1 #include "ec_base_vsx.h" 2 3 void gf_5vect_dot_prod_vsx(int len, int vlen, unsigned char *gftbls, 4 unsigned char **src, unsigned char **dest) 5 { 6 unsigned char *s, *t0, *t1, *t2, *t3, *t4; 7 vector unsigned char vX1, vX2, vX3, vX4; 8 vector unsigned char vY1, vY2, vY3, vY4, vY5, vY6, vY7, vY8, vY9, vYA; 9 vector unsigned char vYD, vYE, vYF, vYG, vYH, vYI, vYJ, vYK, vYL, vYM; 10 vector unsigned char vhi0, vlo0, vhi1, vlo1, vhi2, vlo2, vhi3, vlo3, vhi4, vlo4; 11 int i, j, head; 12 13 if (vlen < 128) { 14 gf_vect_mul_vsx(len, &gftbls[0 * 32 * vlen], src[0], (unsigned char *)dest[0]); 15 gf_vect_mul_vsx(len, &gftbls[1 * 32 * vlen], src[0], (unsigned char *)dest[1]); 16 gf_vect_mul_vsx(len, &gftbls[2 * 32 * vlen], src[0], (unsigned char *)dest[2]); 17 gf_vect_mul_vsx(len, &gftbls[3 * 32 * vlen], src[0], (unsigned char *)dest[3]); 18 gf_vect_mul_vsx(len, &gftbls[4 * 32 * vlen], src[0], (unsigned char *)dest[4]); 19 20 for (j = 1; j < vlen; j++) { 21 gf_5vect_mad_vsx(len, vlen, j, gftbls, src[j], dest); 22 } 23 return; 24 } 25 26 t0 = (unsigned char *)dest[0]; 27 t1 = (unsigned char *)dest[1]; 28 t2 = (unsigned char *)dest[2]; 29 t3 = (unsigned char *)dest[3]; 30 t4 = (unsigned char *)dest[4]; 31 32 head = len % 64; 33 if (head != 0) { 34 gf_vect_dot_prod_base(head, vlen, &gftbls[0 * 32 * vlen], src, t0); 35 gf_vect_dot_prod_base(head, vlen, &gftbls[1 * 32 * vlen], src, t1); 36 gf_vect_dot_prod_base(head, vlen, &gftbls[2 * 32 * vlen], src, t2); 37 gf_vect_dot_prod_base(head, vlen, &gftbls[3 * 32 * vlen], src, t3); 38 gf_vect_dot_prod_base(head, vlen, &gftbls[4 * 32 * vlen], src, t4); 39 } 40 41 for (i = head; i < len - 63; i += 64) { 42 vY1 = vY1 ^ vY1; 43 vY2 = vY2 ^ vY2; 44 vY3 = vY3 ^ vY3; 45 vY4 = vY4 ^ vY4; 46 vY5 = vY5 ^ vY5; 47 vY6 = vY6 ^ vY6; 48 vY7 = vY7 ^ vY7; 49 vY8 = vY8 ^ vY8; 50 vY9 = vY9 ^ vY9; 51 vYA = vYA ^ vYA; 52 53 vYD = vYD ^ vYD; 54 vYE = vYE ^ vYE; 55 vYF = vYF ^ vYF; 56 vYG = vYG ^ vYG; 57 vYH = vYH ^ vYH; 58 vYI = vYI ^ vYI; 59 vYJ = vYJ ^ vYJ; 60 vYK = vYK ^ vYK; 61 vYL = vYL ^ vYL; 62 vYM = vYM ^ vYM; 63 64 unsigned char *g0 = &gftbls[0 * 32 * vlen]; 65 unsigned char *g1 = &gftbls[1 * 32 * vlen]; 66 unsigned char *g2 = &gftbls[2 * 32 * vlen]; 67 unsigned char *g3 = &gftbls[3 * 32 * vlen]; 68 unsigned char *g4 = &gftbls[4 * 32 * vlen]; 69 70 for (j = 0; j < vlen; j++) { 71 s = (unsigned char *)src[j]; 72 vX1 = vec_xl(0, s + i); 73 vX2 = vec_xl(16, s + i); 74 vX3 = vec_xl(32, s + i); 75 vX4 = vec_xl(48, s + i); 76 77 vlo0 = EC_vec_xl(0, g0); 78 vhi0 = EC_vec_xl(16, g0); 79 vlo1 = EC_vec_xl(0, g1); 80 vhi1 = EC_vec_xl(16, g1); 81 82 vY1 = vY1 ^ EC_vec_permxor(vhi0, vlo0, vX1); 83 vY2 = vY2 ^ EC_vec_permxor(vhi0, vlo0, vX2); 84 vYD = vYD ^ EC_vec_permxor(vhi0, vlo0, vX3); 85 vYE = vYE ^ EC_vec_permxor(vhi0, vlo0, vX4); 86 87 vlo2 = vec_xl(0, g2); 88 vhi2 = vec_xl(16, g2); 89 vlo3 = vec_xl(0, g3); 90 vhi3 = vec_xl(16, g3); 91 92 vY3 = vY3 ^ EC_vec_permxor(vhi1, vlo1, vX1); 93 vY4 = vY4 ^ EC_vec_permxor(vhi1, vlo1, vX2); 94 vYF = vYF ^ EC_vec_permxor(vhi1, vlo1, vX3); 95 vYG = vYG ^ EC_vec_permxor(vhi1, vlo1, vX4); 96 97 vlo4 = vec_xl(0, g4); 98 vhi4 = vec_xl(16, g4); 99 100 vY5 = vY5 ^ EC_vec_permxor(vhi2, vlo2, vX1); 101 vY6 = vY6 ^ EC_vec_permxor(vhi2, vlo2, vX2); 102 vYH = vYH ^ EC_vec_permxor(vhi2, vlo2, vX3); 103 vYI = vYI ^ EC_vec_permxor(vhi2, vlo2, vX4); 104 105 vY7 = vY7 ^ EC_vec_permxor(vhi3, vlo3, vX1); 106 vY8 = vY8 ^ EC_vec_permxor(vhi3, vlo3, vX2); 107 vYJ = vYJ ^ EC_vec_permxor(vhi3, vlo3, vX3); 108 vYK = vYK ^ EC_vec_permxor(vhi3, vlo3, vX4); 109 110 vY9 = vY9 ^ EC_vec_permxor(vhi4, vlo4, vX1); 111 vYA = vYA ^ EC_vec_permxor(vhi4, vlo4, vX2); 112 vYL = vYL ^ EC_vec_permxor(vhi4, vlo4, vX3); 113 vYM = vYM ^ EC_vec_permxor(vhi4, vlo4, vX4); 114 115 g0 += 32; 116 g1 += 32; 117 g2 += 32; 118 g3 += 32; 119 g4 += 32; 120 } 121 122 vec_xst(vY1, 0, t0 + i); 123 vec_xst(vY2, 16, t0 + i); 124 vec_xst(vY3, 0, t1 + i); 125 vec_xst(vY4, 16, t1 + i); 126 vec_xst(vY5, 0, t2 + i); 127 vec_xst(vY6, 16, t2 + i); 128 vec_xst(vY7, 0, t3 + i); 129 vec_xst(vY8, 16, t3 + i); 130 vec_xst(vY9, 0, t4 + i); 131 vec_xst(vYA, 16, t4 + i); 132 133 vec_xst(vYD, 32, t0 + i); 134 vec_xst(vYE, 48, t0 + i); 135 vec_xst(vYF, 32, t1 + i); 136 vec_xst(vYG, 48, t1 + i); 137 vec_xst(vYH, 32, t2 + i); 138 vec_xst(vYI, 48, t2 + i); 139 vec_xst(vYJ, 32, t3 + i); 140 vec_xst(vYK, 48, t3 + i); 141 vec_xst(vYL, 32, t4 + i); 142 vec_xst(vYM, 48, t4 + i); 143 } 144 return; 145 } 146