Lines Matching +full:1 +full:b
28 #define B(n) MMX_B(n) macro
36 #define B(n) ZMM_B(n) macro
40 #if SHIFT == 1
61 for (int i = 0; i < 1 << SHIFT; i++) { in glue()
65 shift = c->B(0); in glue()
76 for (int i = 0; i < 1 << SHIFT; i++) { in glue()
80 shift = c->B(0); in glue()
93 shift = c->B(0); in glue()
104 for (int i = 0; i < 1 << SHIFT; i++) { in glue()
108 shift = c->B(0); in glue()
119 for (int i = 0; i < 1 << SHIFT; i++) { in glue()
123 shift = c->B(0); in glue()
136 shift = c->B(0); in glue()
147 for (int i = 0; i < 1 << SHIFT; i++) { in glue()
151 shift = c->B(0); in glue()
152 for (int i = 0; i < 1 << SHIFT; i++) { in glue()
162 for (int i = 0; i < 1 << SHIFT; i++) { in glue()
166 shift = c->B(0); in glue()
167 for (int i = 0; i < 1 << SHIFT; i++) { in glue()
173 #if SHIFT >= 1
184 d->B(j + i) = s->B(j + i + shift); in glue()
187 d->B(j + i) = 0; in glue()
202 d->B(j + i) = s->B(j + i - shift); in glue()
205 d->B(j + i) = 0; in glue()
230 SSE_HELPER_2(name, B, 8 << SHIFT, F)
239 SSE_HELPER_2(name, Q, 1 << SHIFT, F)
286 #define FADD(a, b) ((a) + (b)) argument
287 #define FADDUB(a, b) satub((a) + (b)) argument
288 #define FADDUW(a, b) satuw((a) + (b)) argument
289 #define FADDSB(a, b) satsb((int8_t)(a) + (int8_t)(b)) argument
290 #define FADDSW(a, b) satsw((int16_t)(a) + (int16_t)(b)) argument
292 #define FSUB(a, b) ((a) - (b)) argument
293 #define FSUBUB(a, b) satub((a) - (b)) argument
294 #define FSUBUW(a, b) satuw((a) - (b)) argument
295 #define FSUBSB(a, b) satsb((int8_t)(a) - (int8_t)(b)) argument
296 #define FSUBSW(a, b) satsw((int16_t)(a) - (int16_t)(b)) argument
297 #define FMINUB(a, b) ((a) < (b)) ? (a) : (b) argument
298 #define FMINSW(a, b) ((int16_t)(a) < (int16_t)(b)) ? (a) : (b) argument
299 #define FMAXUB(a, b) ((a) > (b)) ? (a) : (b) argument
300 #define FMAXSW(a, b) ((int16_t)(a) > (int16_t)(b)) ? (a) : (b) argument
302 #define FMULHRW(a, b) (((int16_t)(a) * (int16_t)(b) + 0x8000) >> 16) argument
303 #define FMULHUW(a, b) ((a) * (b) >> 16) argument
304 #define FMULHW(a, b) ((int16_t)(a) * (int16_t)(b) >> 16) argument
306 #define FAVG(a, b) (((a) + (b) + 1) >> 1) argument
316 d->W(1) = FMULHRW(d->W(1), s->W(1)); in SSE_HELPER_W()
329 for (i = 0; i < (1 << SHIFT); i++) { in SSE_HELPER_B()
340 (int16_t)s->W(2 * i + 1) * (int16_t)v->W(2 * i + 1); in glue()
358 for (i = 0; i < (1 << SHIFT); i++) { in glue()
360 val += abs1(v->B(8 * i + 0) - s->B(8 * i + 0)); in glue()
361 val += abs1(v->B(8 * i + 1) - s->B(8 * i + 1)); in glue()
362 val += abs1(v->B(8 * i + 2) - s->B(8 * i + 2)); in glue()
363 val += abs1(v->B(8 * i + 3) - s->B(8 * i + 3)); in glue()
364 val += abs1(v->B(8 * i + 4) - s->B(8 * i + 4)); in glue()
365 val += abs1(v->B(8 * i + 5) - s->B(8 * i + 5)); in glue()
366 val += abs1(v->B(8 * i + 6) - s->B(8 * i + 6)); in glue()
367 val += abs1(v->B(8 * i + 7) - s->B(8 * i + 7)); in glue()
379 if (s->B(i) & 0x80) { in glue()
380 cpu_stb_data_ra(env, a0 + i, d->B(i), GETPC()); in glue()
386 #define SHUFFLE4(F, a, b, offset) do { \ argument
389 r2 = b->F(((order >> 4) & 3) + offset); \
390 r3 = b->F(((order >> 6) & 3) + offset); \
392 d->F(offset + 1) = r1; \
420 for (i = 0; i < 1 << SHIFT; i += 2) { in glue()
421 r0 = v->Q(((order & 1) & 1) + i); in glue()
422 r1 = s->Q(((order >> 1) & 1) + i); in glue()
424 d->Q(i + 1) = r1; in glue()
444 for (i = 0, j = 1; j < 1 << SHIFT; i += 8, j += 2) { in glue()
455 for (i = 4, j = 0; j < 1 << SHIFT; i += 8, j += 2) { in glue()
462 #if SHIFT >= 1
480 for (i = 0; i < 1 << SHIFT; i++) { \
485 #if SHIFT == 1
494 for (i = 1; i < 2 << SHIFT; i++) { \
503 for (i = 1; i < 1 << SHIFT; i++) { \
514 #define FPU_ADD(size, a, b) float ## size ## _add(a, b, &env->sse_status) argument
515 #define FPU_SUB(size, a, b) float ## size ## _sub(a, b, &env->sse_status) argument
516 #define FPU_MUL(size, a, b) float ## size ## _mul(a, b, &env->sse_status) argument
517 #define FPU_DIV(size, a, b) float ## size ## _div(a, b, &env->sse_status) argument
523 #define FPU_MIN(size, a, b) \ argument
524 (float ## size ## _lt(a, b, &env->sse_status) ? (a) : (b))
525 #define FPU_MAX(size, a, b) \ argument
526 (float ## size ## _lt(b, a, &env->sse_status) ? (a) : (b))
546 for (i = 0; i < 1 << SHIFT; i++) { in glue()
551 #if SHIFT == 1
556 for (i = 1; i < 2 << SHIFT; i++) { in helper_sqrtss()
565 for (i = 1; i < 1 << SHIFT; i++) { in helper_sqrtsd()
575 for (i = 1 << SHIFT; --i >= 0; ) { in glue()
583 for (i = 0; i < 1 << SHIFT; i++) { in glue()
586 for (i >>= 1; i < 1 << SHIFT; i++) { in glue()
591 #if SHIFT >= 1
605 if (!(mode & (1 << 2))) { in glue()
612 for (i >>= 2; i < 1 << SHIFT; i++) { in glue()
620 #if SHIFT == 1
625 for (i = 1; i < 1 << SHIFT; i++) { in helper_cvtss2sd()
634 for (i = 1; i < 2 << SHIFT; i++) { in helper_cvtsd2ss()
652 for (i = 1 << SHIFT; --i >= 0; ) { in glue()
658 #if SHIFT == 1
662 d->ZMM_S(1) = int32_to_float32(s->MMX_L(1), &env->sse_status); in helper_cvtpi2ps()
668 d->ZMM_D(1) = int32_to_float64(s->MMX_L(1), &env->sse_status); in helper_cvtpi2pd()
697 #if SHIFT == 1
741 for (i = 0; i < 1 << SHIFT; i++) { in glue()
744 for (i >>= 1; i < 1 << SHIFT; i++) { in glue()
749 #if SHIFT == 1
753 d->MMX_L(1) = x86_float32_to_int32(s->ZMM_S(1), &env->sse_status); in helper_cvtps2pi()
759 d->MMX_L(1) = x86_float64_to_int32(s->ZMM_D(1), &env->sse_status); in helper_cvtpd2pi()
798 for (i = 0; i < 1 << SHIFT; i++) { in glue()
802 for (i >>= 1; i < 1 << SHIFT; i++) { in glue()
807 #if SHIFT == 1
811 d->MMX_L(1) = x86_float32_to_int32_round_to_zero(s->ZMM_S(1), &env->sse_status); in helper_cvttps2pi()
817 d->MMX_L(1) = x86_float64_to_int32_round_to_zero(s->ZMM_D(1), &env->sse_status); in helper_cvttpd2pi()
855 #if SHIFT == 1
864 for (i = 1; i < 2 << SHIFT; i++) { in helper_rsqrtss()
880 #if SHIFT == 1
886 for (i = 1; i < 2 << SHIFT; i++) { in helper_rcpss()
893 #if SHIFT == 1
901 mask = (1ULL << len) - 1; in helper_extrq()
908 d->ZMM_Q(0) = helper_extrq(d->ZMM_Q(0), s->ZMM_B(1) & 63, s->ZMM_B(0) & 63); in helper_extrq_r()
923 mask = (1ULL << len) - 1; in helper_insertq()
946 r[i + k] = F(v->ZMM_S(j + k), v->ZMM_S(j + k + 1), &env->sse_status); \
949 r[i + k] = F(s->ZMM_S(j + k), s->ZMM_S(j + k + 1), &env->sse_status); \
963 float64 r[1 << SHIFT]; \ in SSE_HELPER_HPS()
965 for (k = 0; k < 1 << SHIFT; k += LANE_WIDTH / 8) { \ in SSE_HELPER_HPS()
967 r[i + k] = F(v->ZMM_D(j + k), v->ZMM_D(j + k + 1), &env->sse_status); \ in SSE_HELPER_HPS()
970 r[i + k] = F(s->ZMM_D(j + k), s->ZMM_D(j + k + 1), &env->sse_status); \ in SSE_HELPER_HPS()
973 for (i = 0; i < 1 << SHIFT; i++) { \ in SSE_HELPER_HPS()
986 d->ZMM_S(i+1) = float32_add(v->ZMM_S(i+1), s->ZMM_S(i+1), &env->sse_status);
993 for (i = 0; i < 1 << SHIFT; i += 2) { in glue()
995 d->ZMM_D(i+1) = float64_add(v->ZMM_D(i+1), s->ZMM_D(i+1), &env->sse_status); in glue()
1005 d->ZMM_L(i) = C(F(32, v->ZMM_S(i), s->ZMM_S(i))) ? -1 : 0; \
1013 for (i = 0; i < 1 << SHIFT; i++) { \
1014 d->ZMM_Q(i) = C(F(64, v->ZMM_D(i), s->ZMM_D(i))) ? -1 : 0; \
1018 #if SHIFT == 1
1024 d->ZMM_L(0) = C(F(32, v->ZMM_S(0), s->ZMM_S(0))) ? -1 : 0; \
1025 for (i = 1; i < 2 << SHIFT; i++) { \
1033 d->ZMM_Q(0) = C(F(64, v->ZMM_D(0), s->ZMM_D(0))) ? -1 : 0; \
1034 for (i = 1; i < 1 << SHIFT; i++) { \
1055 #define FPU_CMPQ(size, a, b) \ argument
1056 float ## size ## _compare_quiet(a, b, &env->sse_status)
1057 #define FPU_CMPS(size, a, b) \ argument
1058 float ## size ## _compare(a, b, &env->sse_status)
1102 #if SHIFT == 1
1113 CC_SRC = comis_eflags[ret + 1]; in helper_ucomiss()
1125 CC_SRC = comis_eflags[ret + 1]; in helper_comiss()
1137 CC_SRC = comis_eflags[ret + 1]; in helper_ucomisd()
1149 CC_SRC = comis_eflags[ret + 1]; in helper_comisd()
1161 mask |= (s->ZMM_L(i) >> (31 - i)) & (1 << i); in glue()
1172 for (i = 0; i < 1 << SHIFT; i++) { in glue()
1173 mask |= (s->ZMM_Q(i) >> (63 - i)) & (1 << i); in glue()
1194 d->B(2 * j + k) = r[k]; \
1231 r[2 * i] = v->B(k + i); \
1232 r[2 * i + 1] = s->B(k + i); \
1235 d->B(j) = r[i]; \
1250 r[2 * i + 1] = s->W(k + i); \
1268 r[2 * i + 1] = s->L(k + i); \
1283 for (i = 0; i < 1 << SHIFT; i += 2) { \
1285 r[1] = s->Q(base + i); \
1287 d->Q(i + 1) = r[1]; \
1293 UNPCK_OP(h, 1)
1305 d->MMX_S(1) = int32_to_float32(s->MMX_L(1), &env->mmx_status); in helper_pi2fd()
1311 d->MMX_S(1) = int32_to_float32((int16_t)s->MMX_W(2), &env->mmx_status); in helper_pi2fw()
1317 d->MMX_L(1) = float32_to_int32_round_to_zero(s->MMX_S(1), &env->mmx_status); in helper_pf2id()
1324 d->MMX_L(1) = satsw(float32_to_int32_round_to_zero(s->MMX_S(1), in helper_pf2iw()
1332 r = float32_add(d->MMX_S(0), d->MMX_S(1), &env->mmx_status); in helper_pfacc()
1333 d->MMX_S(1) = float32_add(s->MMX_S(0), s->MMX_S(1), &env->mmx_status); in helper_pfacc()
1340 d->MMX_S(1) = float32_add(d->MMX_S(1), s->MMX_S(1), &env->mmx_status); in helper_pfadd()
1346 &env->mmx_status) ? -1 : 0; in helper_pfcmpeq()
1347 d->MMX_L(1) = float32_eq_quiet(d->MMX_S(1), s->MMX_S(1), in helper_pfcmpeq()
1348 &env->mmx_status) ? -1 : 0; in helper_pfcmpeq()
1354 &env->mmx_status) ? -1 : 0; in helper_pfcmpge()
1355 d->MMX_L(1) = float32_le(s->MMX_S(1), d->MMX_S(1), in helper_pfcmpge()
1356 &env->mmx_status) ? -1 : 0; in helper_pfcmpge()
1362 &env->mmx_status) ? -1 : 0; in helper_pfcmpgt()
1363 d->MMX_L(1) = float32_lt(s->MMX_S(1), d->MMX_S(1), in helper_pfcmpgt()
1364 &env->mmx_status) ? -1 : 0; in helper_pfcmpgt()
1372 if (float32_lt(d->MMX_S(1), s->MMX_S(1), &env->mmx_status)) { in helper_pfmax()
1373 d->MMX_S(1) = s->MMX_S(1); in helper_pfmax()
1382 if (float32_lt(s->MMX_S(1), d->MMX_S(1), &env->mmx_status)) { in helper_pfmin()
1383 d->MMX_S(1) = s->MMX_S(1); in helper_pfmin()
1390 d->MMX_S(1) = float32_mul(d->MMX_S(1), s->MMX_S(1), &env->mmx_status); in helper_pfmul()
1397 r = float32_sub(d->MMX_S(0), d->MMX_S(1), &env->mmx_status); in helper_pfnacc()
1398 d->MMX_S(1) = float32_sub(s->MMX_S(0), s->MMX_S(1), &env->mmx_status); in helper_pfnacc()
1406 r = float32_sub(d->MMX_S(0), d->MMX_S(1), &env->mmx_status); in helper_pfpnacc()
1407 d->MMX_S(1) = float32_add(s->MMX_S(0), s->MMX_S(1), &env->mmx_status); in helper_pfpnacc()
1414 d->MMX_S(1) = d->MMX_S(0); in helper_pfrcp()
1419 d->MMX_L(1) = s->MMX_L(0) & 0x7fffffff; in helper_pfrsqrt()
1420 d->MMX_S(1) = float32_div(float32_one, in helper_pfrsqrt()
1421 float32_sqrt(d->MMX_S(1), &env->mmx_status), in helper_pfrsqrt()
1423 d->MMX_L(1) |= s->MMX_L(0) & 0x80000000; in helper_pfrsqrt()
1424 d->MMX_L(0) = d->MMX_L(1); in helper_pfrsqrt()
1430 d->MMX_S(1) = float32_sub(d->MMX_S(1), s->MMX_S(1), &env->mmx_status); in helper_pfsub()
1436 d->MMX_S(1) = float32_sub(s->MMX_S(1), d->MMX_S(1), &env->mmx_status); in helper_pfsubr()
1444 d->MMX_L(0) = s->MMX_L(1); in helper_pswapd()
1445 d->MMX_L(1) = r; in helper_pswapd()
1457 r[i] = (s->B(i) & 0x80) ? 0 : (v->B(s->B(i) & 7)); in glue()
1460 d->B(i) = r[i]; in glue()
1467 r[i] = (s->B(i) & 0x80) ? 0 : v->B(j | (s->B(i) & 0xf)); in glue()
1470 d->B(i) = r[i]; in glue()
1482 r[i + k] = F(v->W(j + k), v->W(j + k + 1)); \
1485 r[i + k] = F(s->W(j + k), s->W(j + k + 1)); \
1500 r[i + k] = F(v->L(j + k), v->L(j + k + 1)); \
1503 r[i + k] = F(s->L(j + k), s->L(j + k + 1)); \
1525 d->W(i) = satsw((int8_t)s->B(i * 2) * (uint8_t)v->B(i * 2) + in SSE_HELPER_HW()
1526 (int8_t)s->B(i * 2 + 1) * (uint8_t)v->B(i * 2 + 1)); in SSE_HELPER_HW()
1547 for (i = 0; i < (1 << SHIFT); i++) { in SSE_HELPER_W()
1557 for (i = 0; i < (1 << SHIFT); i += 2) { in SSE_HELPER_W()
1561 SHR(s->Q(i + 1), shift - 64) | in SSE_HELPER_W()
1563 SHR(v->Q(i + 1), shift - 192); in SSE_HELPER_W()
1565 SHR(s->Q(i + 1), shift - 0) | in SSE_HELPER_W()
1567 SHR(v->Q(i + 1), shift - 128); in SSE_HELPER_W()
1569 d->Q(i + 1) = r1; in SSE_HELPER_W()
1576 #if SHIFT >= 1
1595 d->elem(i) = F(v->elem(i), s->elem(i), (imm >> j) & 1); \
1599 /* SSE4.1 op helpers */
1603 SSE_HELPER_V(helper_pblendvb, B, 8 << SHIFT, FBLENDVB)
1605 SSE_HELPER_V(helper_blendvpd, Q, 1 << SHIFT, FBLENDVPD)
1612 for (i = 0; i < 1 << SHIFT; i++) { in glue()
1620 #define FMOVSLDUP(i) s->L((i) & ~1)
1621 #define FMOVSHDUP(i) s->L((i) | 1)
1622 #define FMOVDLDUP(i) s->Q((i) & ~1)
1634 SSE_HELPER_F(helper_pmovsxbw, W, 4 << SHIFT, (int8_t) s->B)
1635 SSE_HELPER_F(helper_pmovsxbd, L, 2 << SHIFT, (int8_t) s->B)
1636 SSE_HELPER_F(helper_pmovsxbq, Q, 1 << SHIFT, (int8_t) s->B)
1638 SSE_HELPER_F(helper_pmovsxwq, Q, 1 << SHIFT, (int16_t) s->W)
1639 SSE_HELPER_F(helper_pmovsxdq, Q, 1 << SHIFT, (int32_t) s->L)
1640 SSE_HELPER_F(helper_pmovzxbw, W, 4 << SHIFT, s->B)
1641 SSE_HELPER_F(helper_pmovzxbd, L, 2 << SHIFT, s->B)
1642 SSE_HELPER_F(helper_pmovzxbq, Q, 1 << SHIFT, s->B)
1644 SSE_HELPER_F(helper_pmovzxwq, Q, 1 << SHIFT, s->W)
1645 SSE_HELPER_F(helper_pmovzxdq, Q, 1 << SHIFT, s->L)
1648 SSE_HELPER_F(helper_pmovdldup, Q, 1 << SHIFT, FMOVDLDUP)
1655 for (i = 0; i < 1 << SHIFT; i++) { in glue()
1667 r[1] = satuw(v->L(j + 1)); in glue()
1671 r[5] = satuw(s->L(j + 1)); in glue()
1680 #if SHIFT == 1
1685 if (s->W(1) < s->W(idx)) { in glue()
1686 idx = 1; in glue()
1708 d->W(1) = idx; in glue()
1709 d->L(1) = 0; in glue()
1710 d->Q(1) = 0; in glue()
1722 if (!(mode & (1 << 2))) { in glue()
1730 if (mode & (1 << 3) && !(old_flags & float_flag_inexact)) { in glue()
1746 if (!(mode & (1 << 2))) { in glue()
1750 for (i = 0; i < 1 << SHIFT; i++) { in glue()
1754 if (mode & (1 << 3) && !(old_flags & float_flag_inexact)) { in glue()
1762 #if SHIFT == 1
1771 if (!(mode & (1 << 2))) { in glue()
1776 for (i = 1; i < 2 << SHIFT; i++) { in glue()
1780 if (mode & (1 << 3) && !(old_flags & float_flag_inexact)) { in glue()
1796 if (!(mode & (1 << 2))) { in glue()
1801 for (i = 1; i < 1 << SHIFT; i++) { in glue()
1805 if (mode & (1 << 3) && !(old_flags & float_flag_inexact)) { in glue()
1816 SSE_HELPER_I(helper_blendpd, Q, 1 << SHIFT, FBLENDP)
1827 * We must evaluate (A+B)+(C+D), not ((A+B)+C)+D in glue()
1830 if (mask & (1 << 4)) { in glue()
1835 if (mask & (1 << 5)) { in glue()
1836 prod2 = float32_mul(v->ZMM_S(i+1), s->ZMM_S(i+1), &env->sse_status); in glue()
1841 if (mask & (1 << 6)) { in glue()
1846 if (mask & (1 << 7)) { in glue()
1854 d->ZMM_S(i) = (mask & (1 << 0)) ? temp4 : float32_zero; in glue()
1855 d->ZMM_S(i+1) = (mask & (1 << 1)) ? temp4 : float32_zero; in glue()
1856 d->ZMM_S(i+2) = (mask & (1 << 2)) ? temp4 : float32_zero; in glue()
1857 d->ZMM_S(i+3) = (mask & (1 << 3)) ? temp4 : float32_zero; in glue()
1861 #if SHIFT == 1
1868 if (mask & (1 << 4)) { in glue()
1873 if (mask & (1 << 5)) { in glue()
1874 prod2 = float64_mul(v->ZMM_D(1), s->ZMM_D(1), &env->sse_status); in glue()
1879 d->ZMM_D(0) = (mask & (1 << 0)) ? temp2 : float64_zero; in glue()
1880 d->ZMM_D(1) = (mask & (1 << 1)) ? temp2 : float64_zero; in glue()
1895 r[i] += abs1(v->B(d0 + 0) - s->B(s0 + 0)); in glue()
1896 r[i] += abs1(v->B(d0 + 1) - s->B(s0 + 1)); in glue()
1897 r[i] += abs1(v->B(d0 + 2) - s->B(s0 + 2)); in glue()
1898 r[i] += abs1(v->B(d0 + 3) - s->B(s0 + 3)); in glue()
1908 #if SHIFT == 1
1919 if (ctrl & 1) { in pcmp_elen()
1934 if (ctrl & 1) { in pcmp_ilen()
1939 while (val < 16 && r->B(val)) { in pcmp_ilen()
1951 return r->B(i); in pcmp_val()
1952 case 1: in pcmp_val()
1955 return (int8_t)r->B(i); in pcmp_val()
1968 int upper = (ctrl & 1) ? 7 : 15; in pcmpxstrx()
1979 res <<= 1; in pcmpxstrx()
1986 case 1: in pcmpxstrx()
1988 res <<= 1; in pcmpxstrx()
1990 for (i = ((validd - 1) | 1); i >= 0; i -= 2) { in pcmpxstrx()
1992 pcmp_val(d, ctrl, i - 1) <= v); in pcmpxstrx()
1997 res = (1 << (upper - MAX(valids, validd))) - 1; in pcmpxstrx()
2000 res <<= 1; in pcmpxstrx()
2006 if (validd == -1) { in pcmpxstrx()
2007 res = (2 << upper) - 1; in pcmpxstrx()
2011 res <<= 1; in pcmpxstrx()
2012 v = 1; in pcmpxstrx()
2022 case 1: in pcmpxstrx()
2023 res ^= (2 << upper) - 1; in pcmpxstrx()
2026 res ^= (1 << (valids + 1)) - 1; in pcmpxstrx()
2033 if (res & 1) { in pcmpxstrx()
2048 env->regs[R_ECX] = (ctrl & (1 << 6)) ? 31 - clz32(res) : ctz32(res); in glue()
2050 env->regs[R_ECX] = 16 >> (ctrl & (1 << 0)); in glue()
2062 if ((ctrl >> 6) & 1) { in glue()
2063 if (ctrl & 1) { in glue()
2064 for (i = 0; i < 8; i++, res >>= 1) { in glue()
2065 env->xmm_regs[0].W(i) = (res & 1) ? ~0 : 0; in glue()
2068 for (i = 0; i < 16; i++, res >>= 1) { in glue()
2069 env->xmm_regs[0].B(i) = (res & 1) ? ~0 : 0; in glue()
2073 env->xmm_regs[0].Q(1) = 0; in glue()
2086 env->regs[R_ECX] = (ctrl & (1 << 6)) ? 31 - clz32(res) : ctz32(res); in glue()
2088 env->regs[R_ECX] = 16 >> (ctrl & (1 << 0)); in glue()
2100 if ((ctrl >> 6) & 1) { in glue()
2101 if (ctrl & 1) { in glue()
2102 for (i = 0; i < 8; i++, res >>= 1) { in glue()
2103 env->xmm_regs[0].W(i) = (res & 1) ? ~0 : 0; in glue()
2106 for (i = 0; i < 16; i++, res >>= 1) { in glue()
2107 env->xmm_regs[0].B(i) = (res & 1) ? ~0 : 0; in glue()
2111 env->xmm_regs[0].Q(1) = 0; in glue()
2120 target_ulong crc = (msg & ((target_ulong) -1 >> in helper_crc32()
2124 crc = (crc >> 1) ^ ((crc & 1) ? CRCPOLY_BITREV : 0); in helper_crc32()
2135 int a_idx = (ctrl & 1) != 0; in glue()
2140 uint64_t b = s->Q(2 * i + b_idx); in glue() local
2143 *r = clmul_64(a, b); in glue()
2191 #if SHIFT == 1
2207 d->B(i) = AES_sbox[tmp.B(i + 4)]; in glue()
2208 d->B(i + 8) = AES_sbox[tmp.B(i + 12)]; in glue()
2210 d->L(1) = (d->L(0) << 24 | d->L(0) >> 8) ^ ctrl; in glue()
2216 #if SHIFT >= 1
2222 for (i = 0; i < 1 << SHIFT; i += 2) { in glue()
2223 r0 = v->Q(i + ((s->Q(i) >> 1) & 1)); in glue()
2224 r1 = v->Q(i + ((s->Q(i+1) >> 1) & 1)); in glue()
2226 d->Q(i+1) = r1; in glue()
2237 r1 = v->L(i + (s->L(i+1) & 3)); in glue()
2241 d->L(i+1) = r1; in glue()
2252 for (i = 0; i < 1 << SHIFT; i += 2) { in glue()
2253 r0 = s->Q(i + ((order >> 0) & 1)); in glue()
2254 r1 = s->Q(i + ((order >> 1) & 1)); in glue()
2256 d->Q(i+1) = r1; in glue()
2273 d->L(i+1) = r1; in glue()
2279 #if SHIFT == 1
2314 for (i = 0; i < 1 << SHIFT; i++) { in glue()
2339 for (i = 0; i < (1 << SHIFT); i++) { in glue()
2359 for (i = 0; i < (1 << SHIFT); i++) { in glue()
2382 for (i = 0; i < (1 << SHIFT); i++) { in glue()
2396 for (i = 0; i < (1 << SHIFT); i++) { in glue()
2404 for (i /= 2; i < 1 << SHIFT; i++) { in glue()
2414 for (i = 0; i < (1 << SHIFT); i++) { in glue()
2433 r1 = v->Q(1); in helper_vpermdq_ymm()
2435 case 1: in helper_vpermdq_ymm()
2441 r1 = s->Q(1); in helper_vpermdq_ymm()
2453 r3 = v->Q(1); in helper_vpermdq_ymm()
2455 case 1: in helper_vpermdq_ymm()
2461 r3 = s->Q(1); in helper_vpermdq_ymm()
2471 d->Q(1) = r1; in helper_vpermdq_ymm()
2476 d->Q(1) = 0; in helper_vpermdq_ymm()
2492 d->Q(1) = r1; in helper_vpermq_ymm()
2512 #if SHIFT == 1
2514 void name(CPUX86State *env, Reg *d, Reg *a, Reg *b, Reg *c, int flags) \
2516 d->elem(0) = F(a->elem(0), b->elem(0), c->elem(0), flags, &env->sse_status); \
2519 void glue(name, SUFFIX)(CPUX86State *env, Reg *d, Reg *a, Reg *b, Reg *c, \
2524 d->elem(i) = F(a->elem(i), b->elem(i), c->elem(i), flags, &env->sse_status); \
2533 #if SHIFT >= 1 in SSE_HELPER_FMAS()
2535 SSE_HELPER_FMAP(helper_fma4pd, ZMM_D, 1 << SHIFT, float64_muladd) in SSE_HELPER_FMAS()
2538 #if SHIFT == 1 in SSE_HELPER_FMAS()
2540 void name(Reg *d, Reg *a, Reg *b) \ in SSE_HELPER_FMAS()
2542 uint32_t A, B, C, D, E, t, i; \ in SSE_HELPER_FMAS()
2545 B = a->L(2); \ in SSE_HELPER_FMAS()
2546 C = a->L(1); \ in SSE_HELPER_FMAS()
2551 t = F(B, C, D) + rol32(A, 5) + b->L(3 - i) + E + K; \ in SSE_HELPER_FMAS()
2554 C = rol32(B, 30); \ in SSE_HELPER_FMAS()
2555 B = A; \ in SSE_HELPER_FMAS()
2560 d->L(2) = B; \ in SSE_HELPER_FMAS()
2561 d->L(1) = C; \ in SSE_HELPER_FMAS()
2565 #define SHA1_F0(b, c, d) (((b) & (c)) ^ (~(b) & (d))) argument
2566 #define SHA1_F1(b, c, d) ((b) ^ (c) ^ (d)) argument
2567 #define SHA1_F2(b, c, d) (((b) & (c)) ^ ((b) & (d)) ^ ((c) & (d))) argument
2574 void helper_sha1nexte(Reg *d, Reg *a, Reg *b)
2576 d->L(3) = b->L(3) + rol32(a->L(3), 30);
2577 d->L(2) = b->L(2);
2578 d->L(1) = b->L(1);
2579 d->L(0) = b->L(0);
2582 void helper_sha1msg1(Reg *d, Reg *a, Reg *b) in helper_sha1msg1() argument
2585 uint32_t b3 = b->L(3); in helper_sha1msg1()
2586 uint32_t b2 = b->L(2); in helper_sha1msg1()
2588 d->L(3) = a->L(3) ^ a->L(1); in helper_sha1msg1()
2590 d->L(1) = a->L(1) ^ b3; in helper_sha1msg1()
2594 void helper_sha1msg2(Reg *d, Reg *a, Reg *b) in helper_sha1msg2() argument
2596 d->L(3) = rol32(a->L(3) ^ b->L(2), 1); in helper_sha1msg2()
2597 d->L(2) = rol32(a->L(2) ^ b->L(1), 1); in helper_sha1msg2()
2598 d->L(1) = rol32(a->L(1) ^ b->L(0), 1); in helper_sha1msg2()
2599 d->L(0) = rol32(a->L(0) ^ d->L(3), 1); in helper_sha1msg2()
2603 #define SHA256_MAJ(a, b, c) (((a) & (b)) ^ ((a) & (c)) ^ ((b) & (c))) argument
2610 void helper_sha256rnds2(Reg *d, Reg *a, Reg *b, uint32_t wk0, uint32_t wk1) in helper_sha256rnds2() argument
2614 uint32_t A = b->L(3); in helper_sha256rnds2()
2615 uint32_t B = b->L(2); in helper_sha256rnds2() local
2618 uint32_t E = b->L(1); in helper_sha256rnds2()
2619 uint32_t F = b->L(0); in helper_sha256rnds2()
2620 uint32_t G = a->L(1); in helper_sha256rnds2()
2625 AA = t + SHA256_MAJ(A, B, C) + SHA256_RNDS0(A); in helper_sha256rnds2()
2628 /* These will be B and F at the end of the odd round */ in helper_sha256rnds2()
2632 D = C, C = B, B = A, A = AA; in helper_sha256rnds2()
2637 AA = t + SHA256_MAJ(A, B, C) + SHA256_RNDS0(A); in helper_sha256rnds2()
2641 d->L(1) = EE; in helper_sha256rnds2()
2644 void helper_sha256msg1(Reg *d, Reg *a, Reg *b) in helper_sha256msg1() argument
2646 /* b->L(0) could be overwritten by the first assignment, save it. */ in helper_sha256msg1()
2647 uint32_t b0 = b->L(0); in helper_sha256msg1()
2649 d->L(0) = a->L(0) + SHA256_MSGS0(a->L(1)); in helper_sha256msg1()
2650 d->L(1) = a->L(1) + SHA256_MSGS0(a->L(2)); in helper_sha256msg1()
2655 void helper_sha256msg2(Reg *d, Reg *a, Reg *b) in helper_sha256msg2() argument
2658 d->L(0) = a->L(0) + SHA256_MSGS1(b->L(2)); in helper_sha256msg2()
2659 d->L(1) = a->L(1) + SHA256_MSGS1(b->L(3)); in helper_sha256msg2()
2662 d->L(3) = a->L(3) + SHA256_MSGS1(d->L(1)); in helper_sha256msg2()
2672 #undef B