/* Copyright (C) 2007 Kenta Yoshimura All Rights Reserved. Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: - Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. - Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. - Neither the name of the Kenta Yoshimura nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. */ #include #include #include "bitblt.h" typedef vector unsigned char vecu8_t; typedef vector unsigned short vecu16_t; typedef vector unsigned int vecu32_t; typedef vector signed char veci8_t; typedef vector signed short veci16_t; typedef vector signed int veci32_t; void bitblt4bpp( bitmap_t * src, uint16_t src_x, //!< aligned 32 pixels uint16_t src_y, uint16_t width, //!< aligned 4 pixels uint16_t height, bitmap_t * dst, uint16_t dst_x, //!< aligned 4 pixels uint16_t dst_y, tlb4bpp_t * tlb ) { const vecu8_t * src_gbse = (const vecu8_t *)src->addr; uint8_t * dst_base = ( uint8_t *)dst->addr; const uint32_t src_re = src->row_bytes / sizeof(vecu8_t); const uint32_t dst_rb = dst->row_bytes; const vecu8_t tlb_a = *(const vecu8_t *)tlb->a; const vecu8_t tlb_r = *(const vecu8_t *)tlb->r; const vecu8_t tlb_g = *(const vecu8_t *)tlb->g; const vecu8_t tlb_b = *(const vecu8_t *)tlb->b; const vecu8_t v0x00 = vec_splat_u8(0x00); const vecu8_t v0x04 = vec_splat_u8(0x04); const vecu8_t v0x08 = vec_splat_u8(0x08); const vecu8_t v0x0f = vec_splat_u8(0x0f); const vecu8_t v0xff = (vecu8_t)vec_splat_s8(-1); const vecu16_t v0x00ff = (vecu16_t){ 0x00ff, 0x00ff, 0x00ff, 0x00ff, 0x00ff, 0x00ff, 0x00ff, 0x00ff}; const vecu8_t vmerge = (vecu8_t){ 0x01, 0x11, 0x03, 0x13, 0x05, 0x15, 0x07, 0x17, 0x09, 0x19, 0x0b, 0x1b, 0x0d, 0x1d, 0x0f, 0x1f}; // calculate the base pointer src_gbse = &src_gbse[src_x/2/sizeof(vecu8_t) + src_y*src_re]; dst_base = &dst_base[dst_x*sizeof(argb_t) + dst_y*dst_rb]; for ( int y = 0; y < height; y++ ) { const vecu8_t * src_ptr = &src_gbse[y*src_re]; vecu8_t * dst_ptr = (vecu8_t *)&dst_base[y*dst_rb]; vecu8_t lst_lvsl = vec_lvsl(0, &dst_base[y*dst_rb+width*4]); vecu8_t dst_lvsr = vec_lvsr(0, (uint8_t *)dst_ptr); vecu8_t lst_buf; vecu8_t lst_alp; for ( int x = 0; x < width; x += 32 ) { vecu8_t src_buf[9]; vecu8_t alp_buf[9]; // src_i = 01 23 45 67 89 ab cd ef fe dc ba 98 76 54 32 10 vecu8_t src_i = src_ptr[0]; // src_i0 = 01 03 05 07 09 0b 0d 0f 0e 0c 0a 08 06 04 02 00 // src_i1 = 00 02 04 06 08 0a 0c 0e 0f 0d 0b 09 07 05 03 01 vecu8_t src_i0 = vec_and(src_i, v0x0f); vecu8_t src_i1 = vec_sr(src_i, v0x04); // src_i2 = 00 01 02 03 04 05 06 07 08 09 0a 0b 0c 0d 0e 0f vecu8_t src_i2 = vec_mergeh(src_i1, src_i0); vecu8_t src_i3 = vec_mergel(src_i1, src_i0); // src_r0 = rr rr rr rr rr rr rr rr rr rr rr rr rr rr rr rr vecu8_t src_a0 = vec_perm(tlb_a, tlb_a, src_i2); vecu8_t src_r0 = vec_perm(tlb_r, tlb_r, src_i2); vecu8_t src_g0 = vec_perm(tlb_g, tlb_g, src_i2); vecu8_t src_b0 = vec_perm(tlb_b, tlb_b, src_i2); vecu8_t src_r1 = vec_perm(tlb_r, tlb_r, src_i3); vecu8_t src_g1 = vec_perm(tlb_g, tlb_g, src_i3); vecu8_t src_b1 = vec_perm(tlb_b, tlb_b, src_i3); vecu8_t src_a1 = vec_perm(tlb_a, tlb_a, src_i3); // src_ar0 = rr gg rr gg rr gg rr gg rr gg rr gg rr gg rr gg vecu16_t src_ar0 = (vecu16_t)vec_mergeh(src_a0, src_r0); vecu16_t src_ar1 = (vecu16_t)vec_mergel(src_a0, src_r0); // src_gb0 = bb aa bb aa bb aa bb aa bb aa bb aa bb aa bb aa vecu16_t src_gb0 = (vecu16_t)vec_mergeh(src_g0, src_b0); vecu16_t src_gb1 = (vecu16_t)vec_mergel(src_g0, src_b0); // src_aa0 = a0 a0 a1 a1 a2 a2 a3 a3 a4 a4 a5 a5 a6 a6 a7 a7 vecu16_t src_aa0 = (vecu16_t)vec_mergeh(src_a0, src_a0); vecu16_t src_aa1 = (vecu16_t)vec_mergel(src_a0, src_a0); // src_ar2 = rr gg rr gg rr gg rr gg rr gg rr gg rr gg rr gg vecu16_t src_ar2 = (vecu16_t)vec_mergeh(src_a1, src_r1); vecu16_t src_ar3 = (vecu16_t)vec_mergel(src_a1, src_r1); // src_gb2 = bb aa bb aa bb aa bb aa bb aa bb aa bb aa bb aa vecu16_t src_gb2 = (vecu16_t)vec_mergeh(src_g1, src_b1); vecu16_t src_gb3 = (vecu16_t)vec_mergel(src_b1, src_b1); // src_aa2 = a0 a0 a1 a1 a2 a2 a3 a3 a4 a4 a5 a5 a6 a6 a7 a7 vecu16_t src_aa2 = (vecu16_t)vec_mergeh(src_a1, src_a1); vecu16_t src_aa3 = (vecu16_t)vec_mergel(src_a1, src_a1); // src_buf[0] = rr gg bb aa rr gg bb aa rr gg bb aa rr gg bb aa src_buf[0] = (vecu8_t)vec_mergeh(src_ar0, src_gb0); src_buf[1] = (vecu8_t)vec_mergel(src_ar0, src_gb0); src_buf[2] = (vecu8_t)vec_mergeh(src_ar1, src_gb1); src_buf[3] = (vecu8_t)vec_mergel(src_ar1, src_gb1); src_buf[4] = (vecu8_t)vec_mergeh(src_ar2, src_gb2); src_buf[5] = (vecu8_t)vec_mergel(src_ar2, src_gb2); src_buf[6] = (vecu8_t)vec_mergeh(src_ar3, src_gb3); src_buf[7] = (vecu8_t)vec_mergel(src_ar3, src_gb3); alp_buf[0] = (vecu8_t)vec_mergeh(src_aa0, src_aa0); alp_buf[1] = (vecu8_t)vec_mergel(src_aa0, src_aa0); alp_buf[2] = (vecu8_t)vec_mergeh(src_aa1, src_aa1); alp_buf[3] = (vecu8_t)vec_mergel(src_aa1, src_aa1); alp_buf[4] = (vecu8_t)vec_mergeh(src_aa2, src_aa2); alp_buf[5] = (vecu8_t)vec_mergel(src_aa2, src_aa2); alp_buf[6] = (vecu8_t)vec_mergeh(src_aa3, src_aa3); alp_buf[7] = (vecu8_t)vec_mergel(src_aa3, src_aa3); int i = 0; if ( __builtin_expect(0==x, 0) ) { vecu8_t dst_org = dst_ptr[0]; vecu8_t src_alp = vec_perm(lst_alp, alp_buf[0], dst_lvsr); vecu8_t dst_alp = vec_sub(v0xff, src_alp); /* */ lst_buf = src_buf[0]; vecu8_t src_raw = vec_perm(v0x00, lst_buf, dst_lvsr); // dst_sol = (src*alpha + dst*alpha + alpha) >> 8 vecu16_t src_eve = vec_mule(src_raw, src_alp); vecu16_t src_odd = vec_mulo(src_raw, src_alp); vecu16_t dst_eve = vec_mule(dst_org, dst_alp); vecu16_t dst_odd = vec_mulo(dst_org, dst_alp); /* */ dst_eve = vec_add(dst_eve, src_eve); /* */ dst_odd = vec_add(dst_odd, src_odd); /* */ dst_eve = vec_add(dst_eve, vec_and(v0x00ff, (vecu16_t)src_alp)); /* */ dst_odd = vec_add(dst_odd, vec_and(v0x00ff, (vecu16_t)src_alp)); /* */ dst_eve = vec_sr(dst_eve, (vecu16_t)v0x08); /* */ dst_odd = vec_sr(dst_odd, (vecu16_t)v0x08); vecu8_t dst_sol = (vecu8_t)vec_perm(dst_eve, dst_odd, vmerge); // dst_ptr[0] = dst_lvsr>0x0f ? dst_sol : dst_org vecu8_t dst_sel = (vecu8_t)vec_cmpgt(dst_lvsr, v0x0f); dst_ptr[0] = vec_sel(dst_org, dst_sol, dst_sel); lst_alp = src_alp; i++; } for ( ; i<8 && x+i*4<=width-4; i++ ) { vecu8_t dst_org = dst_ptr[i]; vecu8_t cur_buf = src_buf[i]; vecu8_t src_alp = vec_perm(lst_alp, alp_buf[i], dst_lvsr); vecu8_t dst_alp = vec_sub(v0xff, src_alp); vecu8_t src_raw = vec_perm(lst_buf, cur_buf, dst_lvsr); // dst_sol = (src*alpha + dst*alpha + alpha) >> 8 vecu16_t src_eve = vec_mule(src_raw, src_alp); vecu16_t src_odd = vec_mulo(src_raw, src_alp); vecu16_t dst_eve = vec_mule(dst_org, dst_alp); vecu16_t dst_odd = vec_mulo(dst_org, dst_alp); /* */ dst_eve = vec_add(dst_eve, src_eve); /* */ dst_odd = vec_add(dst_odd, src_odd); /* */ dst_eve = vec_add(dst_eve, vec_and(v0x00ff, (vecu16_t)src_alp)); /* */ dst_odd = vec_add(dst_odd, vec_and(v0x00ff, (vecu16_t)src_alp)); /* */ dst_eve = vec_sr(dst_eve, (vecu16_t)v0x08); /* */ dst_odd = vec_sr(dst_odd, (vecu16_t)v0x08); vecu8_t dst_sol = (vecu8_t)vec_perm(dst_eve, dst_odd, vmerge); dst_ptr[i] = dst_sol; lst_buf = cur_buf; lst_alp = src_alp; } // x+i*4=width が成立すれば右端 int is_inside = (x+i*4> 8 vecu16_t src_eve = vec_mule(src_raw, src_alp); vecu16_t src_odd = vec_mulo(src_raw, src_alp); vecu16_t dst_eve = vec_mule(dst_org, dst_alp); vecu16_t dst_odd = vec_mulo(dst_org, dst_alp); /* */ dst_eve = vec_add(dst_eve, src_eve); /* */ dst_odd = vec_add(dst_odd, src_odd); /* */ dst_eve = vec_add(dst_eve, vec_and(v0x00ff, (vecu16_t)src_alp)); /* */ dst_odd = vec_add(dst_odd, vec_and(v0x00ff, (vecu16_t)src_alp)); /* */ dst_eve = vec_sr(dst_eve, (vecu16_t)v0x08); /* */ dst_odd = vec_sr(dst_odd, (vecu16_t)v0x08); vecu8_t dst_sol = (vecu8_t)vec_perm(dst_eve, dst_odd, vmerge); vecu8_t dst_sel = (vecu8_t)vec_cmpgt(lst_lvsl, v0x0f); dst_ptr[i] = vec_sel(dst_sol, dst_org, dst_sel); } src_ptr++; dst_ptr += 8; } } }