__m128i test_mm_cvtsi64_si128(long long A) {
// DAG-LABEL: test_mm_cvtsi64_si128
// DAG: insertelement <2 x i64> undef, i64 %{{.*}}, i32 0
//
// ASM-LABEL: test_mm_cvtsi64_si128
// ASM: movq
return _mm_cvtsi64_si128(A);
}
TARGET_SSE4_1 static inline __m128i extract(__m128i xmm0, __m128i xmm1, __m128i xmm2, __m128i xmm3)
{
qint64 r0, r1;
r0 = Ax[0][static_cast<quint8>(_mm_extract_epi8(xmm0, row + 0))];
r0 ^= Ax[1][static_cast<quint8>(_mm_extract_epi8(xmm0, row + 8))];
r0 ^= Ax[2][static_cast<quint8>(_mm_extract_epi8(xmm1, row + 0))];
r0 ^= Ax[3][static_cast<quint8>(_mm_extract_epi8(xmm1, row + 8))];
r0 ^= Ax[4][static_cast<quint8>(_mm_extract_epi8(xmm2, row + 0))];
r0 ^= Ax[5][static_cast<quint8>(_mm_extract_epi8(xmm2, row + 8))];
r0 ^= Ax[6][static_cast<quint8>(_mm_extract_epi8(xmm3, row + 0))];
r0 ^= Ax[7][static_cast<quint8>(_mm_extract_epi8(xmm3, row + 8))];
r1 = Ax[0][static_cast<quint8>(_mm_extract_epi8(xmm0, row + 1))];
r1 ^= Ax[1][static_cast<quint8>(_mm_extract_epi8(xmm0, row + 9))];
r1 ^= Ax[2][static_cast<quint8>(_mm_extract_epi8(xmm1, row + 1))];
r1 ^= Ax[3][static_cast<quint8>(_mm_extract_epi8(xmm1, row + 9))];
r1 ^= Ax[4][static_cast<quint8>(_mm_extract_epi8(xmm2, row + 1))];
r1 ^= Ax[5][static_cast<quint8>(_mm_extract_epi8(xmm2, row + 9))];
r1 ^= Ax[6][static_cast<quint8>(_mm_extract_epi8(xmm3, row + 1))];
r1 ^= Ax[7][static_cast<quint8>(_mm_extract_epi8(xmm3, row + 9))];
return _mm_insert_epi64(_mm_cvtsi64_si128(r0), r1, 1);
}
__m128i
test (long long b)
{
return _mm_cvtsi64_si128 (b);
}
}bool validate_utf8_sse(const char *src, size_t len) {
const char *end = src + len;
while (src + 16 < end) {
__m128i chunk = _mm_loadu_si128((const __m128i *)(src));
int asciiMask = _mm_movemask_epi8(chunk);
if (!asciiMask) {
src += 16;
continue;
}
__m128i chunk_signed = _mm_add_epi8(chunk, _mm_set1_epi8(0x80));
__m128i cond2 =
_mm_cmplt_epi8(_mm_set1_epi8(0xc2 - 1 - 0x80), chunk_signed);
__m128i state = _mm_set1_epi8((char)(0x0 | 0x80));
state = _mm_blendv_epi8(state, _mm_set1_epi8((char)(0x2 | 0xc0)), cond2);
__m128i cond3 =
_mm_cmplt_epi8(_mm_set1_epi8(0xe0 - 1 - 0x80), chunk_signed);
state = _mm_blendv_epi8(state, _mm_set1_epi8((char)(0x3 | 0xe0)), cond3);
__m128i mask3 = _mm_slli_si128(cond3, 1);
__m128i cond4 =
_mm_cmplt_epi8(_mm_set1_epi8(0xf0 - 1 - 0x80), chunk_signed);
// Fall back to the scalar processing
if (_mm_movemask_epi8(cond4)) {
break;
}
__m128i count = _mm_and_si128(state, _mm_set1_epi8(0x7));
__m128i count_sub1 = _mm_subs_epu8(count, _mm_set1_epi8(0x1));
__m128i counts = _mm_add_epi8(count, _mm_slli_si128(count_sub1, 1));
__m128i shifts = count_sub1;
shifts = _mm_add_epi8(shifts, _mm_slli_si128(shifts, 1));
counts = _mm_add_epi8(
counts, _mm_slli_si128(_mm_subs_epu8(counts, _mm_set1_epi8(0x2)), 2));
shifts = _mm_add_epi8(shifts, _mm_slli_si128(shifts, 2));
if (asciiMask ^ _mm_movemask_epi8(_mm_cmpgt_epi8(counts, _mm_set1_epi8(0))))
return false; // error
shifts = _mm_add_epi8(shifts, _mm_slli_si128(shifts, 4));
if (_mm_movemask_epi8(_mm_cmpgt_epi8(
_mm_sub_epi8(_mm_slli_si128(counts, 1), counts), _mm_set1_epi8(1))))
return false; // error
shifts = _mm_add_epi8(shifts, _mm_slli_si128(shifts, 8));
__m128i mask = _mm_and_si128(state, _mm_set1_epi8(0xf8));
shifts =
_mm_and_si128(shifts, _mm_cmplt_epi8(counts, _mm_set1_epi8(2))); // <=1
chunk =
_mm_andnot_si128(mask, chunk); // from now on, we only have usefull bits
shifts = _mm_blendv_epi8(shifts, _mm_srli_si128(shifts, 1),
_mm_srli_si128(_mm_slli_epi16(shifts, 7), 1));
__m128i chunk_right = _mm_slli_si128(chunk, 1);
__m128i chunk_low = _mm_blendv_epi8(
chunk,
_mm_or_si128(chunk, _mm_and_si128(_mm_slli_epi16(chunk_right, 6),
_mm_set1_epi8(0xc0))),
_mm_cmpeq_epi8(counts, _mm_set1_epi8(1)));
__m128i chunk_high =
_mm_and_si128(chunk, _mm_cmpeq_epi8(counts, _mm_set1_epi8(2)));
shifts = _mm_blendv_epi8(shifts, _mm_srli_si128(shifts, 2),
_mm_srli_si128(_mm_slli_epi16(shifts, 6), 2));
chunk_high = _mm_srli_epi32(chunk_high, 2);
shifts = _mm_blendv_epi8(shifts, _mm_srli_si128(shifts, 4),
_mm_srli_si128(_mm_slli_epi16(shifts, 5), 4));
chunk_high = _mm_or_si128(
chunk_high, _mm_and_si128(_mm_and_si128(_mm_slli_epi32(chunk_right, 4),
_mm_set1_epi8(0xf0)),
mask3));
int c = _mm_extract_epi16(counts, 7);
int source_advance = !(c & 0x0200) ? 16 : !(c & 0x02) ? 15 : 14;
__m128i high_bits = _mm_and_si128(chunk_high, _mm_set1_epi8(0xf8));
if (!_mm_testz_si128(
mask3,
_mm_or_si128(_mm_cmpeq_epi8(high_bits, _mm_set1_epi8(0x00)),
_mm_cmpeq_epi8(high_bits, _mm_set1_epi8(0xd8)))))
return false;
shifts = _mm_blendv_epi8(shifts, _mm_srli_si128(shifts, 8),
_mm_srli_si128(_mm_slli_epi16(shifts, 4), 8));
chunk_high = _mm_slli_si128(chunk_high, 1);
__m128i shuf =
_mm_add_epi8(shifts, _mm_set_epi8(15, 14, 13, 12, 11, 10, 9, 8, 7, 6, 5,
4, 3, 2, 1, 0));
chunk_low = _mm_shuffle_epi8(chunk_low, shuf);
chunk_high = _mm_shuffle_epi8(chunk_high, shuf);
__m128i utf16_low = _mm_unpacklo_epi8(chunk_low, chunk_high);
__m128i utf16_high = _mm_unpackhi_epi8(chunk_low, chunk_high);
if (_mm_cmpestrc(_mm_cvtsi64_si128(0xfdeffdd0fffffffe), 4, utf16_high, 8,
_SIDD_UWORD_OPS | _SIDD_CMP_RANGES) |
_mm_cmpestrc(_mm_cvtsi64_si128(0xfdeffdd0fffffffe), 4, utf16_low, 8,
_SIDD_UWORD_OPS | _SIDD_CMP_RANGES)) {
return false;
}
src += source_advance;
}
return validate_utf8(src, end - src);
}
