd19, 0x5be0cd19
-
- .data
- .p2align 7
-sha256_4k:
- .long 0x428a2f98, 0x428a2f98, 0x428a2f98, 0x428a2f98
- .long 0x71374491, 0x71374491, 0x71374491, 0x71374491
- .long 0xb5c0fbcf, 0xb5c0fbcf, 0xb5c0fbcf, 0xb5c0fbcf
- .long 0xe9b5dba5, 0xe9b5dba5, 0xe9b5dba5, 0xe9b5dba5
- .long 0x3956c25b, 0x3956c25b, 0x3956c25b, 0x3956c25b
- .long 0x59f111f1, 0x59f111f1, 0x59f111f1, 0x59f111f1
- .long 0x923f82a4, 0x923f82a4, 0x923f82a4, 0x923f82a4
- .long 0xab1c5ed5, 0xab1c5ed5, 0xab1c5ed5, 0xab1c5ed5
- .long 0xd807aa98, 0xd807aa98, 0xd807aa98, 0xd807aa98
- .long 0x12835b01, 0x12835b01, 0x12835b01, 0x12835b01
- .long 0x243185be, 0x243185be, 0x243185be, 0x243185be
- .long 0x550c7dc3, 0x550c7dc3, 0x550c7dc3, 0x550c7dc3
- .long 0x72be5d74, 0x72be5d74, 0x72be5d74, 0x72be5d74
- .long 0x80deb1fe, 0x80deb1fe, 0x80deb1fe, 0x80deb1fe
- .long 0x9bdc06a7, 0x9bdc06a7, 0x9bdc06a7, 0x9bdc06a7
- .long 0xc19bf174, 0xc19bf174, 0xc19bf174, 0xc19bf174
- .long 0xe49b69c1, 0xe49b69c1, 0xe49b69c1, 0xe49b69c1
- .long 0xefbe4786, 0xefbe4786, 0xefbe4786, 0xefbe4786
- .long 0x0fc19dc6, 0x0fc19dc6, 0x0fc19dc6, 0x0fc19dc6
- .long 0x240ca1cc, 0x240ca1cc, 0x240ca1cc, 0x240ca1cc
- .long 0x2de92c6f, 0x2de92c6f, 0x2de92c6f, 0x2de92c6f
- .long 0x4a7484aa, 0x4a7484aa, 0x4a7484aa, 0x4a7484aa
- .long 0x5cb0a9dc, 0x5cb0a9dc, 0x5cb0a9dc, 0x5cb0a9dc
- .long 0x76f988da, 0x76f988da, 0x76f988da, 0x76f988da
- .long 0x983e5152, 0x983e5152, 0x983e5152, 0x983e5152
- .long 0xa831c66d, 0xa831c66d, 0xa831c66d, 0xa831c66d
- .long 0xb00327c8, 0xb00327c8, 0xb00327c8, 0xb00327c8
- .long 0xbf597fc7, 0xbf597fc7, 0xbf597fc7, 0xbf597fc7
- .long 0xc6e00bf3, 0xc6e00bf3, 0xc6e00bf3, 0xc6e00bf3
- .long 0xd5a79147, 0xd5a79147, 0xd5a79147, 0xd5a79147
- .long 0x06ca6351, 0x06ca6351, 0x06ca6351, 0x06ca6351
- .long 0x14292967, 0x14292967, 0x14292967, 0x14292967
- .long 0x27b70a85, 0x27b70a85, 0x27b70a85, 0x27b70a85
- .long 0x2e1b2138, 0x2e1b2138, 0x2e1b2138, 0x2e1b2138
- .long 0x4d2c6dfc, 0x4d2c6dfc, 0x4d2c6dfc, 0x4d2c6dfc
- .long 0x53380d13, 0x53380d13, 0x53380d13, 0x53380d13
- .long 0x650a7354, 0x650a7354, 0x650a7354, 0x650a7354
- .long 0x766a0abb, 0x766a0abb, 0x766a0abb, 0x766a0abb
- .long 0x81c2c92e, 0x81c2c92e, 0x81c2c92e, 0x81c2c92e
- .long 0x92722c85, 0x92722c85, 0x92722c85, 0x92722c85
- .long 0xa2bfe8a1, 0xa2bfe8a1, 0xa2bfe8a1, 0xa2bfe8a1
- .long 0xa81a664b, 0xa81a664b, 0xa81a664b, 0xa81a664b
- .long 0xc24b8b70, 0xc24b8b70, 0xc24b8b70, 0xc24b8b70
- .long 0xc76c51a3, 0xc76c51a3, 0xc76c51a3, 0xc76c51a3
- .long 0xd192e819, 0xd192e819, 0xd192e819, 0xd192e819
- .long 0xd6990624, 0xd6990624, 0xd6990624, 0xd6990624
- .long 0xf40e3585, 0xf40e3585, 0xf40e3585, 0xf40e3585
- .long 0x106aa070, 0x106aa070, 0x106aa070, 0x106aa070
- .long 0x19a4c116, 0x19a4c116, 0x19a4c116, 0x19a4c116
- .long 0x1e376c08, 0x1e376c08, 0x1e376c08, 0x1e376c08
- .long 0x2748774c, 0x2748774c, 0x2748774c, 0x2748774c
- .long 0x34b0bcb5, 0x34b0bcb5, 0x34b0bcb5, 0x34b0bcb5
- .long 0x391c0cb3, 0x391c0cb3, 0x391c0cb3, 0x391c0cb3
- .long 0x4ed8aa4a, 0x4ed8aa4a, 0x4ed8aa4a, 0x4ed8aa4a
- .long 0x5b9cca4f, 0x5b9cca4f, 0x5b9cca4f, 0x5b9cca4f
- .long 0x682e6ff3, 0x682e6ff3, 0x682e6ff3, 0x682e6ff3
- .long 0x748f82ee, 0x748f82ee, 0x748f82ee, 0x748f82ee
- .long 0x78a5636f, 0x78a5636f, 0x78a5636f, 0x78a5636f
- .long 0x84c87814, 0x84c87814, 0x84c87814, 0x84c87814
- .long 0x8cc70208, 0x8cc70208, 0x8cc70208, 0x8cc70208
- .long 0x90befffa, 0x90befffa, 0x90befffa, 0x90befffa
- .long 0xa4506ceb, 0xa4506ceb, 0xa4506ceb, 0xa4506ceb
- .long 0xbef9a3f7, 0xbef9a3f7, 0xbef9a3f7, 0xbef9a3f7
- .long 0xc67178f2, 0xc67178f2, 0xc67178f2, 0xc67178f2
-
- .data
- .p2align 7
-sha256_8h:
- .long 0x6a09e667, 0x6a09e667, 0x6a09e667, 0x6a09e667, 0x6a09e667, 0x6a09e667, 0x6a09e667, 0x6a09e667
- .long 0xbb67ae85, 0xbb67ae85, 0xbb67ae85, 0xbb67ae85, 0xbb67ae85, 0xbb67ae85, 0xbb67ae85, 0xbb67ae85
- .long 0x3c6ef372, 0x3c6ef372, 0x3c6ef372, 0x3c6ef372, 0x3c6ef372, 0x3c6ef372, 0x3c6ef372, 0x3c6ef372
- .long 0xa54ff53a, 0xa54ff53a, 0xa54ff53a, 0xa54ff53a, 0xa54ff53a, 0xa54ff53a, 0xa54ff53a, 0xa54ff53a
- .long 0x510e527f, 0x510e527f, 0x510e527f, 0x510e527f, 0x510e527f, 0x510e527f, 0x510e527f, 0x510e527f
- .long 0x9b05688c, 0x9b05688c, 0x9b05688c, 0x9b05688c, 0x9b05688c, 0x9b05688c, 0x9b05688c, 0x9b05688c
- .long 0x1f83d9ab, 0x1f83d9ab, 0x1f83d9ab, 0x1f83d9ab, 0x1f83d9ab, 0x1f83d9ab, 0x1f83d9ab, 0x1f83d9ab
- .long 0x5be0cd19, 0x5be0cd19, 0x5be0cd19, 0x5be0cd19, 0x5be0cd19, 0x5be0cd19, 0x5be0cd19, 0x5be0cd19
-
- .data
- .p2align 7
-sha256_8k:
- .long 0x428a2f98, 0x428a2f98, 0x428a2f98, 0x428a2f98, 0x428a2f98, 0x428a2f98, 0x428a2f98, 0x428a2f98
- .long 0x71374491, 0x71374491, 0x71374491, 0x71374491, 0x71374491, 0x71374491, 0x71374491, 0x71374491
- .long 0xb5c0fbcf, 0xb5c0fbcf, 0xb5c0fbcf, 0xb5c0fbcf, 0xb5c0fbcf, 0xb5c0fbcf, 0xb5c0fbcf, 0xb5c0fbcf
- .long 0xe9b5dba5, 0xe9b5dba5, 0xe9b5dba5, 0xe9b5dba5, 0xe9b5dba5, 0xe9b5dba5, 0xe9b5dba5, 0xe9b5dba5
- .long 0x3956c25b, 0x3956c25b, 0x3956c25b, 0x3956c25b, 0x3956c25b, 0x3956c25b, 0x3956c25b, 0x3956c25b
- .long 0x59f111f1, 0x59f111f1, 0x59f111f1, 0x59f111f1, 0x59f111f1, 0x59f111f1, 0x59f111f1, 0x59f111f1
- .long 0x923f82a4, 0x923f82a4, 0x923f82a4, 0x923f82a4, 0x923f82a4, 0x923f82a4, 0x923f82a4, 0x923f82a4
- .long 0xab1c5ed5, 0xab1c5ed5, 0xab1c5ed5, 0xab1c5ed5, 0xab1c5ed5, 0xab1c5ed5, 0xab1c5ed5, 0xab1c5ed5
- .long 0xd807aa98, 0xd807aa98, 0xd807aa98, 0xd807aa98, 0xd807aa98, 0xd807aa98, 0xd807aa98, 0xd807aa98
- .long 0x12835b01, 0x12835b01, 0x12835b01, 0x12835b01, 0x12835b01, 0x12835b01, 0x12835b01, 0x12835b01
- .long 0x243185be, 0x243185be, 0x243185be, 0x243185be, 0x243185be, 0x243185be, 0x243185be, 0x243185be
- .long 0x550c7dc3, 0x550c7dc3, 0x550c7dc3, 0x550c7dc3, 0x550c7dc3, 0x550c7dc3, 0x550c7dc3, 0x550c7dc3
- .long 0x72be5d74, 0x72be5d74, 0x72be5d74, 0x72be5d74, 0x72be5d74, 0x72be5d74, 0x72be5d74, 0x72be5d74
- .long 0x80deb1fe, 0x80deb1fe, 0x80deb1fe, 0x80deb1fe, 0x80deb1fe, 0x80deb1fe, 0x80deb1fe, 0x80deb1fe
- .long 0x9bdc06a7, 0x9bdc06a7, 0x9bdc06a7, 0x9bdc06a7, 0x9bdc06a7, 0x9bdc06a7, 0x9bdc06a7, 0x9bdc06a7
- .long 0xc19bf174, 0xc19bf174, 0xc19bf174, 0xc19bf174, 0xc19bf174, 0xc19bf174, 0xc19bf174, 0xc19bf174
- .long 0xe49b69c1, 0xe49b69c1, 0xe49b69c1, 0xe49b69c1, 0xe49b69c1, 0xe49b69c1, 0xe49b69c1, 0xe49b69c1
- .long 0xefbe4786, 0xefbe4786, 0xefbe4786, 0xefbe4786, 0xefbe4786, 0xefbe4786, 0xefbe4786, 0xefbe4786
- .long 0x0fc19dc6, 0x0fc19dc6, 0x0fc19dc6, 0x0fc19dc6, 0x0fc19dc6, 0x0fc19dc6, 0x0fc19dc6, 0x0fc19dc6
- .long 0x240ca1cc, 0x240ca1cc, 0x240ca1cc, 0x240ca1cc, 0x240ca1cc, 0x240ca1cc, 0x240ca1cc, 0x240ca1cc
- .long 0x2de92c6f, 0x2de92c6f, 0x2de92c6f, 0x2de92c6f, 0x2de92c6f, 0x2de92c6f, 0x2de92c6f, 0x2de92c6f
- .long 0x4a7484aa, 0x4a7484aa, 0x4a7484aa, 0x4a7484aa, 0x4a7484aa, 0x4a7484aa, 0x4a7484aa, 0x4a7484aa
- .long 0x5cb0a9dc, 0x5cb0a9dc, 0x5cb0a9dc, 0x5cb0a9dc, 0x5cb0a9dc, 0x5cb0a9dc, 0x5cb0a9dc, 0x5cb0a9dc
- .long 0x76f988da, 0x76f988da, 0x76f988da, 0x76f988da, 0x76f988da, 0x76f988da, 0x76f988da, 0x76f988da
- .long 0x983e5152, 0x983e5152, 0x983e5152, 0x983e5152, 0x983e5152, 0x983e5152, 0x983e5152, 0x983e5152
- .long 0xa831c66d, 0xa831c66d, 0xa831c66d, 0xa831c66d, 0xa831c66d, 0xa831c66d, 0xa831c66d, 0xa831c66d
- .long 0xb00327c8, 0xb00327c8, 0xb00327c8, 0xb00327c8, 0xb00327c8, 0xb00327c8, 0xb00327c8, 0xb00327c8
- .long 0xbf597fc7, 0xbf597fc7, 0xbf597fc7, 0xbf597fc7, 0xbf597fc7, 0xbf597fc7, 0xbf597fc7, 0xbf597fc7
- .long 0xc6e00bf3, 0xc6e00bf3, 0xc6e00bf3, 0xc6e00bf3, 0xc6e00bf3, 0xc6e00bf3, 0xc6e00bf3, 0xc6e00bf3
- .long 0xd5a79147, 0xd5a79147, 0xd5a79147, 0xd5a79147, 0xd5a79147, 0xd5a79147, 0xd5a79147, 0xd5a79147
- .long 0x06ca6351, 0x06ca6351, 0x06ca6351, 0x06ca6351, 0x06ca6351, 0x06ca6351, 0x06ca6351, 0x06ca6351
- .long 0x14292967, 0x14292967, 0x14292967, 0x14292967, 0x14292967, 0x14292967, 0x14292967, 0x14292967
- .long 0x27b70a85, 0x27b70a85, 0x27b70a85, 0x27b70a85, 0x27b70a85, 0x27b70a85, 0x27b70a85, 0x27b70a85
- .long 0x2e1b2138, 0x2e1b2138, 0x2e1b2138, 0x2e1b2138, 0x2e1b2138, 0x2e1b2138, 0x2e1b2138, 0x2e1b2138
- .long 0x4d2c6dfc, 0x4d2c6dfc, 0x4d2c6dfc, 0x4d2c6dfc, 0x4d2c6dfc, 0x4d2c6dfc, 0x4d2c6dfc, 0x4d2c6dfc
- .long 0x53380d13, 0x53380d13, 0x53380d13, 0x53380d13, 0x53380d13, 0x53380d13, 0x53380d13, 0x53380d13
- .long 0x650a7354, 0x650a7354, 0x650a7354, 0x650a7354, 0x650a7354, 0x650a7354, 0x650a7354, 0x650a7354
- .long 0x766a0abb, 0x766a0abb, 0x766a0abb, 0x766a0abb, 0x766a0abb, 0x766a0abb, 0x766a0abb, 0x766a0abb
- .long 0x81c2c92e, 0x81c2c92e, 0x81c2c92e, 0x81c2c92e, 0x81c2c92e, 0x81c2c92e, 0x81c2c92e, 0x81c2c92e
- .long 0x92722c85, 0x92722c85, 0x92722c85, 0x92722c85, 0x92722c85, 0x92722c85, 0x92722c85, 0x92722c85
- .long 0xa2bfe8a1, 0xa2bfe8a1, 0xa2bfe8a1, 0xa2bfe8a1, 0xa2bfe8a1, 0xa2bfe8a1, 0xa2bfe8a1, 0xa2bfe8a1
- .long 0xa81a664b, 0xa81a664b, 0xa81a664b, 0xa81a664b, 0xa81a664b, 0xa81a664b, 0xa81a664b, 0xa81a664b
- .long 0xc24b8b70, 0xc24b8b70, 0xc24b8b70, 0xc24b8b70, 0xc24b8b70, 0xc24b8b70, 0xc24b8b70, 0xc24b8b70
- .long 0xc76c51a3, 0xc76c51a3, 0xc76c51a3, 0xc76c51a3, 0xc76c51a3, 0xc76c51a3, 0xc76c51a3, 0xc76c51a3
- .long 0xd192e819, 0xd192e819, 0xd192e819, 0xd192e819, 0xd192e819, 0xd192e819, 0xd192e819, 0xd192e819
- .long 0xd6990624, 0xd6990624, 0xd6990624, 0xd6990624, 0xd6990624, 0xd6990624, 0xd6990624, 0xd6990624
- .long 0xf40e3585, 0xf40e3585, 0xf40e3585, 0xf40e3585, 0xf40e3585, 0xf40e3585, 0xf40e3585, 0xf40e3585
- .long 0x106aa070, 0x106aa070, 0x106aa070, 0x106aa070, 0x106aa070, 0x106aa070, 0x106aa070, 0x106aa070
- .long 0x19a4c116, 0x19a4c116, 0x19a4c116, 0x19a4c116, 0x19a4c116, 0x19a4c116, 0x19a4c116, 0x19a4c116
- .long 0x1e376c08, 0x1e376c08, 0x1e376c08, 0x1e376c08, 0x1e376c08, 0x1e376c08, 0x1e376c08, 0x1e376c08
- .long 0x2748774c, 0x2748774c, 0x2748774c, 0x2748774c, 0x2748774c, 0x2748774c, 0x2748774c, 0x2748774c
- .long 0x34b0bcb5, 0x34b0bcb5, 0x34b0bcb5, 0x34b0bcb5, 0x34b0bcb5, 0x34b0bcb5, 0x34b0bcb5, 0x34b0bcb5
- .long 0x391c0cb3, 0x391c0cb3, 0x391c0cb3, 0x391c0cb3, 0x391c0cb3, 0x391c0cb3, 0x391c0cb3, 0x391c0cb3
- .long 0x4ed8aa4a, 0x4ed8aa4a, 0x4ed8aa4a, 0x4ed8aa4a, 0x4ed8aa4a, 0x4ed8aa4a, 0x4ed8aa4a, 0x4ed8aa4a
- .long 0x5b9cca4f, 0x5b9cca4f, 0x5b9cca4f, 0x5b9cca4f, 0x5b9cca4f, 0x5b9cca4f, 0x5b9cca4f, 0x5b9cca4f
- .long 0x682e6ff3, 0x682e6ff3, 0x682e6ff3, 0x682e6ff3, 0x682e6ff3, 0x682e6ff3, 0x682e6ff3, 0x682e6ff3
- .long 0x748f82ee, 0x748f82ee, 0x748f82ee, 0x748f82ee, 0x748f82ee, 0x748f82ee, 0x748f82ee, 0x748f82ee
- .long 0x78a5636f, 0x78a5636f, 0x78a5636f, 0x78a5636f, 0x78a5636f, 0x78a5636f, 0x78a5636f, 0x78a5636f
- .long 0x84c87814, 0x84c87814, 0x84c87814, 0x84c87814, 0x84c87814, 0x84c87814, 0x84c87814, 0x84c87814
- .long 0x8cc70208, 0x8cc70208, 0x8cc70208, 0x8cc70208, 0x8cc70208, 0x8cc70208, 0x8cc70208, 0x8cc70208
- .long 0x90befffa, 0x90befffa, 0x90befffa, 0x90befffa, 0x90befffa, 0x90befffa, 0x90befffa, 0x90befffa
- .long 0xa4506ceb, 0xa4506ceb, 0xa4506ceb, 0xa4506ceb, 0xa4506ceb, 0xa4506ceb, 0xa4506ceb, 0xa4506ceb
- .long 0xbef9a3f7, 0xbef9a3f7, 0xbef9a3f7, 0xbef9a3f7, 0xbef9a3f7, 0xbef9a3f7, 0xbef9a3f7, 0xbef9a3f7
- .long 0xc67178f2, 0xc67178f2, 0xc67178f2, 0xc67178f2, 0xc67178f2, 0xc67178f2, 0xc67178f2, 0xc67178f2
-
- .text
- .p2align 6
- .globl sha256_init_4way
- .globl _sha256_init_4way
-sha256_init_4way:
-_sha256_init_4way:
-#if defined(_WIN64) || defined(__CYGWIN__)
- pushq %rdi
- movq %rcx, %rdi
-#endif
- movdqa sha256_4h+0(%rip), %xmm0
- movdqa sha256_4h+16(%rip), %xmm1
- movdqa sha256_4h+32(%rip), %xmm2
- movdqa sha256_4h+48(%rip), %xmm3
- movdqu %xmm0, 0(%rdi)
- movdqu %xmm1, 16(%rdi)
- movdqu %xmm2, 32(%rdi)
- movdqu %xmm3, 48(%rdi)
- movdqa sha256_4h+64(%rip), %xmm0
- movdqa sha256_4h+80(%rip), %xmm1
- movdqa sha256_4h+96(%rip), %xmm2
- movdqa sha256_4h+112(%rip), %xmm3
- movdqu %xmm0, 64(%rdi)
- movdqu %xmm1, 80(%rdi)
- movdqu %xmm2, 96(%rdi)
- movdqu %xmm3, 112(%rdi)
-#if defined(_WIN64) || defined(__CYGWIN__)
- popq %rdi
-#endif
- ret
-
- .text
- .p2align 6
- .globl sha256_init_8way
- .globl _sha256_init_8way
-sha256_init_8way:
-_sha256_init_8way:
-#if defined(_WIN64) || defined(__CYGWIN__)
- pushq %rdi
- movq %rcx, %rdi
-#endif
- vpbroadcastd sha256_4h+0(%rip), %ymm0
- vpbroadcastd sha256_4h+16(%rip), %ymm1
- vpbroadcastd sha256_4h+32(%rip), %ymm2
- vpbroadcastd sha256_4h+48(%rip), %ymm3
- vmovdqu %ymm0, 0*32(%rdi)
- vmovdqu %ymm1, 1*32(%rdi)
- vmovdqu %ymm2, 2*32(%rdi)
- vmovdqu %ymm3, 3*32(%rdi)
- vpbroadcastd sha256_4h+64(%rip), %ymm0
- vpbroadcastd sha256_4h+80(%rip), %ymm1
- vpbroadcastd sha256_4h+96(%rip), %ymm2
- vpbroadcastd sha256_4h+112(%rip), %ymm3
- vmovdqu %ymm0, 4*32(%rdi)
- vmovdqu %ymm1, 5*32(%rdi)
- vmovdqu %ymm2, 6*32(%rdi)
- vmovdqu %ymm3, 7*32(%rdi)
-#if defined(_WIN64) || defined(__CYGWIN__)
- popq %rdi
-#endif
- ret
-
-.macro sha256_sse2_extend_round i
- movdqa (\i-15)*16(%rax), %xmm0
- movdqa %xmm0, %xmm2
- psrld $3, %xmm0
- movdqa %xmm0, %xmm1
- pslld $14, %xmm2
- psrld $4, %xmm1
- pxor %xmm1, %xmm0
- pxor %xmm2, %xmm0
- psrld $11, %xmm1
- pslld $11, %xmm2
- pxor %xmm1, %xmm0
- pxor %xmm2, %xmm0
- paddd (\i-16)*16(%rax), %xmm0
- paddd (\i-7)*16(%rax), %xmm0
-
- movdqa %xmm3, %xmm2
- psrld $10, %xmm3
- pslld $13, %xmm2
- movdqa %xmm3, %xmm1
- psrld $7, %xmm1
- pxor %xmm1, %xmm3
- pxor %xmm2, %xmm3
- psrld $2, %xmm1
- pslld $2, %xmm2
- pxor %xmm1, %xmm3
- pxor %xmm2, %xmm3
- paddd %xmm0, %xmm3
- movdqa %xmm3, \i*16(%rax)
-.endm
-
-.macro sha256_sse2_extend_doubleround i
- movdqa (\i-15)*16(%rax), %xmm0
- movdqa (\i-14)*16(%rax), %xmm4
- movdqa %xmm0, %xmm2
- movdqa %xmm4, %xmm6
- psrld $3, %xmm0
- psrld $3, %xmm4
- movdqa %xmm0, %xmm1
- movdqa %xmm4, %xmm5
- pslld $14, %xmm2
- pslld $14, %xmm6
- psrld $4, %xmm1
- psrld $4, %xmm5
- pxor %xmm1, %xmm0
- pxor %xmm5, %xmm4
- psrld $11, %xmm1
- psrld $11, %xmm5
- pxor %xmm2, %xmm0
- pxor %xmm6, %xmm4
- pslld $11, %xmm2
- pslld $11, %xmm6
- pxor %xmm1, %xmm0
- pxor %xmm5, %xmm4
- pxor %xmm2, %xmm0
- pxor %xmm6, %xmm4
-
- paddd (\i-16)*16(%rax), %xmm0
- paddd (\i-15)*16(%rax), %xmm4
-
- movdqa %xmm3, %xmm2
- movdqa %xmm7, %xmm6
- psrld $10, %xmm3
- psrld $10, %xmm7
- movdqa %xmm3, %xmm1
- movdqa %xmm7, %xmm5
- pslld $13, %xmm2
- pslld $13, %xmm6
- psrld $7, %xmm1
- psrld $7, %xmm5
-
- paddd (\i-7)*16(%rax), %xmm0
- paddd (\i-6)*16(%rax), %xmm4
-
- pxor %xmm1, %xmm3
- pxor %xmm5, %xmm7
- psrld $2, %xmm1
- psrld $2, %xmm5
- pxor %xmm2, %xmm3
- pxor %xmm6, %xmm7
- pslld $2, %xmm2
- pslld $2, %xmm6
- pxor %xmm1, %xmm3
- pxor %xmm5, %xmm7
- pxor %xmm2, %xmm3
- pxor %xmm6, %xmm7
-
- paddd %xmm0, %xmm3
- paddd %xmm4, %xmm7
- movdqa %xmm3, \i*16(%rax)
- movdqa %xmm7, (\i+1)*16(%rax)
-.endm
-
-.macro sha256_sse2_main_round i
- movdqa 16*(\i)(%rax), %xmm6
-
- movdqa %xmm0, %xmm1
- movdqa 16(%rsp), %xmm2
- pandn %xmm2, %xmm1
- paddd 32(%rsp), %xmm6
-
- movdqa %xmm2, 32(%rsp)
- movdqa 0(%rsp), %xmm2
- movdqa %xmm2, 16(%rsp)
-
- pand %xmm0, %xmm2
- pxor %xmm2, %xmm1
- movdqa %xmm0, 0(%rsp)
-
- paddd %xmm1, %xmm6
-
- movdqa %xmm0, %xmm1
- psrld $6, %xmm0
- paddd 16*(\i)(%rcx), %xmm6
- movdqa %xmm0, %xmm2
- pslld $7, %xmm1
- psrld $5, %xmm2
- pxor %xmm1, %xmm0
- pxor %xmm2, %xmm0
- pslld $14, %xmm1
- psrld $14, %xmm2
- pxor %xmm1, %xmm0
- pslld $5, %xmm1
- pxor %xmm2, %xmm0
- pxor %xmm1, %xmm0
- movdqa %xmm5, %xmm1
- paddd %xmm0, %xmm6
-
- movdqa %xmm3, %xmm0
- movdqa %xmm4, %xmm3
- movdqa %xmm4, %xmm2
- paddd %xmm6, %xmm0
- pand %xmm5, %xmm2
- pand %xmm7, %xmm1
- pand %xmm7, %xmm4
- pxor %xmm4, %xmm1
- movdqa %xmm5, %xmm4
- movdqa %xmm7, %xmm5
- pxor %xmm2, %xmm1
- paddd %xmm1, %xmm6
-
- movdqa %xmm7, %xmm2
- psrld $2, %xmm7
- movdqa %xmm7, %xmm1
- pslld $10, %xmm2
- psrld $11, %xmm1
- pxor %xmm2, %xmm7
- pslld $9, %xmm2
- pxor %xmm1, %xmm7
- psrld $9, %xmm1
- pxor %xmm2, %xmm7
- pslld $11, %xmm2
- pxor %xmm1, %xmm7
- pxor %xmm2, %xmm7
- paddd %xmm6, %xmm7
-.endm
-
-.macro sha256_sse2_main_quadround i
- sha256_sse2_main_round \i+0
- sha256_sse2_main_round \i+1
- sha256_sse2_main_round \i+2
- sha256_sse2_main_round \i+3
-.endm
-
-
-.macro sha256_avx_extend_round i
- vmovdqa (\i-15)*16(%rax), %xmm0
- vpslld $14, %xmm0, %xmm2
- vpsrld $3, %xmm0, %xmm0
- vpsrld $4, %xmm0, %xmm1
- vpxor %xmm1, %xmm0, %xmm0
- vpxor %xmm2, %xmm0, %xmm0
- vpsrld $11, %xmm1, %xmm1
- vpslld $11, %xmm2, %xmm2
- vpxor %xmm1, %xmm0, %xmm0
- vpxor %xmm2, %xmm0, %xmm0
- vpaddd (\i-16)*16(%rax), %xmm0, %xmm0
- vpaddd (\i-7)*16(%rax), %xmm0, %xmm0
-
- vpslld $13, %xmm3, %xmm2
- vpsrld $10, %xmm3, %xmm3
- vpsrld $7, %xmm3, %xmm1
- vpxor %xmm1, %xmm3, %xmm3
- vpxor %xmm2, %xmm3, %xmm3
- vpsrld $2, %xmm1, %xmm1
- vpslld $2, %xmm2, %xmm2
- vpxor %xmm1, %xmm3, %xmm3
- vpxor %xmm2, %xmm3, %xmm3
- vpaddd %xmm0, %xmm3, %xmm3
- vmovdqa %xmm3, \i*16(%rax)
-.endm
-
-.macro sha256_avx_extend_doubleround i
- vmovdqa (\i-15)*16(%rax), %xmm0
- vmovdqa (\i-14)*16(%rax), %xmm4
- vpslld $14, %xmm0, %xmm2
- vpslld $14, %xmm4, %xmm6
- vpsrld $3, %xmm0, %xmm8
- vpsrld $3, %xmm4, %xmm4
- vpsrld $7, %xmm0, %xmm1
- vpsrld $4, %xmm4, %xmm5
- vpxor %xmm1, %xmm8, %xmm8
- vpxor %xmm5, %xmm4, %xmm4
- vpsrld $11, %xmm1, %xmm1
- vpsrld $11, %xmm5, %xmm5
- vpxor %xmm2, %xmm8, %xmm8
- vpxor %xmm6, %xmm4, %xmm4
- vpslld $11, %xmm2, %xmm2
- vpslld $11, %xmm6, %xmm6
- vpxor %xmm1, %xmm8, %xmm8
- vpxor %xmm5, %xmm4, %xmm4
- vpxor %xmm2, %xmm8, %xmm8
- vpxor %xmm6, %xmm4, %xmm4
-
- vpaddd %xmm0, %xmm4, %xmm4
- vpaddd (\i-16)*16(%rax), %xmm8, %xmm0
-
- vpslld $13, %xmm3, %xmm2
- vpslld $13, %xmm7, %xmm6
- vpsrld $10, %xmm3, %xmm3
- vpsrld $10, %xmm7, %xmm7
-
- vpaddd (\i-7)*16(%rax), %xmm0, %xmm0
- vpaddd (\i-6)*16(%rax), %xmm4, %xmm4
-
- vpsrld $7, %xmm3, %xmm1
- vpsrld $7, %xmm7, %xmm5
- vpxor %xmm1, %xmm3, %xmm3
- vpxor %xmm5, %xmm7, %xmm7
- vpsrld $2, %xmm1, %xmm1
- vpsrld $2, %xmm5, %xmm5
- vpxor %xmm2, %xmm3, %xmm3
- vpxor %xmm6, %xmm7, %xmm7
- vpslld $2, %xmm2, %xmm2
- vpslld $2, %xmm6, %xmm6
- vpxor %xmm1, %xmm3, %xmm3
- vpxor %xmm5, %xmm7, %xmm7
- vpxor %xmm2, %xmm3, %xmm3
- vpxor %xmm6, %xmm7, %xmm7
-
- vpaddd %xmm0, %xmm3, %xmm3
- vpaddd %xmm4, %xmm7, %xmm7
- vmovdqa %xmm3, \i*16(%rax)
- vmovdqa %xmm7, (\i+1)*16(%rax)
-.endm
-
-.macro sha256_avx_main_round i, r0, r1, r2, r3, r4, r5, r6, r7
- vpaddd 16*(\i)(%rax), \r0, %xmm6
- vpaddd 16*(\i)(%rcx), %xmm6, %xmm6
-
- vpandn \r1, \r3, %xmm1
- vpand \r3, \r2, %xmm2
- vpxor %xmm2, %xmm1, %xmm1
- vpaddd %xmm1, %xmm6, %xmm6
-
- vpslld $7, \r3, %xmm1
- vpsrld $6, \r3, \r0
- vpsrld $5, \r0, %xmm2
- vpxor %xmm1, \r0, \r0
- vpxor %xmm2, \r0, \r0
- vpslld $14, %xmm1, %xmm1
- vpsrld $14, %xmm2, %xmm2
- vpxor %xmm1, \r0, \r0
- vpxor %xmm2, \r0, \r0
- vpslld $5, %xmm1, %xmm1
- vpxor %xmm1, \r0, \r0
- vpaddd \r0, %xmm6, %xmm6
- vpaddd %xmm6, \r4, \r0
-
- vpand \r6, \r5, %xmm2
- vpand \r7, \r5, \r4
- vpand \r7, \r6, %xmm1
- vpxor \r4, %xmm1, %xmm1
- vpxor %xmm2, %xmm1, %xmm1
- vpaddd %xmm1, %xmm6, %xmm6
-
- vpslld $10, \r7, %xmm2
- vpsrld $2, \r7, \r4
- vpsrld $11, \r4, %xmm1
- vpxor %xmm2, \r4, \r4
- vpxor %xmm1, \r4, \r4
- vpslld $9, %xmm2, %xmm2
- vpsrld $9, %xmm1, %xmm1
- vpxor %xmm2, \r4, \r4
- vpxor %xmm1, \r4, \r4
- vpslld $11, %xmm2, %xmm2
- vpxor %xmm2, \r4, \r4
- vpaddd %xmm6, \r4, \r4
-.endm
-
-.macro sha256_avx_main_quadround i
- sha256_avx_main_round \i+0, %xmm10, %xmm9, %xmm8, %xmm0, %xmm3, %xmm4, %xmm5, %xmm7
- sha256_avx_main_round \i+1, %xmm9, %xmm8, %xmm0, %xmm10, %xmm4, %xmm5, %xmm7, %xmm3
- sha256_avx_main_round \i+2, %xmm8, %xmm0, %xmm10, %xmm9, %xmm5, %xmm7, %xmm3, %xmm4
- sha256_avx_main_round \i+3, %xmm0, %xmm10, %xmm9, %xmm8, %xmm7, %xmm3, %xmm4, %xmm5
-.endm
-
-
-.macro sha256_avx2_extend_round i
- vmovdqa (\i-15)*32(%rax), %ymm0
- vpslld $14, %ymm0, %ymm2
- vpsrld $3, %ymm0, %ymm0
- vpsrld $4, %ymm0, %ymm1
- vpxor %ymm1, %ymm0, %ymm0
- vpxor %ymm2, %ymm0, %ymm0
- vpsrld $11, %ymm1, %ymm1
- vpslld $11, %ymm2, %ymm2
- vpxor %ymm1, %ymm0, %ymm0
- vpxor %ymm2, %ymm0, %ymm0
- vpaddd (\i-16)*32(%rax), %ymm0, %ymm0
- vpaddd (\i-7)*32(%rax), %ymm0, %ymm0
-
- vpslld $13, %ymm3, %ymm2
- vpsrld $10, %ymm3, %ymm3
- vpsrld $7, %ymm3, %ymm1
- vpxor %ymm1, %ymm3, %ymm3
- vpxor %ymm2, %ymm3, %ymm3
- vpsrld $2, %ymm1, %ymm1
- vpslld $2, %ymm2, %ymm2
- vpxor %ymm1, %ymm3, %ymm3
- vpxor %ymm2, %ymm3, %ymm3
- vpaddd %ymm0, %ymm3, %ymm3
- vmovdqa %ymm3, \i*32(%rax)
-.endm
-
-.macro sha256_avx2_extend_doubleround i
- vmovdqa (\i-15)*32(%rax), %ymm0
- vmovdqa (\i-14)*32(%rax), %ymm4
- vpslld $14, %ymm0, %ymm2
- vpslld $14, %ymm4, %ymm6
- vpsrld $3, %ymm0, %ymm8
- vpsrld $3, %ymm4, %ymm4
- vpsrld $7, %ymm0, %ymm1
- vpsrld $4, %ymm4, %ymm5
- vpxor %ymm1, %ymm8, %ymm8
- vpxor %ymm5, %ymm4, %ymm4
- vpsrld $11, %ymm1, %ymm1
- vpsrld $11, %ymm5, %ymm5
- vpxor %ymm2, %ymm8, %ymm8
- vpxor %ymm6, %ymm4, %ymm4
- vpslld $11, %ymm2, %ymm2
- vpslld $11, %ymm6, %ymm6
- vpxor %ymm1, %ymm8, %ymm8
- vpxor %ymm5, %ymm4, %ymm4
- vpxor %ymm2, %ymm8, %ymm8
- vpxor %ymm6, %ymm4, %ymm4
-
- vpaddd %ymm0, %ymm4, %ymm4
- vpaddd (\i-16)*32(%rax), %ymm8, %ymm0
-
- vpslld $13, %ymm3, %ymm2
- vpslld $13, %ymm7, %ymm6
- vpsrld $10, %ymm3, %ymm3
- vpsrld $10, %ymm7, %ymm7
-
- vpaddd (\i-7)*32(%rax), %ymm0, %ymm0
- vpaddd (\i-6)*32(%rax), %ymm4, %ymm4
-
- vpsrld $7, %ymm3, %ymm1
- vpsrld $7, %ymm7, %ymm5
- vpxor %ymm1, %ymm3, %ymm3
- vpxor %ymm5, %ymm7, %ymm7
- vpsrld $2, %ymm1, %ymm1
- vpsrld $2, %ymm5, %ymm5
- vpxor %ymm2, %ymm3, %ymm3
- vpxor %ymm6, %ymm7, %ymm7
- vpslld $2, %ymm2, %ymm2
- vpslld $2, %ymm6, %ymm6
- vpxor %ymm1, %ymm3, %ymm3
- vpxor %ymm5, %ymm7, %ymm7
- vpxor %ymm2, %ymm3, %ymm3
- vpxor %ymm6, %ymm7, %ymm7
-
- vpaddd %ymm0, %ymm3, %ymm3
- vpaddd %ymm4, %ymm7, %ymm7
- vmovdqa %ymm3, \i*32(%rax)
- vmovdqa %ymm7, (\i+1)*32(%rax)
-.endm
-
-.macro sha256_avx2_main_round i, r0, r1, r2, r3, r4, r5, r6, r7
- vpaddd 32*(\i)(%rax), \r0, %ymm6
- vpaddd 32*(\i)(%rcx), %ymm6, %ymm6
-
- vpandn \r1, \r3, %ymm1
- vpand \r3, \r2, %ymm2
- vpxor %ymm2, %ymm1, %ymm1
- vpaddd %ymm1, %ymm6, %ymm6
-
- vpslld $7, \r3, %ymm1
- vpsrld $6, \r3, \r0
- vpsrld $5, \r0, %ymm2
- vpxor %ymm1, \r0, \r0
- vpxor %ymm2, \r0, \r0
- vpslld $14, %ymm1, %ymm1
- vpsrld $14, %ymm2, %ymm2
- vpxor %ymm1, \r0, \r0
- vpxor %ymm2, \r0, \r0
- vpslld $5, %ymm1, %ymm1
- vpxor %ymm1, \r0, \r0
- vpaddd \r0, %ymm6, %ymm6
- vpaddd %ymm6, \r4, \r0
-
- vpand \r6, \r5, %ymm2
- vpand \r7, \r5, \r4
- vpand \r7, \r6, %ymm1
- vpxor \r4, %ymm1, %ymm1
- vpxor %ymm2, %ymm1, %ymm1
- vpaddd %ymm1, %ymm6, %ymm6
-
- vpslld $10, \r7, %ymm2
- vpsrld $2, \r7, \r4
- vpsrld $11, \r4, %ymm1
- vpxor %ymm2, \r4, \r4
- vpxor %ymm1, \r4, \r4
- vpslld $9, %ymm2, %ymm2
- vpsrld $9, %ymm1, %ymm1
- vpxor %ymm2, \r4, \r4
- vpxor %ymm1, \r4, \r4
- vpslld $11, %ymm2, %ymm2
- vpxor %ymm2, \r4, \r4
- vpaddd %ymm6, \r4, \r4
-.endm
-
-.macro sha256_avx2_main_quadround i
- sha256_avx2_main_round \i+0, %ymm10, %ymm9, %ymm8, %ymm0, %ymm3, %ymm4, %ymm5, %ymm7
- sha256_avx2_main_round \i+1, %ymm9, %ymm8, %ymm0, %ymm10, %ymm4, %ymm5, %ymm7, %ymm3
- sha256_avx2_main_round \i+2, %ymm8, %ymm0, %ymm10, %ymm9, %ymm5, %ymm7, %ymm3, %ymm4
- sha256_avx2_main_round \i+3, %ymm0, %ymm10, %ymm9, %ymm8, %ymm7, %ymm3, %ymm4, %ymm5
-.endm
-
-.macro sha256_xop_extend_round i
- vmovdqa (\i-15)*16(%rax), %xmm0
- vprotd $25, %xmm0, %xmm1
- vprotd $14, %xmm0, %xmm2
- vpsrld $3, %xmm0, %xmm0
- vpxor %xmm1, %xmm2, %xmm2
- vpxor %xmm2, %xmm0, %xmm0
-
- vpaddd (\i-16)*16(%rax), %xmm0, %xmm0
- vpaddd (\i-7)*16(%rax), %xmm0, %xmm0
-
- vprotd $15, %xmm3, %xmm1
- vprotd $13, %xmm3, %xmm2
- vpsrld $10, %xmm3, %xmm3
- vpxor %xmm1, %xmm2, %xmm2
- vpxor %xmm2, %xmm3, %xmm3
- vpaddd %xmm0, %xmm3, %xmm3
- vmovdqa %xmm3, \i*16(%rax)
-.endm
-
-.macro sha256_xop_extend_doubleround i
- vmovdqa (\i-15)*16(%rax), %xmm0
- vmovdqa (\i-14)*16(%rax), %xmm4
- vprotd $25, %xmm0, %xmm1
- vprotd $25, %xmm4, %xmm5
- vprotd $14, %xmm0, %xmm2
- vprotd $14, %xmm4, %xmm6
- vpxor %xmm1, %xmm2, %xmm2
- vpxor %xmm5, %xmm6, %xmm6
- vpsrld $3, %xmm0, %xmm0
- vpsrld $3, %xmm4, %xmm4
- vpxor %xmm2, %xmm0, %xmm0
- vpxor %xmm6, %xmm4, %xmm4
-
- vpaddd (\i-16)*16(%rax), %xmm0, %xmm0
- vpaddd (\i-15)*16(%rax), %xmm4, %xmm4
-
- vprotd $15, %xmm3, %xmm1
- vprotd $15, %xmm7, %xmm5
- vprotd $13, %xmm3, %xmm2
- vprotd $13, %xmm7, %xmm6
- vpxor %xmm1, %xmm2, %xmm2
- vpxor %xmm5, %xmm6, %xmm6
-
- vpaddd (\i-7)*16(%rax), %xmm0, %xmm0
- vpaddd (\i-6)*16(%rax), %xmm4, %xmm4
-
- vpsrld $10, %xmm3, %xmm3
- vpsrld $10, %xmm7, %xmm7
- vpxor %xmm2, %xmm3, %xmm3
- vpxor %xmm6, %xmm7, %xmm7
-
- vpaddd %xmm0, %xmm3, %xmm3
- vpaddd %xmm4, %xmm7, %xmm7
- vmovdqa %xmm3, \i*16(%rax)
- vmovdqa %xmm7, (\i+1)*16(%rax)
-.endm
-
-.macro sha256_xop_main_round i, r0, r1, r2, r3, r4, r5, r6, r7
- vpaddd 16*(\i)(%rax), \r0, %xmm6
- vpaddd 16*(\i)(%rcx), %xmm6, %xmm6
-
- vpandn \r1, \r3, %xmm1
- vpand \r3, \r2, %xmm2
- vpxor %xmm2, %xmm1, %xmm1
- vpaddd %xmm1, %xmm6, %xmm6
-
- vprotd $26, \r3, %xmm1
- vprotd $21, \r3, %xmm2
- vpxor %xmm1, %xmm2, %xmm2
- vprotd $7, \r3, \r0
- vpxor %xmm2, \r0, \r0
- vpaddd \r0, %xmm6, %xmm6
- vpaddd %xmm6, \r4, \r0
-
- vpand \r6, \r5, %xmm2
- vpand \r7, \r5, \r4
- vpand \r7, \r6, %xmm1
- vpxor \r4, %xmm1, %xmm1
- vpxor %xmm2, %xmm1, %xmm1
- vpaddd %xmm1, %xmm6, %xmm6
-
- vprotd $30, \r7, %xmm1
- vprotd $19, \r7, %xmm2
- vpxor %xmm1, %xmm2, %xmm2
- vprotd $10, \r7, \r4
- vpxor %xmm2, \r4, \r4
- vpaddd %xmm6, \r4, \r4
-.endm
-
-.macro sha256_xop_main_quadround i
- sha256_xop_main_round \i+0, %xmm10, %xmm9, %xmm8, %xmm0, %xmm3, %xmm4, %xmm5, %xmm7
- sha256_xop_main_round \i+1, %xmm9, %xmm8, %xmm0, %xmm10, %xmm4, %xmm5, %xmm7, %xmm3
- sha256_xop_main_round \i+2, %xmm8, %xmm0, %xmm10, %xmm9, %xmm5, %xmm7, %xmm3, %xmm4
- sha256_xop_main_round \i+3, %xmm0, %xmm10, %xmm9, %xmm8, %xmm7, %xmm3, %xmm4, %xmm5
-.endm
-
- .text
- .p2align 6
-sha256_transform_4way_core_sse2:
- leaq 256(%rsp), %rcx
- leaq 48*16(%rcx), %rax
- movdqa -2*16(%rcx), %xmm3
- movdqa -1*16(%rcx), %xmm7
-sha256_transform_4way_sse2_extend_loop:
- movdqa -15*16(%rcx), %xmm0
- movdqa -14*16(%rcx), %xmm4
- movdqa %xmm0, %xmm2
- movdqa %xmm4, %xmm6
- psrld $3, %xmm0
- psrld $3, %xmm4
- movdqa %xmm0, %xmm1
- movdqa %xmm4, %xmm5
- pslld $14, %xmm2
- pslld $14, %xmm6
- psrld $4, %xmm1
- psrld $4, %xmm5
- pxor %xmm1, %xmm0
- pxor %xmm5, %xmm4
- psrld $11, %xmm1
- psrld $11, %xmm5
- pxor %xmm2, %xmm0
- pxor %xmm6, %xmm4
- pslld $11, %xmm2
- pslld $11, %xmm6
- pxor %xmm1, %xmm0
- pxor %xmm5, %xmm4
- pxor %xmm2, %xmm0
- pxor %xmm6, %xmm4
-
- paddd -16*16(%rcx), %xmm0
- paddd -15*16(%rcx), %xmm4
-
- movdqa %xmm3, %xmm2
- movdqa %xmm7, %xmm6
- psrld $10, %xmm3
- psrld $10, %xmm7
- movdqa %xmm3, %xmm1
- movdqa %xmm7, %xmm5
- pslld $13, %xmm2
- pslld $13, %xmm6
- psrld $7, %xmm1
- psrld $7, %xmm5
-
- paddd -7*16(%rcx), %xmm0
- paddd -6*16(%rcx), %xmm4
-
- pxor %xmm1, %xmm3
- pxor %xmm5, %xmm7
- psrld $2, %xmm1
- psrld $2, %xmm5
- pxor %xmm2, %xmm3
- pxor %xmm6, %xmm7
- pslld $2, %xmm2
- pslld $2, %xmm6
- pxor %xmm1, %xmm3
- pxor %xmm5, %xmm7
- pxor %xmm2, %xmm3
- pxor %xmm6, %xmm7
-
- paddd %xmm0, %xmm3
- paddd %xmm4, %xmm7
- movdqa %xmm3, (%rcx)
- movdqa %xmm7, 16(%rcx)
- addq $2*16, %rcx
- cmpq %rcx, %rax
- jne sha256_transform_4way_sse2_extend_loop
-
- movdqu 0(%rdi), %xmm7
- movdqu 16(%rdi), %xmm5
- movdqu 32(%rdi), %xmm4
- movdqu 48(%rdi), %xmm3
- movdqu 64(%rdi), %xmm0
- movdqu 80(%rdi), %xmm8
- movdqu 96(%rdi), %xmm9
- movdqu 112(%rdi), %xmm10
-
- leaq sha256_4k(%rip), %rcx
- xorq %rax, %rax
-sha256_transform_4way_sse2_main_loop:
- movdqa (%rsp, %rax), %xmm6
- paddd (%rcx, %rax), %xmm6
- paddd %xmm10, %xmm6
-
- movdqa %xmm0, %xmm1
- movdqa %xmm9, %xmm2
- pandn %xmm2, %xmm1
-
- movdqa %xmm2, %xmm10
- movdqa %xmm8, %xmm2
- movdqa %xmm2, %xmm9
-
- pand %xmm0, %xmm2
- pxor %xmm2, %xmm1
- movdqa %xmm0, %xmm8
-
- paddd %xmm1, %xmm6
-
- movdqa %xmm0, %xmm1
- psrld $6, %xmm0
- movdqa %xmm0, %xmm2
- pslld $7, %xmm1
- psrld $5, %xmm2
- pxor %xmm1, %xmm0
- pxor %xmm2, %xmm0
- pslld $14, %xmm1
- psrld $14, %xmm2
- pxor %xmm1, %xmm0
- pxor %xmm2, %xmm0
- pslld $5, %xmm1
- pxor %xmm1, %xmm0
- paddd %xmm0, %xmm6
-
- movdqa %xmm3, %xmm0
- paddd %xmm6, %xmm0
-
- movdqa %xmm5, %xmm1
- movdqa %xmm4, %xmm3
- movdqa %xmm4, %xmm2
- pand %xmm5, %xmm2
- pand %xmm7, %xmm4
- pand %xmm7, %xmm1
- pxor %xmm4, %xmm1
- movdqa %xmm5, %xmm4
- movdqa %xmm7, %xmm5
- pxor %xmm2, %xmm1
- paddd %xmm1, %xmm6
-
- movdqa %xmm7, %xmm2
- psrld $2, %xmm7
- movdqa %xmm7, %xmm1
- pslld $10, %xmm2
- psrld $11, %xmm1
- pxor %xmm2, %xmm7
- pxor %xmm1, %xmm7
- pslld $9, %xmm2
- psrld $9, %xmm1
- pxor %xmm2, %xmm7
- pxor %xmm1, %xmm7
- pslld $11, %xmm2
- pxor %xmm2, %xmm7
- paddd %xmm6, %xmm7
-
- addq $16, %rax
- cmpq $16*64, %rax
- jne sha256_transform_4way_sse2_main_loop
- jmp sha256_transform_4way_finish
-
- .text
- .p2align 6
-sha256_transform_4way_core_avx:
- leaq 256(%rsp), %rax
- movdqa -2*16(%rax), %xmm3
- movdqa -1*16(%rax), %xmm7
- sha256_avx_extend_doubleround 0
- sha256_avx_extend_doubleround 2
- sha256_avx_extend_doubleround 4
- sha256_avx_extend_doubleround 6
- sha256_avx_extend_doubleround 8
- sha256_avx_extend_doubleround 10
- sha256_avx_extend_doubleround 12
- sha256_avx_extend_doubleround 14
- sha256_avx_extend_doubleround 16
- sha256_avx_extend_doubleround 18
- sha256_avx_extend_doubleround 20
- sha256_avx_extend_doubleround 22
- sha256_avx_extend_doubleround 24
- sha256_avx_extend_doubleround 26
- sha256_avx_extend_doubleround 28
- sha256_avx_extend_doubleround 30
- sha256_avx_extend_doubleround 32
- sha256_avx_extend_doubleround 34
- sha256_avx_extend_doubleround 36
- sha256_avx_extend_doubleround 38
- sha256_avx_extend_doubleround 40
- sha256_avx_extend_doubleround 42
- sha256_avx_extend_doubleround 44
- sha256_avx_extend_doubleround 46
- movdqu 0(%rdi), %xmm7
- movdqu 16(%rdi), %xmm5
- movdqu 32(%rdi), %xmm4
- movdqu 48(%rdi), %xmm3
- movdqu 64(%rdi), %xmm0
- movdqu 80(%rdi), %xmm8
- movdqu 96(%rdi), %xmm9
- movdqu 112(%rdi), %xmm10
- movq %rsp, %rax
- leaq sha256_4k(%rip), %rcx
- sha256_avx_main_quadround 0
- sha256_avx_main_quadround 4
- sha256_avx_main_quadround 8
- sha256_avx_main_quadround 12
- sha256_avx_main_quadround 16
- sha256_avx_main_quadround 20
- sha256_avx_main_quadround 24
- sha256_avx_main_quadround 28
- sha256_avx_main_quadround 32
- sha256_avx_main_quadround 36
- sha256_avx_main_quadround 40
- sha256_avx_main_quadround 44
- sha256_avx_main_quadround 48
- sha256_avx_main_quadround 52
- sha256_avx_main_quadround 56
- sha256_avx_main_quadround 60
- jmp sha256_transform_4way_finish
-
- .text
- .p2align 6
-sha256_transform_4way_core_xop:
- leaq 256(%rsp), %rax
- movdqa -2*16(%rax), %xmm3
- movdqa -1*16(%rax), %xmm7
- sha256_xop_extend_doubleround 0
- sha256_xop_extend_doubleround 2
- sha256_xop_extend_doubleround 4
- sha256_xop_extend_doubleround 6
- sha256_xop_extend_doubleround 8
- sha256_xop_extend_doubleround 10
- sha256_xop_extend_doubleround 12
- sha256_xop_extend_doubleround 14
- sha256_xop_extend_doubleround 16
- sha256_xop_extend_doubleround 18
- sha256_xop_extend_doubleround 20
- sha256_xop_extend_doubleround 22
- sha256_xop_extend_doubleround 24
- sha256_xop_extend_doubleround 26
- sha256_xop_extend_doubleround 28
- sha256_xop_extend_doubleround 30
- sha256_xop_extend_doubleround 32
- sha256_xop_extend_doubleround 34
- sha256_xop_extend_doubleround 36
- sha256_xop_extend_doubleround 38
- sha256_xop_extend_doubleround 40
- sha256_xop_extend_doubleround 42
- sha256_xop_extend_doubleround 44
- sha256_xop_extend_doubleround 46
- movdqu 0(%rdi), %xmm7
- movdqu 16(%rdi), %xmm5
- movdqu 32(%rdi), %xmm4
- movdqu 48(%rdi), %xmm3
- movdqu 64(%rdi), %xmm0
- movdqu 80(%rdi), %xmm8
- movdqu 96(%rdi), %xmm9
- movdqu 112(%rdi), %xmm10
- movq %rsp, %rax
- leaq sha256_4k(%rip), %rcx
- sha256_xop_main_quadround 0
- sha256_xop_main_quadround 4
- sha256_xop_main_quadround 8
- sha256_xop_main_quadround 12
- sha256_xop_main_quadround 16
- sha256_xop_main_quadround 20
- sha256_xop_main_quadround 24
- sha256_xop_main_quadround 28
- sha256_xop_main_quadround 32
- sha256_xop_main_quadround 36
- sha256_xop_main_quadround 40
- sha256_xop_main_quadround 44
- sha256_xop_main_quadround 48
- sha256_xop_main_quadround 52
- sha256_xop_main_quadround 56
- sha256_xop_main_quadround 60
- jmp sha256_transform_4way_finish
-
- .data
- .p2align 3
-sha256_transform_4way_core_addr:
- .quad 0x0
-
-.macro p2bswap_rsi_rsp i
- movdqu \i*16(%rsi), %xmm0
- movdqu (\i+1)*16(%rsi), %xmm2
- pshuflw $0xb1, %xmm0, %xmm0
- pshuflw $0xb1, %xmm2, %xmm2
- pshufhw $0xb1, %xmm0, %xmm0
- pshufhw $0xb1, %xmm2, %xmm2
- movdqa %xmm0, %xmm1
- movdqa %xmm2, %xmm3
- psrlw $8, %xmm1
- psrlw $8, %xmm3
- psllw $8, %xmm0
- psllw $8, %xmm2
- pxor %xmm1, %xmm0
- pxor %xmm3, %xmm2
- movdqa %xmm0, \i*16(%rsp)
- movdqa %xmm2, (\i+1)*16(%rsp)
-.endm
-
- .text
- .p2align 6
- .globl sha256_transform_4way
- .globl _sha256_transform_4way
-sha256_transform_4way:
-_sha256_transform_4way:
-#if defined(_WIN64) || defined(__CYGWIN__)
- pushq %rdi
- subq $96, %rsp
- movdqa %xmm6, 0(%rsp)
- movdqa %xmm7, 16(%rsp)
- movdqa %xmm8, 32(%rsp)
- movdqa %xmm9, 48(%rsp)
- movdqa %xmm10, 64(%rsp)
- movdqa %xmm11, 80(%rsp)
- pushq %rsi
- movq %rcx, %rdi
- movq %rdx, %rsi
- movq %r8, %rdx
-#endif
- movq %rsp, %r8
- subq $1032, %rsp
- andq $-128, %rsp
-
- testq %rdx, %rdx
- jnz sha256_transform_4way_swap
-
- movdqu 0*16(%rsi), %xmm0
- movdqu 1*16(%rsi), %xmm1
- movdqu 2*16(%rsi), %xmm2
- movdqu 3*16(%rsi), %xmm3
- movdqu 4*16(%rsi), %xmm4
- movdqu 5*16(%rsi), %xmm5
- movdqu 6*16(%rsi), %xmm6
- movdqu 7*16(%rsi), %xmm7
- movdqa %xmm0, 0*16(%rsp)
- movdqa %xmm1, 1*16(%rsp)
- movdqa %xmm2, 2*16(%rsp)
- movdqa %xmm3, 3*16(%rsp)
- movdqa %xmm4, 4*16(%rsp)
- movdqa %xmm5, 5*16(%rsp)
- movdqa %xmm6, 6*16(%rsp)
- movdqa %xmm7, 7*16(%rsp)
- movdqu 8*16(%rsi), %xmm0
- movdqu 9*16(%rsi), %xmm1
- movdqu 10*16(%rsi), %xmm2
- movdqu 11*16(%rsi), %xmm3
- movdqu 12*16(%rsi), %xmm4
- movdqu 13*16(%rsi), %xmm5
- movdqu 14*16(%rsi), %xmm6
- movdqu 15*16(%rsi), %xmm7
- movdqa %xmm0, 8*16(%rsp)
- movdqa %xmm1, 9*16(%rsp)
- movdqa %xmm2, 10*16(%rsp)
- movdqa %xmm3, 11*16(%rsp)
- movdqa %xmm4, 12*16(%rsp)
- movdqa %xmm5, 13*16(%rsp)
- movdqa %xmm6, 14*16(%rsp)
- movdqa %xmm7, 15*16(%rsp)
- jmp *sha256_transform_4way_core_addr(%rip)
-
- .p2align 6
-sha256_transform_4way_swap:
- p2bswap_rsi_rsp 0
- p2bswap_rsi_rsp 2
- p2bswap_rsi_rsp 4
- p2bswap_rsi_rsp 6
- p2bswap_rsi_rsp 8
- p2bswap_rsi_rsp 10
- p2bswap_rsi_rsp 12
- p2bswap_rsi_rsp 14
- jmp *sha256_transform_4way_core_addr(%rip)
-
- .p2align 6
-sha256_transform_4way_finish:
- movdqu 0(%rdi), %xmm2
- movdqu 16(%rdi), %xmm6
- movdqu 32(%rdi), %xmm11
- movdqu 48(%rdi), %xmm1
- paddd %xmm2, %xmm7
- paddd %xmm6, %xmm5
- paddd %xmm11, %xmm4
- paddd %xmm1, %xmm3
- movdqu 64(%rdi), %xmm2
- movdqu 80(%rdi), %xmm6
- movdqu 96(%rdi), %xmm11
- movdqu 112(%rdi), %xmm1
- paddd %xmm2, %xmm0
- paddd %xmm6, %xmm8
- paddd %xmm11, %xmm9
- paddd %xmm1, %xmm10
-
- movdqu %xmm7, 0(%rdi)
- movdqu %xmm5, 16(%rdi)
- movdqu %xmm4, 32(%rdi)
- movdqu %xmm3, 48(%rdi)
- movdqu %xmm0, 64(%rdi)
- movdqu %xmm8, 80(%rdi)
- movdqu %xmm9, 96(%rdi)
- movdqu %xmm10, 112(%rdi)
-
- movq %r8, %rsp
-#if defined(_WIN64) || defined(__CYGWIN__)
- popq %rsi
- movdqa 0(%rsp), %xmm6
- movdqa 16(%rsp), %xmm7
- movdqa 32(%rsp), %xmm8
- movdqa 48(%rsp), %xmm9
- movdqa 64(%rsp), %xmm10
- movdqa 80(%rsp), %xmm11
- addq $96, %rsp
- popq %rdi
-#endif
- ret
-
- .text
- .p2align 6
-sha256_transform_8way_core_avx2:
- leaq 8*64(%rsp), %rax
- vmovdqa -2*32(%rax), %ymm3
- vmovdqa -1*32(%rax), %ymm7
- sha256_avx2_extend_doubleround 0
- sha256_avx2_extend_doubleround 2
- sha256_avx2_extend_doubleround 4
- sha256_avx2_extend_doubleround 6
- sha256_avx2_extend_doubleround 8
- sha256_avx2_extend_doubleround 10
- sha256_avx2_extend_doubleround 12
- sha256_avx2_extend_doubleround 14
- sha256_avx2_extend_doubleround 16
- sha256_avx2_extend_doubleround 18
- sha256_avx2_extend_doubleround 20
- sha256_avx2_extend_doubleround 22
- sha256_avx2_extend_doubleround 24
- sha256_avx2_extend_doubleround 26
- sha256_avx2_extend_doubleround 28
- sha256_avx2_extend_doubleround 30
- sha256_avx2_extend_doubleround 32
- sha256_avx2_extend_doubleround 34
- sha256_avx2_extend_doubleround 36
- sha256_avx2_extend_doubleround 38
- sha256_avx2_extend_doubleround 40
- sha256_avx2_extend_doubleround 42
- sha256_avx2_extend_doubleround 44
- sha256_avx2_extend_doubleround 46
- vmovdqu 0*32(%rdi), %ymm7
- vmovdqu 1*32(%rdi), %ymm5
- vmovdqu 2*32(%rdi), %ymm4
- vmovdqu 3*32(%rdi), %ymm3
- vmovdqu 4*32(%rdi), %ymm0
- vmovdqu 5*32(%rdi), %ymm8
- vmovdqu 6*32(%rdi), %ymm9
- vmovdqu 7*32(%rdi), %ymm10
- movq %rsp, %rax
- leaq sha256_8k(%rip), %rcx
- sha256_avx2_main_quadround 0
- sha256_avx2_main_quadround 4
- sha256_avx2_main_quadround 8
- sha256_avx2_main_quadround 12
- sha256_avx2_main_quadround 16
- sha256_avx2_main_quadround 20
- sha256_avx2_main_quadround 24
- sha256_avx2_main_quadround 28
- sha256_avx2_main_quadround 32
- sha256_avx2_main_quadround 36
- sha256_avx2_main_quadround 40
- sha256_avx2_main_quadround 44
- sha256_avx2_main_quadround 48
- sha256_avx2_main_quadround 52
- sha256_avx2_main_quadround 56
- sha256_avx2_main_quadround 60
- jmp sha256_transform_8way_finish
-
-.macro p2bswap_avx2_rsi_rsp i
- vmovdqu \i*32(%rsi), %ymm0
- vmovdqu (\i+1)*32(%rsi), %ymm2
- vpshuflw $0xb1, %ymm0, %ymm0
- vpshuflw $0xb1, %ymm2, %ymm2
- vpshufhw $0xb1, %ymm0, %ymm0
- vpshufhw $0xb1, %ymm2, %ymm2
- vpsrlw $8, %ymm0, %ymm1
- vpsrlw $8, %ymm2, %ymm3
- vpsllw $8, %ymm0, %ymm0
- vpsllw $8, %ymm2, %ymm2
- vpxor %ymm1, %ymm0, %ymm0
- vpxor %ymm3, %ymm2, %ymm2
- vmovdqa %ymm0, \i*32(%rsp)
- vmovdqa %ymm2, (\i+1)*32(%rsp)
-.endm
-
- .text
- .p2align 6
- .globl sha256_transform_8way
- .globl _sha256_transform_8way
-sha256_transform_8way:
-_sha256_transform_8way:
-#if defined(_WIN64) || defined(__CYGWIN__)
- pushq %rdi
- subq $96, %rsp
- vmovdqa %xmm6, 0(%rsp)
- vmovdqa %xmm7, 16(%rsp)
- vmovdqa %xmm8, 32(%rsp)
- vmovdqa %xmm9, 48(%rsp)
- vmovdqa %xmm10, 64(%rsp)
- vmovdqa %xmm11, 80(%rsp)
- pushq %rsi
- movq %rcx, %rdi
- movq %rdx, %rsi
- movq %r8, %rdx
-#endif
- movq %rsp, %r8
- subq $64*32, %rsp
- andq $-128, %rsp
-
- testq %rdx, %rdx
- jnz sha256_transform_8way_swap
-
- vmovdqu 0*32(%rsi), %ymm0
- vmovdqu 1*32(%rsi), %ymm1
- vmovdqu 2*32(%rsi), %ymm2
- vmovdqu 3*32(%rsi), %ymm3
- vmovdqu 4*32(%rsi), %ymm4
- vmovdqu 5*32(%rsi), %ymm5
- vmovdqu 6*32(%rsi), %ymm6
- vmovdqu 7*32(%rsi), %ymm7
- vmovdqa %ymm0, 0*32(%rsp)
- vmovdqa %ymm1, 1*32(%rsp)
- vmovdqa %ymm2, 2*32(%rsp)
- vmovdqa %ymm3, 3*32(%rsp)
- vmovdqa %ymm4, 4*32(%rsp)
- vmovdqa %ymm5, 5*32(%rsp)
- vmovdqa %ymm6, 6*32(%rsp)
- vmovdqa %ymm7, 7*32(%rsp)
- vmovdqu 8*32(%rsi), %ymm0
- vmovdqu 9*32(%rsi), %ymm1
- vmovdqu 10*32(%rsi), %ymm2
- vmovdqu 11*32(%rsi), %ymm3
- vmovdqu 12*32(%rsi), %ymm4
- vmovdqu 13*32(%rsi), %ymm5
- vmovdqu 14*32(%rsi), %ymm6
- vmovdqu 15*32(%rsi), %ymm7
- vmovdqa %ymm0, 8*32(%rsp)
- vmovdqa %ymm1, 9*32(%rsp)
- vmovdqa %ymm2, 10*32(%rsp)
- vmovdqa %ymm3, 11*32(%rsp)
- vmovdqa %ymm4, 12*32(%rsp)
- vmovdqa %ymm5, 13*32(%rsp)
- vmovdqa %ymm6, 14*32(%rsp)
- vmovdqa %ymm7, 15*32(%rsp)
- jmp sha256_transform_8way_core_avx2
-
- .p2align 6
-sha256_transform_8way_swap:
- p2bswap_avx2_rsi_rsp 0
- p2bswap_avx2_rsi_rsp 2
- p2bswap_avx2_rsi_rsp 4
- p2bswap_avx2_rsi_rsp 6
- p2bswap_avx2_rsi_rsp 8
- p2bswap_avx2_rsi_rsp 10
- p2bswap_avx2_rsi_rsp 12
- p2bswap_avx2_rsi_rsp 14
- jmp sha256_transform_8way_core_avx2
-
- .p2align 6
-sha256_transform_8way_finish:
- vmovdqu 0*32(%rdi), %ymm2
- vmovdqu 1*32(%rdi), %ymm6
- vmovdqu 2*32(%rdi), %ymm11
- vmovdqu 3*32(%rdi), %ymm1
- vpaddd %ymm2, %ymm7, %ymm7
- vpaddd %ymm6, %ymm5, %ymm5
- vpaddd %ymm11, %ymm4, %ymm4
- vpaddd %ymm1, %ymm3, %ymm3
- vmovdqu 4*32(%rdi), %ymm2
- vmovdqu 5*32(%rdi), %ymm6
- vmovdqu 6*32(%rdi), %ymm11
- vmovdqu 7*32(%rdi), %ymm1
- vpaddd %ymm2, %ymm0, %ymm0
- vpaddd %ymm6, %ymm8, %ymm8
- vpaddd %ymm11, %ymm9, %ymm9
- vpaddd %ymm1, %ymm10, %ymm10
-
- vmovdqu %ymm7, 0*32(%rdi)
- vmovdqu %ymm5, 1*32(%rdi)
- vmovdqu %ymm4, 2*32(%rdi)
- vmovdqu %ymm3, 3*32(%rdi)
- vmovdqu %ymm0, 4*32(%rdi)
- vmovdqu %ymm8, 5*32(%rdi)
- vmovdqu %ymm9, 6*32(%rdi)
- vmovdqu %ymm10, 7*32(%rdi)
-
- movq %r8, %rsp
-#if defined(_WIN64) || defined(__CYGWIN__)
- popq %rsi
- vmovdqa 0(%rsp), %xmm6
- vmovdqa 16(%rsp), %xmm7
- vmovdqa 32(%rsp), %xmm8
- vmovdqa 48(%rsp), %xmm9
- vmovdqa 64(%rsp), %xmm10
- vmovdqa 80(%rsp), %xmm11
- addq $96, %rsp
- popq %rdi
-#endif
- ret
-
-
-.macro sha256_sse2_main_round_red i, r7
- movdqa 16*\i(%rax), %xmm6
- paddd 16*\i(%rcx), %xmm6
- paddd 32(%rsp), %xmm6
- movdqa %xmm0, %xmm1
- movdqa 16(%rsp), %xmm2
- paddd \r7, %xmm6
- pandn %xmm2, %xmm1
- movdqa %xmm2, 32(%rsp)
- movdqa 0(%rsp), %xmm2
- movdqa %xmm2, 16(%rsp)
- pand %xmm0, %xmm2
- pxor %xmm2, %xmm1
- movdqa %xmm0, 0(%rsp)
- paddd %xmm1, %xmm6
- movdqa %xmm0, %xmm1
- psrld $6, %xmm0
- movdqa %xmm0, %xmm2
- pslld $7, %xmm1
- psrld $5, %xmm2
- pxor %xmm1, %xmm0
- pxor %xmm2, %xmm0
- pslld $14, %xmm1
- psrld $14, %xmm2
- pxor %xmm1, %xmm0
- pxor %xmm2, %xmm0
- pslld $5, %xmm1
- pxor %xmm1, %xmm0
- paddd %xmm6, %xmm0
-.endm
-
-.macro sha256_avx_main_round_red i, r0, r1, r2, r3, r4
- vpaddd 16*\i(%rax), \r0, %xmm6
- vpaddd 16*\i(%rcx), %xmm6, %xmm6
- vpandn \r1, \r3, %xmm1
- vpand \r3, \r2, %xmm2
- vpxor %xmm2, %xmm1, %xmm1
- vpaddd %xmm1, %xmm6, %xmm6
- vpslld $7, \r3, %xmm1
- vpsrld $6, \r3, \r0
- vpsrld $5, \r0, %xmm2
- vpxor %xmm1, \r0, \r0
- vpxor %xmm2, \r0, \r0
- vpslld $14, %xmm1, %xmm1
- vpsrld $14, %xmm2, %xmm2
- vpxor %xmm1, \r0, \r0
- vpxor %xmm2, \r0, \r0
- vpslld $5, %xmm1, %xmm1
- vpxor %xmm1, \r0, \r0
- vpaddd \r0, %xmm6, %xmm6
- vpaddd %xmm6, \r4, \r0
-.endm
-
-.macro sha256_xop_main_round_red i, r0, r1, r2, r3, r4
- vpaddd 16*\i(%rax), \r0, %xmm6
- vpaddd 16*\i(%rcx), %xmm6, %xmm6
- vpandn \r1, \r3, %xmm1
- vpand \r3, \r2, %xmm2
- vpxor %xmm2, %xmm1, %xmm1
- vpaddd %xmm1, %xmm6, %xmm6
- vprotd $26, \r3, %xmm1
- vprotd $21, \r3, %xmm2
- vpxor %xmm1, %xmm2, %xmm2
- vprotd $7, \r3, \r0
- vpxor %xmm2, \r0, \r0
- vpaddd \r0, %xmm6, %xmm6
- vpaddd %xmm6, \r4, \r0
-.endm
-
- .text
- .p2align 6
- .globl sha256_use_4way
- .globl _sha256_use_4way
-sha256_use_4way:
-_sha256_use_4way:
- pushq %rbx
- pushq %rcx
- pushq %rdx
-
- /* Check for VIA PadLock Hash Engine */
- movl $0xc0000000, %eax
- cpuid
- cmpl $0xc0000001, %eax
- jb sha256_use_4way_no_phe
- movl $0xc0000001, %eax
- cpuid
- andl $0x00000c00, %edx
- cmpl $0x00000c00, %edx
- jne sha256_use_4way_no_phe
- leaq sha256_transform_phe(%rip), %rdx
- movq %rdx, sha256_transform_addr(%rip)
- xorl %eax, %eax
- jmp sha256_use_4way_exit
-sha256_use_4way_no_phe:
- /* Check for AVX and OSXSAVE support */
- movl $1, %eax
- cpuid
- andl $0x18000000, %ecx
- cmpl $0x18000000, %ecx
- jne sha256_use_4way_base
- /* Check for XMM and YMM state support */
- xorl %ecx, %ecx
- xgetbv
- andl $0x00000006, %eax
- cmpl $0x00000006, %eax
- jne sha256_use_4way_base
- /* Check for XOP support */
- movl $0x80000001, %eax
- cpuid
- andl $0x00000800, %ecx
- jz sha256_use_4way_avx
-
-sha256_use_4way_xop:
- leaq sha256_transform_4way_core_xop(%rip), %rdx
- jmp sha256_use_4way_done
-
-sha256_use_4way_avx:
- leaq sha256_transform_4way_core_avx(%rip), %rdx
- jmp sha256_use_4way_done
-
-sha256_use_4way_base:
- leaq sha256_transform_4way_core_sse2(%rip), %rdx
-
-sha256_use_4way_done:
- movq %rdx, sha256_transform_4way_core_addr(%rip)
- movl $1, %eax
-sha256_use_4way_exit:
- popq %rdx
- popq %rcx
- popq %rbx
- ret
-
- .text
- .p2align 6
- .globl sha256_use_ssse3
- .globl _sha256_use_ssse3
-sha256_use_ssse3:
-_sha256_use_ssse3:
- pushq %rbx
- pushq %rcx
- pushq %rdx
- cpuid
- andl $0x00000200, %ecx
- jz sha256_use_ssse3_done
- xorl %eax, %eax
- popq %rdx
- popq %rcx
- popq %rbx
- ret
-
-sha256_use_ssse3_done:
- movl $1, %eax
- popq %rdx
- popq %rcx
- popq %rbx
- ret
-
-.macro sha256_avx2_main_round_red i, r0, r1, r2, r3, r4
- vpaddd 32*\i(%rax), \r0, %ymm6
- vpaddd 32*\i(%rcx), %ymm6, %ymm6
- vpandn \r1, \r3, %ymm1
- vpand \r3, \r2, %ymm2
- vpxor %ymm2, %ymm1, %ymm1
- vpaddd %ymm1, %ymm6, %ymm6
- vpslld $7, \r3, %ymm1
- vpsrld $6, \r3, \r0
- vpsrld $5, \r0, %ymm2
- vpxor %ymm1, \r0, \r0
- vpxor %ymm2, \r0, \r0
- vpslld $14, %ymm1, %ymm1
- vpsrld $14, %ymm2, %ymm2
- vpxor %ymm1, \r0, \r0
- vpxor %ymm2, \r0, \r0
- vpslld $5, %ymm1, %ymm1
- vpxor %ymm1, \r0, \r0
- vpaddd \r0, %ymm6, %ymm6
- vpaddd %ymm6, \r4, \r0
-.endm
-
- .text
- .p2align 6
- .globl sha256_use_8way
- .globl _sha256_use_8way
-sha256_use_8way:
-_sha256_use_8way:
-
- pushq %rbx
- /* Check for AVX and OSXSAVE support */
- movl $1, %eax
- cpuid
- andl $0x18000000, %ecx
- cmpl $0x18000000, %ecx
- jne sha256_use_8way_no
- /* Check for AVX2 support */
- movl $7, %eax
- xorl %ecx, %ecx
- cpuid
- andl $0x00000020, %ebx
- cmpl $0x00000020, %ebx
- jne sha256_use_8way_no
- /* Check for XMM and YMM state support */
- xorl %ecx, %ecx
- xgetbv
- andl $0x00000006, %eax
- cmpl $0x00000006, %eax
- jne sha256_use_8way_no
-
-sha256_use_8way_yes:
- movl $1, %eax
- jmp sha256_use_8way_done
-
-sha256_use_8way_no:
- xorl %eax, %eax
-
-sha256_use_8way_done:
- popq %rbx
- ret
-
-#endif
using namespace std;
-#ifdef USE_ASM
-
-#ifdef _MSC_VER
-#include <stdlib.h>
-#define __builtin_bswap32 _byteswap_ulong
-#endif
-
-#if defined(__i386__) || defined(__x86_64__)
-#include <immintrin.h>
-#endif
-
-#ifndef __i386__
-// kernel padding
-static const uint32_t block1_suffix[9] = { 0x80000000, 0, 0, 0, 0, 0, 0, 0, 0x000000e0 };
-// hash padding
-static const uint32_t block2_suffix[8] = { 0x80000000, 0, 0, 0, 0, 0, 0, 0x00000100 };
-
-// Sha256 initial state
-static const uint32_t sha256_initial[8] = { 0x6a09e667, 0xbb67ae85, 0x3c6ef372, 0xa54ff53a, 0x510e527f, 0x9b05688c, 0x1f83d9ab, 0x5be0cd19 };
-
-extern "C" void sha256_transform(uint32_t *state, const uint32_t *block, int swap);
-#endif
-
-// 4-way kernel padding
-static const uint32_t block1_suffix_4way[4 * 9] = {
- 0x80000000, 0x80000000, 0x80000000, 0x80000000,
- 0, 0, 0, 0,
- 0, 0, 0, 0,
- 0, 0, 0, 0,
- 0, 0, 0, 0,
- 0, 0, 0, 0,
- 0, 0, 0, 0,
- 0, 0, 0, 0,
- 0x000000e0, 0x000000e0, 0x000000e0, 0x000000e0
-};
-
-// 4-way hash padding
-static const uint32_t block2_suffix_4way[4 * 8] = {
- 0x80000000, 0x80000000, 0x80000000, 0x80000000,
- 0, 0, 0, 0,
- 0, 0, 0, 0,
- 0, 0, 0, 0,
- 0, 0, 0, 0,
- 0, 0, 0, 0,
- 0, 0, 0, 0,
- 0x00000100, 0x00000100, 0x00000100, 0x00000100
-};
-
-extern "C" int sha256_use_4way();
-extern "C" void sha256_init_4way(uint32_t *state);
-extern "C" void sha256_transform_4way(uint32_t *state, const uint32_t *block, int swap);
-bool fUse4Way = sha256_use_4way() != 0;
-
-#ifdef __x86_64__
-// 8-way kernel padding
-static const uint32_t block1_suffix_8way[8 * 9] = {
- 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000,
- 0, 0, 0, 0, 0, 0, 0, 0,
- 0, 0, 0, 0, 0, 0, 0, 0,
- 0, 0, 0, 0, 0, 0, 0, 0,
- 0, 0, 0, 0, 0, 0, 0, 0,
- 0, 0, 0, 0, 0, 0, 0, 0,
- 0, 0, 0, 0, 0, 0, 0, 0,
- 0, 0, 0, 0, 0, 0, 0, 0,
- 0x000000e0, 0x000000e0, 0x000000e0, 0x000000e0, 0x000000e0, 0x000000e0, 0x000000e0, 0x000000e0
-};
-
-// 8-way hash padding
-static const uint32_t block2_suffix_8way[8 * 8] = {
- 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000, 0x80000000,
- 0, 0, 0, 0, 0, 0, 0, 0,
- 0, 0, 0, 0, 0, 0, 0, 0,
- 0, 0, 0, 0, 0, 0, 0, 0,
- 0, 0, 0, 0, 0, 0, 0, 0,
- 0, 0, 0, 0, 0, 0, 0, 0,
- 0, 0, 0, 0, 0, 0, 0, 0,
- 0x000000e0, 0x000000e0, 0x000000e0, 0x000000e0, 0x000000e0, 0x000000e0, 0x000000e0, 0x000000e0
-};
-
-extern "C" int sha256_use_8way();
-extern "C" void sha256_init_8way(uint32_t *state);
-extern "C" void sha256_transform_8way(uint32_t *state, const uint32_t *block, int swap);
-bool fUse8Way = sha256_use_8way() != 0;
-
-inline void copyrow8_swap32(uint32_t *to, uint32_t *from)
-{
- // There are no AVX2 CPUs without SSSE3 support, so we don't need any conditions here.
- __m128i mask = _mm_set_epi8(12, 13, 14, 15, 8, 9, 10, 11, 4, 5, 6, 7, 0, 1, 2, 3);
- _mm_storeu_si128((__m128i *)&to[0], _mm_shuffle_epi8(_mm_loadu_si128((__m128i *)&from[0]), mask));
- _mm_storeu_si128((__m128i *)&to[4], _mm_shuffle_epi8(_mm_loadu_si128((__m128i *)&from[4]), mask));
-}
-#endif
-
-#if defined(__i386__) || defined(__x86_64__)
-extern "C" int sha256_use_ssse3();
-bool fUseSSSE3 = sha256_use_ssse3() != 0;
-
-inline void copyrow4_swap32(uint32_t *to, uint32_t *from)
-{
- if (!fUseSSSE3)
- {
- for (int i = 0; i < 4; i++)
- to[i] = __builtin_bswap32(from[i]);
- }
- else
- {
- __m128i mask = _mm_set_epi8(12, 13, 14, 15, 8, 9, 10, 11, 4, 5, 6, 7, 0, 1, 2, 3);
- _mm_storeu_si128((__m128i *)&to[0], _mm_shuffle_epi8(_mm_loadu_si128((__m128i *)&from[0]), mask));
- }
-}
-#else
-inline void copyrow4_swap32(uint32_t *to, uint32_t *from)
-{
- for (int i = 0; i < 4; i++)
- to[i] = __builtin_bswap32(from[i]);
-}
-#endif
-#endif
-
KernelWorker::KernelWorker(unsigned char *kernel, uint32_t nBits, uint32_t nInputTxTime, int64_t nValueIn, uint32_t nIntervalBegin, uint32_t nIntervalEnd)
: kernel(kernel), nBits(nBits), nInputTxTime(nInputTxTime), bnValueIn(nValueIn), nIntervalBegin(nIntervalBegin), nIntervalEnd(nIntervalEnd)
{
solutions = vector<std::pair<uint256,uint32_t> >();
}
-#ifdef USE_ASM
-#ifdef __x86_64__
-void KernelWorker::Do_8way()
-{
- SetThreadPriority(THREAD_PRIORITY_LOWEST);
-
- // Compute maximum possible target to filter out majority of obviously insufficient hashes
- CBigNum bnTargetPerCoinDay;
- bnTargetPerCoinDay.SetCompact(nBits);
- uint256 nMaxTarget = (bnTargetPerCoinDay * bnValueIn * nStakeMaxAge / COIN / nOneDay).getuint256();
-
-#ifdef _MSC_VER
- __declspec(align(16)) uint32_t blocks1[8 * 16];
- __declspec(align(16)) uint32_t blocks2[8 * 16];
- __declspec(align(16)) uint32_t candidates[8 * 8];
-#else
- uint32_t blocks1[8 * 16] __attribute__((aligned(16)));
- uint32_t blocks2[8 * 16] __attribute__((aligned(16)));
- uint32_t candidates[8 * 8] __attribute__((aligned(16)));
-#endif
-
- vector<uint32_t> vRow = vector<uint32_t>(8);
- uint32_t *pnKernel = (uint32_t *) kernel;
-
- for(int i = 0; i < 7; i++)
- {
- fill(vRow.begin(), vRow.end(), pnKernel[i]);
- copyrow8_swap32(&blocks1[i*8], &vRow[0]);
- }
-
- memcpy(&blocks1[56], &block1_suffix_8way[0], 36*8); // sha256 padding
- memcpy(&blocks2[64], &block2_suffix_8way[0], 32*8);
-
- uint32_t nHashes[8];
- uint32_t nTimeStamps[8];
-
- // Search forward in time from the given timestamp
- // Stopping search in case of shutting down
- for (uint32_t nTimeTx=nIntervalBegin, nMaxTarget32 = nMaxTarget.Get32(7); nTimeTx<nIntervalEnd && !fShutdown; nTimeTx +=8)
- {
- sha256_init_8way(blocks2);
- sha256_init_8way(candidates);
-
- nTimeStamps[0] = nTimeTx;
- nTimeStamps[1] = nTimeTx+1;
- nTimeStamps[2] = nTimeTx+2;
- nTimeStamps[3] = nTimeTx+3;
- nTimeStamps[4] = nTimeTx+4;
- nTimeStamps[5] = nTimeTx+5;
- nTimeStamps[6] = nTimeTx+6;
- nTimeStamps[7] = nTimeTx+7;
-
- copyrow8_swap32(&blocks1[24], &nTimeStamps[0]); // Kernel timestamps
- sha256_transform_8way(&blocks2[0], &blocks1[0], 0); // first hashing
- sha256_transform_8way(&candidates[0], &blocks2[0], 0); // second hashing
- copyrow8_swap32(&nHashes[0], &candidates[56]);
-
- for(int nResult = 0; nResult < 8; nResult++)
- {
- if (nHashes[nResult] <= nMaxTarget32) // Possible hit
- {
- uint256 nHashProofOfStake = 0;
- uint32_t *pnHashProofOfStake = (uint32_t *) &nHashProofOfStake;
-
- for (int i = 0; i < 7; i++)
- pnHashProofOfStake[i] = __builtin_bswap32(candidates[(i*8) + nResult]);
- pnHashProofOfStake[7] = nHashes[nResult];
-
- CBigNum bnCoinDayWeight = bnValueIn * GetWeight((int64_t)nInputTxTime, (int64_t)nTimeStamps[nResult]) / COIN / nOneDay;
- CBigNum bnTargetProofOfStake = bnCoinDayWeight * bnTargetPerCoinDay;
-
- if (bnTargetProofOfStake >= CBigNum(nHashProofOfStake))
- solutions.push_back(std::pair<uint256,uint32_t>(nHashProofOfStake, nTimeStamps[nResult]));
- }
- }
- }
-}
-#endif
-
-void KernelWorker::Do_4way()
-{
- SetThreadPriority(THREAD_PRIORITY_LOWEST);
-
- // Compute maximum possible target to filter out majority of obviously insufficient hashes
- CBigNum bnTargetPerCoinDay;
- bnTargetPerCoinDay.SetCompact(nBits);
- uint256 nMaxTarget = (bnTargetPerCoinDay * bnValueIn * nStakeMaxAge / COIN / nOneDay).getuint256();
-
-#ifdef _MSC_VER
- __declspec(align(16)) uint32_t blocks1[4 * 16];
- __declspec(align(16)) uint32_t blocks2[4 * 16];
- __declspec(align(16)) uint32_t candidates[4 * 8];
-#else
- uint32_t blocks1[4 * 16] __attribute__((aligned(16)));
- uint32_t blocks2[4 * 16] __attribute__((aligned(16)));
- uint32_t candidates[4 * 8] __attribute__((aligned(16)));
-#endif
-
- vector<uint32_t> vRow = vector<uint32_t>(4);
- uint32_t *pnKernel = (uint32_t *) kernel;
-
- for(int i = 0; i < 7; i++)
- {
- fill(vRow.begin(), vRow.end(), pnKernel[i]);
- copyrow4_swap32(&blocks1[i*4], &vRow[0]);
- }
-
- memcpy(&blocks1[28], &block1_suffix_4way[0], 36*4); // sha256 padding
- memcpy(&blocks2[32], &block2_suffix_4way[0], 32*4);
-
- uint32_t nHashes[4];
- uint32_t nTimeStamps[4];
-
- // Search forward in time from the given timestamp
- // Stopping search in case of shutting down
- for (uint32_t nTimeTx=nIntervalBegin, nMaxTarget32 = nMaxTarget.Get32(7); nTimeTx<nIntervalEnd && !fShutdown; nTimeTx +=4)
- {
- sha256_init_4way(blocks2);
- sha256_init_4way(candidates);
-
- nTimeStamps[0] = nTimeTx;
- nTimeStamps[1] = nTimeTx+1;
- nTimeStamps[2] = nTimeTx+2;
- nTimeStamps[3] = nTimeTx+3;
-
- copyrow4_swap32(&blocks1[24], &nTimeStamps[0]); // Kernel timestamps
- sha256_transform_4way(&blocks2[0], &blocks1[0], 0); // first hashing
- sha256_transform_4way(&candidates[0], &blocks2[0], 0); // second hashing
- copyrow4_swap32(&nHashes[0], &candidates[28]);
-
- for(int nResult = 0; nResult < 4; nResult++)
- {
- if (nHashes[nResult] <= nMaxTarget32) // Possible hit
- {
- uint256 nHashProofOfStake = 0;
- uint32_t *pnHashProofOfStake = (uint32_t *) &nHashProofOfStake;
-
- for (int i = 0; i < 7; i++)
- pnHashProofOfStake[i] = __builtin_bswap32(candidates[(i*4) + nResult]);
- pnHashProofOfStake[7] = nHashes[nResult];
-
- CBigNum bnCoinDayWeight = bnValueIn * GetWeight((int64_t)nInputTxTime, (int64_t)nTimeStamps[nResult]) / COIN / nOneDay;
- CBigNum bnTargetProofOfStake = bnCoinDayWeight * bnTargetPerCoinDay;
-
- if (bnTargetProofOfStake >= CBigNum(nHashProofOfStake))
- solutions.push_back(std::pair<uint256,uint32_t>(nHashProofOfStake, nTimeStamps[nResult]));
- }
- }
- }
-}
-#endif
-
void KernelWorker::Do_generic()
{
SetThreadPriority(THREAD_PRIORITY_LOWEST);
bnTargetPerCoinDay.SetCompact(nBits);
uint256 nMaxTarget = (bnTargetPerCoinDay * bnValueIn * nStakeMaxAge / COIN / nOneDay).getuint256();
-#if !defined(USE_ASM) || defined(__i386__)
SHA256_CTX ctx, workerCtx;
// Init new sha256 context and update it
// with first 24 bytes of kernel
if (bnTargetProofOfStake >= CBigNum(*pnHashProofOfStake))
solutions.push_back(std::pair<uint256,uint32_t>(*pnHashProofOfStake, nTimeTx));
}
-#else
-
-#ifdef _MSC_VER
- __declspec(align(16)) uint32_t block1[16];
- __declspec(align(16)) uint32_t block2[16];
- __declspec(align(16)) uint32_t candidate[8];
-#else
- uint32_t block1[16] __attribute__((aligned(16)));
- uint32_t block2[16] __attribute__((aligned(16)));
- uint32_t candidate[8] __attribute__((aligned(16)));
-#endif
-
- memcpy(&block1[7], &block1_suffix[0], 36); // sha256 padding
- memcpy(&block2[8], &block2_suffix[0], 32);
-
- uint32_t *pnKernel = (uint32_t *) kernel;
-
- for (int i = 0; i < 6; i++)
- block1[i] = __builtin_bswap32(pnKernel[i]);
-
- // Search forward in time from the given timestamp
- // Stopping search in case of shutting down
- for (uint32_t nTimeTx=nIntervalBegin, nMaxTarget32 = nMaxTarget.Get32(7); nTimeTx<nIntervalEnd && !fShutdown; nTimeTx++)
- {
- memcpy(&block2[0], &sha256_initial[0], 32);
- memcpy(&candidate[0], &sha256_initial[0], 32);
-
- block1[6] = __builtin_bswap32(nTimeTx);
-
- sha256_transform(&block2[0], &block1[0], 0); // first hashing
- sha256_transform(&candidate[0], &block2[0], 0); // second hashing
-
- uint32_t nHash7 = __builtin_bswap32(candidate[7]);
-
- // Skip if hash doesn't satisfy the maximum target
- if (nHash7 > nMaxTarget32)
- continue;
-
- uint256 nHashProofOfStake;
- uint32_t *pnHashProofOfStake = (uint32_t *) &nHashProofOfStake;
-
- for (int i = 0; i < 7; i++)
- pnHashProofOfStake[i] = __builtin_bswap32(candidate[i]);
- pnHashProofOfStake[7] = nHash7;
-
- CBigNum bnCoinDayWeight = bnValueIn * GetWeight((int64_t)nInputTxTime, (int64_t)nTimeTx) / COIN / nOneDay;
- CBigNum bnTargetProofOfStake = bnCoinDayWeight * bnTargetPerCoinDay;
-
- if (bnTargetProofOfStake >= CBigNum(nHashProofOfStake))
- solutions.push_back(std::pair<uint256,uint32_t>(nHashProofOfStake, nTimeTx));
- }
-#endif
}
void KernelWorker::Do()
{
-#ifdef USE_ASM
-#ifdef __x86_64__
- if (false && fUse8Way) // disable for now
- {
- Do_8way();
- return;
- }
-#endif
- if (fUse4Way)
- {
- Do_4way();
- return;
- }
-#endif
-
Do_generic();
}
}
// Scan given kernel for solutions
-#ifdef USE_ASM
-
-#ifdef __x86_64__
-bool ScanKernelBackward_8Way(unsigned char *kernel, uint32_t nBits, uint32_t nInputTxTime, int64_t nValueIn, std::pair<uint32_t, uint32_t> &SearchInterval, std::pair<uint256, uint32_t> &solution)
-{
- CBigNum bnTargetPerCoinDay;
- bnTargetPerCoinDay.SetCompact(nBits);
-
- CBigNum bnValueIn(nValueIn);
-
- // Get maximum possible target to filter out the majority of obviously insufficient hashes
- uint256 nMaxTarget = (bnTargetPerCoinDay * bnValueIn * nStakeMaxAge / COIN / nOneDay).getuint256();
-
-#ifdef _MSC_VER
- __declspec(align(16)) uint32_t blocks1[8 * 16];
- __declspec(align(16)) uint32_t blocks2[8 * 16];
- __declspec(align(16)) uint32_t candidates[8 * 8];
-#else
- uint32_t blocks1[8 * 16] __attribute__((aligned(16)));
- uint32_t blocks2[8 * 16] __attribute__((aligned(16)));
- uint32_t candidates[8 * 8] __attribute__((aligned(16)));
-#endif
-
- vector<uint32_t> vRow = vector<uint32_t>(8);
- uint32_t *pnKernel = (uint32_t *) kernel;
-
- for(int i = 0; i < 7; i++)
- {
- fill(vRow.begin(), vRow.end(), pnKernel[i]);
- copyrow8_swap32(&blocks1[i*8], &vRow[0]);
- }
-
- memcpy(&blocks1[56], &block1_suffix_8way[0], 36*8); // sha256 padding
- memcpy(&blocks2[64], &block2_suffix_8way[0], 32*8);
-
- uint32_t nHashes[8];
- uint32_t nTimeStamps[8];
-
- // Search forward in time from the given timestamp
- // Stopping search in case of shutting down
- for (uint32_t nTimeTx=SearchInterval.first, nMaxTarget32 = nMaxTarget.Get32(7); nTimeTx>SearchInterval.second && !fShutdown; nTimeTx -=8)
- {
- sha256_init_8way(blocks2);
- sha256_init_8way(candidates);
-
- nTimeStamps[0] = nTimeTx;
- nTimeStamps[1] = nTimeTx-1;
- nTimeStamps[2] = nTimeTx-2;
- nTimeStamps[3] = nTimeTx-3;
- nTimeStamps[4] = nTimeTx-4;
- nTimeStamps[5] = nTimeTx-5;
- nTimeStamps[6] = nTimeTx-6;
- nTimeStamps[7] = nTimeTx-7;
-
- copyrow8_swap32(&blocks1[24], &nTimeStamps[0]); // Kernel timestamps
- sha256_transform_8way(&blocks2[0], &blocks1[0], 0); // first hashing
- sha256_transform_8way(&candidates[0], &blocks2[0], 0); // second hashing
- copyrow8_swap32(&nHashes[0], &candidates[56]);
-
- for(int nResult = 0; nResult < 8; nResult++)
- {
- if (nHashes[nResult] <= nMaxTarget32) // Possible hit
- {
- uint256 nHashProofOfStake = 0;
- uint32_t *pnHashProofOfStake = (uint32_t *) &nHashProofOfStake;
-
- for (int i = 0; i < 7; i++)
- pnHashProofOfStake[i] = __builtin_bswap32(candidates[(i*8) + nResult]);
- pnHashProofOfStake[7] = nHashes[nResult];
-
- CBigNum bnCoinDayWeight = bnValueIn * GetWeight((int64_t)nInputTxTime, (int64_t)nTimeStamps[nResult]) / COIN / nOneDay;
- CBigNum bnTargetProofOfStake = bnCoinDayWeight * bnTargetPerCoinDay;
-
- if (bnTargetProofOfStake >= CBigNum(nHashProofOfStake))
- {
- solution.first = nHashProofOfStake;
- solution.second = nTimeStamps[nResult];
-
- return true;
- }
- }
- }
- }
-
- return false;
-}
-#endif
-
-bool ScanKernelBackward_4Way(unsigned char *kernel, uint32_t nBits, uint32_t nInputTxTime, int64_t nValueIn, std::pair<uint32_t, uint32_t> &SearchInterval, std::pair<uint256, uint32_t> &solution)
-{
- CBigNum bnTargetPerCoinDay;
- bnTargetPerCoinDay.SetCompact(nBits);
-
- CBigNum bnValueIn(nValueIn);
-
- // Get maximum possible target to filter out the majority of obviously insufficient hashes
- uint256 nMaxTarget = (bnTargetPerCoinDay * bnValueIn * nStakeMaxAge / COIN / nOneDay).getuint256();
-
-#ifdef _MSC_VER
- __declspec(align(16)) uint32_t blocks1[4 * 16];
- __declspec(align(16)) uint32_t blocks2[4 * 16];
- __declspec(align(16)) uint32_t candidates[4 * 8];
-#else
- uint32_t blocks1[4 * 16] __attribute__((aligned(16)));
- uint32_t blocks2[4 * 16] __attribute__((aligned(16)));
- uint32_t candidates[4 * 8] __attribute__((aligned(16)));
-#endif
-
- vector<uint32_t> vRow = vector<uint32_t>(4);
- uint32_t *pnKernel = (uint32_t *) kernel;
-
- for(int i = 0; i < 7; i++)
- {
- fill(vRow.begin(), vRow.end(), pnKernel[i]);
- copyrow4_swap32(&blocks1[i*4], &vRow[0]);
- }
-
- memcpy(&blocks1[28], &block1_suffix_4way[0], 36*4); // sha256 padding
- memcpy(&blocks2[32], &block2_suffix_4way[0], 32*4);
-
- uint32_t nHashes[4];
- uint32_t nTimeStamps[4];
-
- // Search forward in time from the given timestamp
- // Stopping search in case of shutting down
- for (uint32_t nTimeTx=SearchInterval.first, nMaxTarget32 = nMaxTarget.Get32(7); nTimeTx>SearchInterval.second && !fShutdown; nTimeTx -=4)
- {
- sha256_init_4way(blocks2);
- sha256_init_4way(candidates);
-
- nTimeStamps[0] = nTimeTx;
- nTimeStamps[1] = nTimeTx-1;
- nTimeStamps[2] = nTimeTx-2;
- nTimeStamps[3] = nTimeTx-3;
-
- copyrow4_swap32(&blocks1[24], &nTimeStamps[0]); // Kernel timestamps
- sha256_transform_4way(&blocks2[0], &blocks1[0], 0); // first hashing
- sha256_transform_4way(&candidates[0], &blocks2[0], 0); // second hashing
- copyrow4_swap32(&nHashes[0], &candidates[28]);
-
- for(int nResult = 0; nResult < 4; nResult++)
- {
- if (nHashes[nResult] <= nMaxTarget32) // Possible hit
- {
- uint256 nHashProofOfStake = 0;
- uint32_t *pnHashProofOfStake = (uint32_t *) &nHashProofOfStake;
-
- for (int i = 0; i < 7; i++)
- pnHashProofOfStake[i] = __builtin_bswap32(candidates[(i*4) + nResult]);
- pnHashProofOfStake[7] = nHashes[nResult];
-
- CBigNum bnCoinDayWeight = bnValueIn * GetWeight((int64_t)nInputTxTime, (int64_t)nTimeStamps[nResult]) / COIN / nOneDay;
- CBigNum bnTargetProofOfStake = bnCoinDayWeight * bnTargetPerCoinDay;
-
- if (bnTargetProofOfStake >= CBigNum(nHashProofOfStake))
- {
- solution.first = nHashProofOfStake;
- solution.second = nTimeStamps[nResult];
-
- return true;
- }
- }
- }
- }
-
- return false;
-}
-#endif
bool ScanKernelBackward(unsigned char *kernel, uint32_t nBits, uint32_t nInputTxTime, int64_t nValueIn, std::pair<uint32_t, uint32_t> &SearchInterval, std::pair<uint256, uint32_t> &solution)
{
-#ifdef USE_ASM
-#ifdef __x86_64__
- if (false && fUse8Way) // disable for now
- {
- return ScanKernelBackward_8Way(kernel, nBits, nInputTxTime, nValueIn, SearchInterval, solution);
- }
-#endif
- if (fUse4Way)
- {
- return ScanKernelBackward_4Way(kernel, nBits, nInputTxTime, nValueIn, SearchInterval, solution);
- }
-#endif
-
CBigNum bnTargetPerCoinDay;
bnTargetPerCoinDay.SetCompact(nBits);
git://git.novaco.in / novacoin.git / commitdiff