ops_sse.h

xen 3.2.2 源码

    Reg *d, *s;
    d = (Reg *)((char *)env + PARAM1);
    s = (Reg *)((char *)env + PARAM2);

    d->Q(0) = (uint64_t)s->L(0) * (uint64_t)d->L(0);
#if SHIFT == 1
    d->Q(1) = (uint64_t)s->L(2) * (uint64_t)d->L(2);
#endif
}

void OPPROTO glue(op_pmaddwd, SUFFIX) (void)
{
    int i;
    Reg *d, *s;
    d = (Reg *)((char *)env + PARAM1);
    s = (Reg *)((char *)env + PARAM2);

    for(i = 0; i < (2 << SHIFT); i++) {
        d->L(i) = (int16_t)s->W(2*i) * (int16_t)d->W(2*i) +
            (int16_t)s->W(2*i+1) * (int16_t)d->W(2*i+1);
    }
    FORCE_RET();
}

#if SHIFT == 0
static inline int abs1(int a)
{
    if (a < 0)
        return -a;
    else
        return a;
}
#endif
void OPPROTO glue(op_psadbw, SUFFIX) (void)
{
    unsigned int val;
    Reg *d, *s;
    d = (Reg *)((char *)env + PARAM1);
    s = (Reg *)((char *)env + PARAM2);

    val = 0;
    val += abs1(d->B(0) - s->B(0));
    val += abs1(d->B(1) - s->B(1));
    val += abs1(d->B(2) - s->B(2));
    val += abs1(d->B(3) - s->B(3));
    val += abs1(d->B(4) - s->B(4));
    val += abs1(d->B(5) - s->B(5));
    val += abs1(d->B(6) - s->B(6));
    val += abs1(d->B(7) - s->B(7));
    d->Q(0) = val;
#if SHIFT == 1
    val = 0;
    val += abs1(d->B(8) - s->B(8));
    val += abs1(d->B(9) - s->B(9));
    val += abs1(d->B(10) - s->B(10));
    val += abs1(d->B(11) - s->B(11));
    val += abs1(d->B(12) - s->B(12));
    val += abs1(d->B(13) - s->B(13));
    val += abs1(d->B(14) - s->B(14));
    val += abs1(d->B(15) - s->B(15));
    d->Q(1) = val;
#endif
}

void OPPROTO glue(op_maskmov, SUFFIX) (void)
{
    int i;
    Reg *d, *s;
    d = (Reg *)((char *)env + PARAM1);
    s = (Reg *)((char *)env + PARAM2);
    for(i = 0; i < (8 << SHIFT); i++) {
        if (s->B(i) & 0x80)
            stb(A0 + i, d->B(i));
    }
    FORCE_RET();
}

void OPPROTO glue(op_movl_mm_T0, SUFFIX) (void)
{
    Reg *d;
    d = (Reg *)((char *)env + PARAM1);
    d->L(0) = T0;
    d->L(1) = 0;
#if SHIFT == 1
    d->Q(1) = 0;
#endif
}

void OPPROTO glue(op_movl_T0_mm, SUFFIX) (void)
{
    Reg *s;
    s = (Reg *)((char *)env + PARAM1);
    T0 = s->L(0);
}

#ifdef TARGET_X86_64
void OPPROTO glue(op_movq_mm_T0, SUFFIX) (void)
{
    Reg *d;
    d = (Reg *)((char *)env + PARAM1);
    d->Q(0) = T0;
#if SHIFT == 1
    d->Q(1) = 0;
#endif
}

void OPPROTO glue(op_movq_T0_mm, SUFFIX) (void)
{
    Reg *s;
    s = (Reg *)((char *)env + PARAM1);
    T0 = s->Q(0);
}
#endif

#if SHIFT == 0
void OPPROTO glue(op_pshufw, SUFFIX) (void)
{
    Reg r, *d, *s;
    int order;
    d = (Reg *)((char *)env + PARAM1);
    s = (Reg *)((char *)env + PARAM2);
    order = PARAM3;
    r.W(0) = s->W(order & 3);
    r.W(1) = s->W((order >> 2) & 3);
    r.W(2) = s->W((order >> 4) & 3);
    r.W(3) = s->W((order >> 6) & 3);
    *d = r;
}
#else
void OPPROTO op_shufps(void)
{
    Reg r, *d, *s;
    int order;
    d = (Reg *)((char *)env + PARAM1);
    s = (Reg *)((char *)env + PARAM2);
    order = PARAM3;
    r.L(0) = d->L(order & 3);
    r.L(1) = d->L((order >> 2) & 3);
    r.L(2) = s->L((order >> 4) & 3);
    r.L(3) = s->L((order >> 6) & 3);
    *d = r;
}

void OPPROTO op_shufpd(void)
{
    Reg r, *d, *s;
    int order;
    d = (Reg *)((char *)env + PARAM1);
    s = (Reg *)((char *)env + PARAM2);
    order = PARAM3;
    r.Q(0) = d->Q(order & 1);
    r.Q(1) = s->Q((order >> 1) & 1);
    *d = r;
}

void OPPROTO glue(op_pshufd, SUFFIX) (void)
{
    Reg r, *d, *s;
    int order;
    d = (Reg *)((char *)env + PARAM1);
    s = (Reg *)((char *)env + PARAM2);
    order = PARAM3;
    r.L(0) = s->L(order & 3);
    r.L(1) = s->L((order >> 2) & 3);
    r.L(2) = s->L((order >> 4) & 3);
    r.L(3) = s->L((order >> 6) & 3);
    *d = r;
}

void OPPROTO glue(op_pshuflw, SUFFIX) (void)
{
    Reg r, *d, *s;
    int order;
    d = (Reg *)((char *)env + PARAM1);
    s = (Reg *)((char *)env + PARAM2);
    order = PARAM3;
    r.W(0) = s->W(order & 3);
    r.W(1) = s->W((order >> 2) & 3);
    r.W(2) = s->W((order >> 4) & 3);
    r.W(3) = s->W((order >> 6) & 3);
    r.Q(1) = s->Q(1);
    *d = r;
}

void OPPROTO glue(op_pshufhw, SUFFIX) (void)
{
    Reg r, *d, *s;
    int order;
    d = (Reg *)((char *)env + PARAM1);
    s = (Reg *)((char *)env + PARAM2);
    order = PARAM3;
    r.Q(0) = s->Q(0);
    r.W(4) = s->W(4 + (order & 3));
    r.W(5) = s->W(4 + ((order >> 2) & 3));
    r.W(6) = s->W(4 + ((order >> 4) & 3));
    r.W(7) = s->W(4 + ((order >> 6) & 3));
    *d = r;
}
#endif

#if SHIFT == 1
/* FPU ops */
/* XXX: not accurate */

#define SSE_OP_S(name, F)\
void OPPROTO op_ ## name ## ps (void)\
{\
    Reg *d, *s;\
    d = (Reg *)((char *)env + PARAM1);\
    s = (Reg *)((char *)env + PARAM2);\
    d->XMM_S(0) = F(32, d->XMM_S(0), s->XMM_S(0));\
    d->XMM_S(1) = F(32, d->XMM_S(1), s->XMM_S(1));\
    d->XMM_S(2) = F(32, d->XMM_S(2), s->XMM_S(2));\
    d->XMM_S(3) = F(32, d->XMM_S(3), s->XMM_S(3));\
}\
\
void OPPROTO op_ ## name ## ss (void)\
{\
    Reg *d, *s;\
    d = (Reg *)((char *)env + PARAM1);\
    s = (Reg *)((char *)env + PARAM2);\
    d->XMM_S(0) = F(32, d->XMM_S(0), s->XMM_S(0));\
}\
void OPPROTO op_ ## name ## pd (void)\
{\
    Reg *d, *s;\
    d = (Reg *)((char *)env + PARAM1);\
    s = (Reg *)((char *)env + PARAM2);\
    d->XMM_D(0) = F(64, d->XMM_D(0), s->XMM_D(0));\
    d->XMM_D(1) = F(64, d->XMM_D(1), s->XMM_D(1));\
}\
\
void OPPROTO op_ ## name ## sd (void)\
{\
    Reg *d, *s;\
    d = (Reg *)((char *)env + PARAM1);\
    s = (Reg *)((char *)env + PARAM2);\
    d->XMM_D(0) = F(64, d->XMM_D(0), s->XMM_D(0));\
}

#define FPU_ADD(size, a, b) float ## size ## _add(a, b, &env->sse_status)
#define FPU_SUB(size, a, b) float ## size ## _sub(a, b, &env->sse_status)
#define FPU_MUL(size, a, b) float ## size ## _mul(a, b, &env->sse_status)
#define FPU_DIV(size, a, b) float ## size ## _div(a, b, &env->sse_status)
#define FPU_MIN(size, a, b) (a) < (b) ? (a) : (b)
#define FPU_MAX(size, a, b) (a) > (b) ? (a) : (b)
#define FPU_SQRT(size, a, b) float ## size ## _sqrt(b, &env->sse_status)

SSE_OP_S(add, FPU_ADD)
SSE_OP_S(sub, FPU_SUB)
SSE_OP_S(mul, FPU_MUL)
SSE_OP_S(div, FPU_DIV)
SSE_OP_S(min, FPU_MIN)
SSE_OP_S(max, FPU_MAX)
SSE_OP_S(sqrt, FPU_SQRT)


/* float to float conversions */
void OPPROTO op_cvtps2pd(void)
{
    float32 s0, s1;
    Reg *d, *s;
    d = (Reg *)((char *)env + PARAM1);
    s = (Reg *)((char *)env + PARAM2);
    s0 = s->XMM_S(0);
    s1 = s->XMM_S(1);
    d->XMM_D(0) = float32_to_float64(s0, &env->sse_status);
    d->XMM_D(1) = float32_to_float64(s1, &env->sse_status);
}

void OPPROTO op_cvtpd2ps(void)
{
    Reg *d, *s;
    d = (Reg *)((char *)env + PARAM1);
    s = (Reg *)((char *)env + PARAM2);
    d->XMM_S(0) = float64_to_float32(s->XMM_D(0), &env->sse_status);
    d->XMM_S(1) = float64_to_float32(s->XMM_D(1), &env->sse_status);
    d->Q(1) = 0;
}

void OPPROTO op_cvtss2sd(void)
{
    Reg *d, *s;
    d = (Reg *)((char *)env + PARAM1);
    s = (Reg *)((char *)env + PARAM2);
    d->XMM_D(0) = float32_to_float64(s->XMM_S(0), &env->sse_status);
}

void OPPROTO op_cvtsd2ss(void)
{
    Reg *d, *s;
    d = (Reg *)((char *)env + PARAM1);
    s = (Reg *)((char *)env + PARAM2);
    d->XMM_S(0) = float64_to_float32(s->XMM_D(0), &env->sse_status);
}

/* integer to float */
void OPPROTO op_cvtdq2ps(void)
{
    XMMReg *d = (XMMReg *)((char *)env + PARAM1);
    XMMReg *s = (XMMReg *)((char *)env + PARAM2);
    d->XMM_S(0) = int32_to_float32(s->XMM_L(0), &env->sse_status);
    d->XMM_S(1) = int32_to_float32(s->XMM_L(1), &env->sse_status);
    d->XMM_S(2) = int32_to_float32(s->XMM_L(2), &env->sse_status);
    d->XMM_S(3) = int32_to_float32(s->XMM_L(3), &env->sse_status);
}

void OPPROTO op_cvtdq2pd(void)
{
    XMMReg *d = (XMMReg *)((char *)env + PARAM1);
    XMMReg *s = (XMMReg *)((char *)env + PARAM2);
    int32_t l0, l1;
    l0 = (int32_t)s->XMM_L(0);
    l1 = (int32_t)s->XMM_L(1);
    d->XMM_D(0) = int32_to_float64(l0, &env->sse_status);
    d->XMM_D(1) = int32_to_float64(l1, &env->sse_status);
}

void OPPROTO op_cvtpi2ps(void)
{
    XMMReg *d = (Reg *)((char *)env + PARAM1);
    MMXReg *s = (MMXReg *)((char *)env + PARAM2);
    d->XMM_S(0) = int32_to_float32(s->MMX_L(0), &env->sse_status);
    d->XMM_S(1) = int32_to_float32(s->MMX_L(1), &env->sse_status);
}

void OPPROTO op_cvtpi2pd(void)
{
    XMMReg *d = (Reg *)((char *)env + PARAM1);
    MMXReg *s = (MMXReg *)((char *)env + PARAM2);
    d->XMM_D(0) = int32_to_float64(s->MMX_L(0), &env->sse_status);
    d->XMM_D(1) = int32_to_float64(s->MMX_L(1), &env->sse_status);
}

void OPPROTO op_cvtsi2ss(void)
{
    XMMReg *d = (Reg *)((char *)env + PARAM1);
    d->XMM_S(0) = int32_to_float32(T0, &env->sse_status);
}

void OPPROTO op_cvtsi2sd(void)
{
    XMMReg *d = (Reg *)((char *)env + PARAM1);
    d->XMM_D(0) = int32_to_float64(T0, &env->sse_status);
}

#ifdef TARGET_X86_64
void OPPROTO op_cvtsq2ss(void)
{
    XMMReg *d = (Reg *)((char *)env + PARAM1);
    d->XMM_S(0) = int64_to_float32(T0, &env->sse_status);
}

void OPPROTO op_cvtsq2sd(void)
{
    XMMReg *d = (Reg *)((char *)env + PARAM1);
    d->XMM_D(0) = int64_to_float64(T0, &env->sse_status);
}
#endif

/* float to integer */
void OPPROTO op_cvtps2dq(void)
{
    XMMReg *d = (XMMReg *)((char *)env + PARAM1);
    XMMReg *s = (XMMReg *)((char *)env + PARAM2);
    d->XMM_L(0) = float32_to_int32(s->XMM_S(0), &env->sse_status);
    d->XMM_L(1) = float32_to_int32(s->XMM_S(1), &env->sse_status);
    d->XMM_L(2) = float32_to_int32(s->XMM_S(2), &env->sse_status);
    d->XMM_L(3) = float32_to_int32(s->XMM_S(3), &env->sse_status);
}

void OPPROTO op_cvtpd2dq(void)
{
    XMMReg *d = (XMMReg *)((char *)env + PARAM1);
    XMMReg *s = (XMMReg *)((char *)env + PARAM2);
    d->XMM_L(0) = float64_to_int32(s->XMM_D(0), &env->sse_status);
    d->XMM_L(1) = float64_to_int32(s->XMM_D(1), &env->sse_status);
    d->XMM_Q(1) = 0;
}

void OPPROTO op_cvtps2pi(void)
{
    MMXReg *d = (MMXReg *)((char *)env + PARAM1);
    XMMReg *s = (XMMReg *)((char *)env + PARAM2);
    d->MMX_L(0) = float32_to_int32(s->XMM_S(0), &env->sse_status);
    d->MMX_L(1) = float32_to_int32(s->XMM_S(1), &env->sse_status);
}

void OPPROTO op_cvtpd2pi(void)
{
    MMXReg *d = (MMXReg *)((char *)env + PARAM1);
    XMMReg *s = (XMMReg *)((char *)env + PARAM2);
    d->MMX_L(0) = float64_to_int32(s->XMM_D(0), &env->sse_status);
    d->MMX_L(1) = float64_to_int32(s->XMM_D(1), &env->sse_status);
}

void OPPROTO op_cvtss2si(void)
{
    XMMReg *s = (XMMReg *)((char *)env + PARAM1);
    T0 = float32_to_int32(s->XMM_S(0), &env->sse_status);
}

void OPPROTO op_cvtsd2si(void)
{
    XMMReg *s = (XMMReg *)((char *)env + PARAM1);
    T0 = float64_to_int32(s->XMM_D(0), &env->sse_status);
}

#ifdef TARGET_X86_64
void OPPROTO op_cvtss2sq(void)
{
    XMMReg *s = (XMMReg *)((char *)env + PARAM1);
    T0 = float32_to_int64(s->XMM_S(0), &env->sse_status);
}

void OPPROTO op_cvtsd2sq(void)
{
    XMMReg *s = (XMMReg *)((char *)env + PARAM1);
    T0 = float64_to_int64(s->XMM_D(0), &env->sse_status);
}
#endif

/* float to integer truncated */
void OPPROTO op_cvttps2dq(void)
{
    XMMReg *d = (XMMReg *)((char *)env + PARAM1);
    XMMReg *s = (XMMReg *)((char *)env + PARAM2);
    d->XMM_L(0) = float32_to_int32_round_to_zero(s->XMM_S(0), &env->sse_status);
    d->XMM_L(1) = float32_to_int32_round_to_zero(s->XMM_S(1), &env->sse_status);
    d->XMM_L(2) = float32_to_int32_round_to_zero(s->XMM_S(2), &env->sse_status);
    d->XMM_L(3) = float32_to_int32_round_to_zero(s->XMM_S(3), &env->sse_status);
}

void OPPROTO op_cvttpd2dq(void)
{
    XMMReg *d = (XMMReg *)((char *)env + PARAM1);
    XMMReg *s = (XMMReg *)((char *)env + PARAM2);
    d->XMM_L(0) = float64_to_int32_round_to_zero(s->XMM_D(0), &env->sse_status);
    d->XMM_L(1) = float64_to_int32_round_to_zero(s->XMM_D(1), &env->sse_status);
    d->XMM_Q(1) = 0;
}

void OPPROTO op_cvttps2pi(void)
{
    MMXReg *d = (MMXReg *)((char *)env + PARAM1);
    XMMReg *s = (XMMReg *)((char *)env + PARAM2);
    d->MMX_L(0) = float32_to_int32_round_to_zero(s->XMM_S(0), &env->sse_status);
    d->MMX_L(1) = float32_to_int32_round_to_zero(s->XMM_S(1), &env->sse_status);
}

void OPPROTO op_cvttpd2pi(void)
{
    MMXReg *d = (MMXReg *)((char *)env + PARAM1);
    XMMReg *s = (XMMReg *)((char *)env + PARAM2);
    d->MMX_L(0) = float64_to_int32_round_to_zero(s->XMM_D(0), &env->sse_status);
    d->MMX_L(1) = float64_to_int32_round_to_zero(s->XMM_D(1), &env->sse_status);
}

void OPPROTO op_cvttss2si(void)
{
    XMMReg *s = (XMMReg *)((char *)env + PARAM1);
    T0 = float32_to_int32_round_to_zero(s->XMM_S(0), &env->sse_status);
}

void OPPROTO op_cvttsd2si(void)