像素操作库pixman

/* * Copyright © 2008 Rodrigo Kumpera * Copyright © 2008 André Tupinambá * * Permission to use, copy, modify, distribute, and sell this software and its * documentation for any purpose is hereby granted without fee, provided that * the above copyright notice appear in all copies and that both that * copyright notice and this permission notice appear in supporting * documentation, and that the name of Red Hat not be used in advertising or * publicity pertaining to distribution of the software without specific, * written prior permission. Red Hat makes no representations about the * suitability of this software for any purpose. It is provided "as is" * without express or implied warranty. * * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS * SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND * FITNESS, IN NO EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY * SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN * AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING * OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS * SOFTWARE. * * Author: Rodrigo Kumpera ([email protected]) * André Tupinambá ([email protected]) * * Based on work by Owen Taylor and Søren Sandmann */ #ifdef HAVE_CONFIG_H #include <config.h> #endif #include <xmmintrin.h> /* for _mm_shuffle_pi16 and _MM_SHUFFLE */ #include <emmintrin.h> /* for SSE2 intrinsics */ #include "pixman-private.h" #include "pixman-combine32.h" #include "pixman-inlines.h" static __m128i mask_0080; static __m128i mask_00ff; static __m128i mask_0101; static __m128i mask_ffff; static __m128i mask_ff000000; static __m128i mask_alpha; static __m128i mask_565_r; static __m128i mask_565_g1, mask_565_g2; static __m128i mask_565_b; static __m128i mask_red; static __m128i mask_green; static __m128i mask_blue; static __m128i mask_565_fix_rb; static __m128i mask_565_fix_g; static force_inline __m128i unpack_32_1x128 (uint32_t data) { return _mm_unpacklo_epi8 (_mm_cvtsi32_si128 (data), _mm_setzero_si128 ()); } static force_inline void unpack_128_2x128 (__m128i data, __m128i* data_lo, __m128i* data_hi) { *data_lo = _mm_unpacklo_epi8 (data, _mm_setzero_si128 ()); *data_hi = _mm_unpackhi_epi8 (data, _mm_setzero_si128 ()); } static force_inline __m128i unpack_565_to_8888 (__m128i lo) { __m128i r, g, b, rb, t; r = _mm_and_si128 (_mm_slli_epi32 (lo, 8), mask_red); g = _mm_and_si128 (_mm_slli_epi32 (lo, 5), mask_green); b = _mm_and_si128 (_mm_slli_epi32 (lo, 3), mask_blue); rb = _mm_or_si128 (r, b); t = _mm_and_si128 (rb, mask_565_fix_rb); t = _mm_srli_epi32 (t, 5); rb = _mm_or_si128 (rb, t); t = _mm_and_si128 (g, mask_565_fix_g); t = _mm_srli_epi32 (t, 6); g = _mm_or_si128 (g, t); return _mm_or_si128 (rb, g); } static force_inline void unpack_565_128_4x128 (__m128i data, __m128i* data0, __m128i* data1, __m128i* data2, __m128i* data3) { __m128i lo, hi; lo = _mm_unpacklo_epi16 (data, _mm_setzero_si128 ()); hi = _mm_unpackhi_epi16 (data, _mm_setzero_si128 ()); lo = unpack_565_to_8888 (lo); hi = unpack_565_to_8888 (hi); unpack_128_2x128 (lo, data0, data1); unpack_128_2x128 (hi, data2, data3); } static force_inline uint16_t pack_565_32_16 (uint32_t pixel) { return (uint16_t) (((pixel >> 8) & 0xf800) | ((pixel >> 5) & 0x07e0) | ((pixel >> 3) & 0x001f)); } static force_inline __m128i pack_2x128_128 (__m128i lo, __m128i hi) { return _mm_packus_epi16 (lo, hi); } static force_inline __m128i pack_565_2x128_128 (__m128i lo, __m128i hi) { __m128i data; __m128i r, g1, g2, b; data = pack_2x128_128 (lo, hi); r = _mm_and_si128 (data, mask_565_r); g1 = _mm_and_si128 (_mm_slli_epi32 (data, 3), mask_565_g1); g2 = _mm_and_si128 (_mm_srli_epi32 (data, 5), mask_565_g2); b = _mm_and_si128 (_mm_srli_epi32 (data, 3), mask_565_b); return _mm_or_si128 (_mm_or_si128 (_mm_or_si128 (r, g1), g2), b); } static force_inline __m128i pack_565_4x128_128 (__m128i* xmm0, __m128i* xmm1, __m128i* xmm2, __m128i* xmm3) { return _mm_packus_epi16 (pack_565_2x128_128 (*xmm0, *xmm1), pack_565_2x128_128 (*xmm2, *xmm3)); } static force_inline int is_opaque (__m128i x) { __m128i ffs = _mm_cmpeq_epi8 (x, x); return (_mm_movemask_epi8 (_mm_cmpeq_epi8 (x, ffs)) & 0x8888) == 0x8888; } static force_inline int is_zero (__m128i x) { return _mm_movemask_epi8 ( _mm_cmpeq_epi8 (x, _mm_setzero_si128 ())) == 0xffff; } static force_inline int is_transparent (__m128i x) { return (_mm_movemask_epi8 ( _mm_cmpeq_epi8 (x, _mm_setzero_si128 ())) & 0x8888) == 0x8888; } static force_inline __m128i expand_pixel_32_1x128 (uint32_t data) { return _mm_shuffle_epi32 (unpack_32_1x128 (data), _MM_SHUFFLE (1, 0, 1, 0)); } static force_inline __m128i expand_alpha_1x128 (__m128i data) { return _mm_shufflehi_epi16 (_mm_shufflelo_epi16 (data, _MM_SHUFFLE (3, 3, 3, 3)), _MM_SHUFFLE (3, 3, 3, 3)); } static force_inline void expand_alpha_2x128 (__m128i data_lo, __m128i data_hi, __m128i* alpha_lo, __m128i* alpha_hi) { __m128i lo, hi; lo = _mm_shufflelo_epi16 (data_lo, _MM_SHUFFLE (3, 3, 3, 3)); hi = _mm_shufflelo_epi16 (data_hi, _MM_SHUFFLE (3, 3, 3, 3)); *alpha_lo = _mm_shufflehi_epi16 (lo, _MM_SHUFFLE (3, 3, 3, 3)); *alpha_hi = _mm_shufflehi_epi16 (hi, _MM_SHUFFLE (3, 3, 3, 3)); } static force_inline void expand_alpha_rev_2x128 (__m128i data_lo, __m128i data_hi, __m128i* alpha_lo, __m128i* alpha_hi) { __m128i lo, hi; lo = _mm_shufflelo_epi16 (data_lo, _MM_SHUFFLE (0, 0, 0, 0)); hi = _mm_shufflelo_epi16 (data_hi, _MM_SHUFFLE (0, 0, 0, 0)); *alpha_lo = _mm_shufflehi_epi16 (lo, _MM_SHUFFLE (0, 0, 0, 0)); *alpha_hi = _mm_shufflehi_epi16 (hi, _MM_SHUFFLE (0, 0, 0, 0)); } static force_inline void pix_multiply_2x128 (__m128i* data_lo, __m128i* data_hi, __m128i* alpha_lo, __m128i* alpha_hi, __m128i* ret_lo, __m128i* ret_hi) { __m128i lo, hi; lo = _mm_mullo_epi16 (*data_lo, *alpha_lo); hi = _mm_mullo_epi16 (*data_hi, *alpha_hi); lo = _mm_adds_epu16 (lo, mask_0080); hi = _mm_adds_epu16 (hi, mask_0080); *ret_lo = _mm_mulhi_epu16 (lo, mask_0101); *ret_hi = _mm_mulhi_epu16 (hi, mask_0101); } static force_inline void pix_add_multiply_2x128 (__m128i* src_lo, __m128i* src_hi, __m128i* alpha_dst_lo, __m128i* alpha_dst_hi, __m128i* dst_lo, __m128i* dst_hi, __m128i* alpha_src_lo, __m128i* alpha_src_hi, __m128i* ret_lo, __m128i* ret_hi) { __m128i t1_lo, t1_hi; __m128i t2_lo, t2_hi; pix_multiply_2x128 (src_lo, src_hi, alpha_dst_lo, alpha_dst_hi, &t1_lo, &t1_hi); pix_multiply_2x128 (dst_lo, dst_hi, alpha_src_lo, alpha_src_hi, &t2_lo, &t2_hi); *ret_lo = _mm_adds_epu8 (t1_lo, t2_lo); *ret_hi = _mm_adds_epu8 (t1_hi, t2_hi); } static force_inline void negate_2x128 (__m128i data_lo, __m128i data_hi, __m128i* neg_lo, __m128i* neg_hi) { *neg_lo = _mm_xor_si128 (data_lo, mask_00ff); *neg_hi = _mm_xor_si128 (data_hi, mask_00ff); } static force_inline void invert_colors_2x128 (__m128i data_lo, __m128i data_hi, __m128i* inv_lo, __m128i* inv_hi) { __m128i lo, hi; lo = _mm_shufflelo_ep