2 * Copyright (c) 2008 Mans Rullgard <mans@mansr.com>
4 * This file is part of FFmpeg.
6 * FFmpeg is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation; either
9 * version 2.1 of the License, or (at your option) any later version.
11 * FFmpeg is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Lesser General Public License for more details.
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with FFmpeg; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
26 function ff_h264_idct_add_neon, export=1
27 vld1.64 {d0-d3}, [r1,:128]
44 vld1.32 {d18[0]}, [r0,:32], r2
47 vld1.32 {d19[1]}, [r0,:32], r2
49 vld1.32 {d18[1]}, [r0,:32], r2
51 vld1.32 {d19[0]}, [r0,:32], r2
53 sub r0, r0, r2, lsl #2
66 vst1.32 {d0[0]}, [r0,:32], r2
67 vst1.32 {d1[1]}, [r0,:32], r2
68 vst1.32 {d0[1]}, [r0,:32], r2
69 vst1.32 {d1[0]}, [r0,:32], r2
74 function ff_h264_idct_dc_add_neon, export=1
75 vld1.16 {d2[],d3[]}, [r1,:16]
77 vld1.32 {d0[0]}, [r0,:32], r2
78 vld1.32 {d0[1]}, [r0,:32], r2
80 vld1.32 {d1[0]}, [r0,:32], r2
81 vld1.32 {d1[1]}, [r0,:32], r2
85 sub r0, r0, r2, lsl #2
86 vst1.32 {d0[0]}, [r0,:32], r2
87 vst1.32 {d0[1]}, [r0,:32], r2
88 vst1.32 {d1[0]}, [r0,:32], r2
89 vst1.32 {d1[1]}, [r0,:32], r2
93 function ff_h264_idct_add16_neon, export=1
111 adrne lr, ff_h264_idct_dc_add_neon
112 adreq lr, ff_h264_idct_add_neon
120 function ff_h264_idct_add16intra_neon, export=1
135 adrne lr, ff_h264_idct_add_neon
136 adreq lr, ff_h264_idct_dc_add_neon
145 function ff_h264_idct_add8_neon, export=1
162 adrne lr, ff_h264_idct_add_neon
163 adreq lr, ff_h264_idct_dc_add_neon
173 scan8: .byte 4+1*8, 5+1*8, 4+2*8, 5+2*8
174 .byte 6+1*8, 7+1*8, 6+2*8, 7+2*8
175 .byte 4+3*8, 5+3*8, 4+4*8, 5+4*8
176 .byte 6+3*8, 7+3*8, 6+4*8, 7+4*8