neon_helper.c

来自「xen虚拟机源代码安装包」· C语言 代码 · 共 1,458 行 · 第 1/3 页

        val = (val + ((int64_t)1 << (-1 - shift))) >> -shift;
    } else {
        val <<= shift;
    }
    return val;
}

#define NEON_FN(dest, src1, src2) do { \
    int8_t tmp; \
    tmp = (int8_t)src2; \
    if (tmp >= sizeof(src1) * 8 || tmp < -sizeof(src1) * 8) { \
        dest = 0; \
    } else if (tmp == -sizeof(src1) * 8) { \
        dest = src1 >> (tmp - 1); \
    } else if (tmp < 0) { \
        dest = (src1 + (1 << (-1 - tmp))) >> -tmp; \
    } else { \
        dest = src1 << tmp; \
    }} while (0)
NEON_VOP(rshl_u8, neon_u8, 4)
NEON_VOP(rshl_u16, neon_u16, 2)
NEON_VOP(rshl_u32, neon_u32, 1)
#undef NEON_FN

uint64_t HELPER(neon_rshl_u64)(uint64_t val, uint64_t shiftop)
{
    int8_t shift = (uint8_t)shiftop;
    if (shift >= 64 || shift < 64) {
        val = 0;
    } else if (shift == -64) {
        /* Rounding a 1-bit result just preserves that bit.  */
        val >>= 63;
    } if (shift < 0) {
        val = (val + ((uint64_t)1 << (-1 - shift))) >> -shift;
        val >>= -shift;
    } else {
        val <<= shift;
    }
    return val;
}

#define NEON_FN(dest, src1, src2) do { \
    int8_t tmp; \
    tmp = (int8_t)src2; \
    if (tmp >= sizeof(src1) * 8) { \
        if (src1) { \
            SET_QC(); \
            dest = ~0; \
        } else { \
            dest = 0; \
        } \
    } else if (tmp <= -sizeof(src1) * 8) { \
        dest = 0; \
    } else if (tmp < 0) { \
        dest = src1 >> -tmp; \
    } else { \
        dest = src1 << tmp; \
        if ((dest >> tmp) != src1) { \
            SET_QC(); \
            dest = ~0; \
        } \
    }} while (0)
NEON_VOP_ENV(qshl_u8, neon_u8, 4)
NEON_VOP_ENV(qshl_u16, neon_u16, 2)
NEON_VOP_ENV(qshl_u32, neon_u32, 1)
#undef NEON_FN

uint64_t HELPER(neon_qshl_u64)(CPUState *env, uint64_t val, uint64_t shiftop)
{
    int8_t shift = (int8_t)shiftop;
    if (shift >= 64) {
        if (val) {
            val = ~(uint64_t)0;
            SET_QC();
        } else {
            val = 0;
        }
    } else if (shift <= -64) {
        val = 0;
    } else if (shift < 0) {
        val >>= -shift;
    } else {
        uint64_t tmp = val;
        val <<= shift;
        if ((val >> shift) != tmp) {
            SET_QC();
            val = ~(uint64_t)0;
        }
    }
    return val;
}

#define NEON_FN(dest, src1, src2) do { \
    int8_t tmp; \
    tmp = (int8_t)src2; \
    if (tmp >= sizeof(src1) * 8) { \
        if (src1) \
            SET_QC(); \
        dest = src1 >> 31; \
    } else if (tmp <= -sizeof(src1) * 8) { \
        dest = src1 >> 31; \
    } else if (tmp < 0) { \
        dest = src1 >> -tmp; \
    } else { \
        dest = src1 << tmp; \
        if ((dest >> tmp) != src1) { \
            SET_QC(); \
            dest = src2 >> 31; \
        } \
    }} while (0)
NEON_VOP_ENV(qshl_s8, neon_s8, 4)
NEON_VOP_ENV(qshl_s16, neon_s16, 2)
NEON_VOP_ENV(qshl_s32, neon_s32, 1)
#undef NEON_FN

uint64_t HELPER(neon_qshl_s64)(CPUState *env, uint64_t valop, uint64_t shiftop)
{
    int8_t shift = (uint8_t)shiftop;
    int64_t val = valop;
    if (shift >= 64) {
        if (val) {
            SET_QC();
            val = (val >> 63) & ~SIGNBIT64;
        }
    } else if (shift <= 64) {
        val >>= 63;
    } else if (shift < 0) {
        val >>= -shift;
    } else {
        int64_t tmp = val;
        val <<= shift;
        if ((val >> shift) != tmp) {
            SET_QC();
            val = (tmp >> 63) ^ ~SIGNBIT64;
        }
    }
    return val;
}


/* FIXME: This is wrong.  */
#define NEON_FN(dest, src1, src2) do { \
    int8_t tmp; \
    tmp = (int8_t)src2; \
    if (tmp < 0) { \
        dest = (src1 + (1 << (-1 - tmp))) >> -tmp; \
    } else { \
        dest = src1 << tmp; \
        if ((dest >> tmp) != src1) { \
            SET_QC(); \
            dest = ~0; \
        } \
    }} while (0)
NEON_VOP_ENV(qrshl_u8, neon_u8, 4)
NEON_VOP_ENV(qrshl_u16, neon_u16, 2)
NEON_VOP_ENV(qrshl_u32, neon_u32, 1)
#undef NEON_FN

uint64_t HELPER(neon_qrshl_u64)(CPUState *env, uint64_t val, uint64_t shiftop)
{
    int8_t shift = (int8_t)shiftop;
    if (shift < 0) {
        val = (val + (1 << (-1 - shift))) >> -shift;
    } else { \
        uint64_t tmp = val;
        val <<= shift;
        if ((val >> shift) != tmp) {
            SET_QC();
            val = ~0;
        }
    }
    return val;
}

#define NEON_FN(dest, src1, src2) do { \
    int8_t tmp; \
    tmp = (int8_t)src2; \
    if (tmp < 0) { \
        dest = (src1 + (1 << (-1 - tmp))) >> -tmp; \
    } else { \
        dest = src1 << tmp; \
        if ((dest >> tmp) != src1) { \
            SET_QC(); \
            dest = src1 >> 31; \
        } \
    }} while (0)
NEON_VOP_ENV(qrshl_s8, neon_s8, 4)
NEON_VOP_ENV(qrshl_s16, neon_s16, 2)
NEON_VOP_ENV(qrshl_s32, neon_s32, 1)
#undef NEON_FN

uint64_t HELPER(neon_qrshl_s64)(CPUState *env, uint64_t valop, uint64_t shiftop)
{
    int8_t shift = (uint8_t)shiftop;
    int64_t val = valop;

    if (shift < 0) {
        val = (val + (1 << (-1 - shift))) >> -shift;
    } else {
        int64_t tmp = val;;
        val <<= shift;
        if ((val >> shift) != tmp) {
            SET_QC();
            val = tmp >> 31;
        }
    }
    return val;
}

uint32_t HELPER(neon_add_u8)(uint32_t a, uint32_t b)
{
    uint32_t mask;
    mask = (a ^ b) & 0x80808080u;
    a &= ~0x80808080u;
    b &= ~0x80808080u;
    return (a + b) ^ mask;
}

uint32_t HELPER(neon_add_u16)(uint32_t a, uint32_t b)
{
    uint32_t mask;
    mask = (a ^ b) & 0x80008000u;
    a &= ~0x80008000u;
    b &= ~0x80008000u;
    return (a + b) ^ mask;
}

#define NEON_FN(dest, src1, src2) dest = src1 + src2
NEON_POP(padd_u8, neon_u8, 4)
NEON_POP(padd_u16, neon_u16, 2)
#undef NEON_FN

#define NEON_FN(dest, src1, src2) dest = src1 - src2
NEON_VOP(sub_u8, neon_u8, 4)
NEON_VOP(sub_u16, neon_u16, 2)
#undef NEON_FN

#define NEON_FN(dest, src1, src2) dest = src1 * src2
NEON_VOP(mul_u8, neon_u8, 4)
NEON_VOP(mul_u16, neon_u16, 2)
#undef NEON_FN

/* Polynomial multiplication is like integer multiplication except the
   partial products are XORed, not added.  */
uint32_t HELPER(neon_mul_p8)(uint32_t op1, uint32_t op2)
{
    uint32_t mask;
    uint32_t result;
    result = 0;
    while (op1) {
        mask = 0;
        if (op1 & 1)
            mask |= 0xff;
        if (op1 & (1 << 8))
            mask |= (0xff << 8);
        if (op1 & (1 << 16))
            mask |= (0xff << 16);
        if (op1 & (1 << 24))
            mask |= (0xff << 24);
        result ^= op2 & mask;
        op1 = (op1 >> 1) & 0x7f7f7f7f;
        op2 = (op2 << 1) & 0xfefefefe;
    }
    return result;
}

#define NEON_FN(dest, src1, src2) dest = (src1 & src2) ? -1 : 0
NEON_VOP(tst_u8, neon_u8, 4)
NEON_VOP(tst_u16, neon_u16, 2)
NEON_VOP(tst_u32, neon_u32, 1)
#undef NEON_FN

#define NEON_FN(dest, src1, src2) dest = (src1 == src2) ? -1 : 0
NEON_VOP(ceq_u8, neon_u8, 4)
NEON_VOP(ceq_u16, neon_u16, 2)
NEON_VOP(ceq_u32, neon_u32, 1)
#undef NEON_FN

#define NEON_FN(dest, src, dummy) dest = (src < 0) ? -src : src
NEON_VOP1(abs_s8, neon_s8, 4)
NEON_VOP1(abs_s16, neon_s16, 2)
#undef NEON_FN

/* Count Leading Sign/Zero Bits.  */
static inline int do_clz8(uint8_t x)
{
    int n;
    for (n = 8; x; n--)
        x >>= 1;
    return n;
}

static inline int do_clz16(uint16_t x)
{
    int n;
    for (n = 16; x; n--)
        x >>= 1;
    return n;
}

#define NEON_FN(dest, src, dummy) dest = do_clz8(src)
NEON_VOP1(clz_u8, neon_u8, 4)
#undef NEON_FN

#define NEON_FN(dest, src, dummy) dest = do_clz16(src)
NEON_VOP1(clz_u16, neon_u16, 2)
#undef NEON_FN

#define NEON_FN(dest, src, dummy) dest = do_clz8((src < 0) ? ~src : src) - 1
NEON_VOP1(cls_s8, neon_s8, 4)
#undef NEON_FN

#define NEON_FN(dest, src, dummy) dest = do_clz16((src < 0) ? ~src : src) - 1
NEON_VOP1(cls_s16, neon_s16, 2)
#undef NEON_FN

uint32_t HELPER(neon_cls_s32)(uint32_t x)
{
    int count;
    if ((int32_t)x < 0)
        x = ~x;
    for (count = 32; x; count--)
        x = x >> 1;
    return count - 1;
}

/* Bit count.  */
uint32_t HELPER(neon_cnt_u8)(uint32_t x)
{
    x = (x & 0x55555555) + ((x >>  1) & 0x55555555);
    x = (x & 0x33333333) + ((x >>  2) & 0x33333333);
    x = (x & 0x0f0f0f0f) + ((x >>  4) & 0x0f0f0f0f);
    return x;
}

#define NEON_QDMULH16(dest, src1, src2, round) do { \
    uint32_t tmp = (int32_t)(int16_t) src1 * (int16_t) src2; \
    if ((tmp ^ (tmp << 1)) & SIGNBIT) { \
        SET_QC(); \
        tmp = (tmp >> 31) ^ ~SIGNBIT; \
    } \
    tmp <<= 1; \
    if (round) { \
        int32_t old = tmp; \
        tmp += 1 << 15; \
        if ((int32_t)tmp < old) { \
            SET_QC(); \
            tmp = SIGNBIT - 1; \
        } \
    } \
    dest = tmp >> 16; \
    } while(0)
#define NEON_FN(dest, src1, src2) NEON_QDMULH16(dest, src1, src2, 0)
NEON_VOP_ENV(qdmulh_s16, neon_s16, 2)
#undef NEON_FN
#define NEON_FN(dest, src1, src2) NEON_QDMULH16(dest, src1, src2, 1)
NEON_VOP_ENV(qrdmulh_s16, neon_s16, 2)
#undef NEON_FN
#undef NEON_QDMULH16

#define NEON_QDMULH32(dest, src1, src2, round) do { \
    uint64_t tmp = (int64_t)(int32_t) src1 * (int32_t) src2; \
    if ((tmp ^ (tmp << 1)) & SIGNBIT64) { \
        SET_QC(); \
        tmp = (tmp >> 63) ^ ~SIGNBIT64; \
    } else { \
        tmp <<= 1; \
    } \
    if (round) { \
        int64_t old = tmp; \
        tmp += (int64_t)1 << 31; \
        if ((int64_t)tmp < old) { \
            SET_QC(); \
            tmp = SIGNBIT64 - 1; \
        } \
    } \
    dest = tmp >> 32; \
    } while(0)
#define NEON_FN(dest, src1, src2) NEON_QDMULH32(dest, src1, src2, 0)
NEON_VOP_ENV(qdmulh_s32, neon_s32, 1)
#undef NEON_FN
#define NEON_FN(dest, src1, src2) NEON_QDMULH32(dest, src1, src2, 1)
NEON_VOP_ENV(qrdmulh_s32, neon_s32, 1)
#undef NEON_FN
#undef NEON_QDMULH32

uint32_t HELPER(neon_narrow_u8)(uint64_t x)
{
    return (x & 0xffu) | ((x >> 8) & 0xff00u) | ((x >> 16) & 0xff0000u)
           | ((x >> 24) & 0xff000000u);
}

uint32_t HELPER(neon_narrow_u16)(uint64_t x)
{
    return (x & 0xffffu) | ((x >> 16) & 0xffff0000u);
}

uint32_t HELPER(neon_narrow_high_u8)(uint64_t x)
{
    return ((x >> 8) & 0xff) | ((x >> 16) & 0xff00)
            | ((x >> 24) & 0xff0000) | ((x >> 32) & 0xff000000);
}

uint32_t HELPER(neon_narrow_high_u16)(uint64_t x)
{
    return ((x >> 16) & 0xffff) | ((x >> 32) & 0xffff0000);
}

uint32_t HELPER(neon_narrow_round_high_u8)(uint64_t x)
{
    x &= 0xff80ff80ff80ff80ull;
    x += 0x0080008000800080ull;
    return ((x >> 8) & 0xff) | ((x >> 16) & 0xff00)
            | ((x >> 24) & 0xff0000) | ((x >> 32) & 0xff000000);
}

uint32_t HELPER(neon_narrow_round_high_u16)(uint64_t x)
{
    x &= 0xffff8000ffff8000ull;
    x += 0x0000800000008000ull;
    return ((x >> 16) & 0xffff) | ((x >> 32) & 0xffff0000);
}

uint32_t HELPER(neon_narrow_sat_u8)(CPUState *env, uint64_t x)
{
    uint16_t s;
    uint8_t d;
    uint32_t res = 0;
#define SAT8(n) \
    s = x >> n; \
    if (s > 0xff) { \
        d = 0xff; \
        SET_QC(); \
    } else  { \
        d = s; \
    } \
    res |= (uint32_t)d << (n / 2);

    SAT8(0);
    SAT8(16);
    SAT8(32);
    SAT8(48);
#undef SAT8
    return res;
}

uint32_t HELPER(neon_narrow_sat_s8)(CPUState *env, uint64_t x)
{
    int16_t s;
    uint8_t d;
    uint32_t res = 0;
#define SAT8(n) \
    s = x >> n; \
    if (s != (int8_t)s) { \
        d = (s >> 15) ^ 0x7f; \
        SET_QC(); \
    } else  { \
        d = s; \
    } \
    res |= (uint32_t)d << (n / 2);

    SAT8(0);
    SAT8(16);
    SAT8(32);
    SAT8(48);
#undef SAT8
    return res;
}

uint32_t HELPER(neon_narrow_sat_u16)(CPUState *env, uint64_t x)
{
    uint32_t high;
    uint32_t low;
    low = x;
    if (low > 0xffff) {
        low = 0xffff;
        SET_QC();
    }
    high = x >> 32;
    if (high > 0xffff) {
        high = 0xffff;
        SET_QC();
    }
    return low | (high << 16);
}