3
3
* Copyright (c) 2000, 2001, 2002 Fabrice Bellard.
5
* This library is free software; you can redistribute it and/or
4
* Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at>
6
* This file is part of FFmpeg.
8
* FFmpeg is free software; you can redistribute it and/or
6
9
* modify it under the terms of the GNU Lesser General Public
7
10
* License as published by the Free Software Foundation; either
8
* version 2 of the License, or (at your option) any later version.
11
* version 2.1 of the License, or (at your option) any later version.
10
* This library is distributed in the hope that it will be useful,
13
* FFmpeg is distributed in the hope that it will be useful,
11
14
* but WITHOUT ANY WARRANTY; without even the implied warranty of
12
15
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13
16
* Lesser General Public License for more details.
15
18
* You should have received a copy of the GNU Lesser General Public
16
* License along with this library; if not, write to the Free Software
17
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
19
* License along with FFmpeg; if not, write to the Free Software
20
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
36
39
typedef short DCTELEM;
38
42
void fdct_ifast (DCTELEM *data);
43
void fdct_ifast248 (DCTELEM *data);
39
44
void ff_jpeg_fdct_islow (DCTELEM *data);
45
void ff_fdct248_islow (DCTELEM *data);
41
47
void j_rev_dct (DCTELEM *data);
48
void j_rev_dct4 (DCTELEM *data);
49
void j_rev_dct2 (DCTELEM *data);
50
void j_rev_dct1 (DCTELEM *data);
43
52
void ff_fdct_mmx(DCTELEM *block);
53
void ff_fdct_mmx2(DCTELEM *block);
54
void ff_fdct_sse2(DCTELEM *block);
56
void ff_h264_idct8_add_c(uint8_t *dst, DCTELEM *block, int stride);
57
void ff_h264_idct_add_c(uint8_t *dst, DCTELEM *block, int stride);
58
void ff_h264_idct8_dc_add_c(uint8_t *dst, DCTELEM *block, int stride);
59
void ff_h264_idct_dc_add_c(uint8_t *dst, DCTELEM *block, int stride);
60
void ff_h264_lowres_idct_add_c(uint8_t *dst, int stride, DCTELEM *block);
61
void ff_h264_lowres_idct_put_c(uint8_t *dst, int stride, DCTELEM *block);
63
void ff_vector_fmul_add_add_c(float *dst, const float *src0, const float *src1,
64
const float *src2, int src3, int blocksize, int step);
65
void ff_float_to_int16_c(int16_t *dst, const float *src, int len);
45
67
/* encoding scans */
46
68
extern const uint8_t ff_alternate_horizontal_scan[64];
47
69
extern const uint8_t ff_alternate_vertical_scan[64];
48
70
extern const uint8_t ff_zigzag_direct[64];
71
extern const uint8_t ff_zigzag248_direct[64];
50
73
/* pixel operations */
51
#define MAX_NEG_CROP 384
74
#define MAX_NEG_CROP 1024
54
extern uint32_t squareTbl[512];
55
extern uint8_t cropTbl[256 + 2 * MAX_NEG_CROP];
77
extern uint32_t ff_squareTbl[512];
78
extern uint8_t ff_cropTbl[256 + 2 * MAX_NEG_CROP];
80
/* VP3 DSP functions */
81
void ff_vp3_idct_c(DCTELEM *block/* align 16*/);
82
void ff_vp3_idct_put_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/);
83
void ff_vp3_idct_add_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/);
85
/* 1/2^n downscaling functions from imgconvert.c */
86
void ff_img_copy_plane(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height);
87
void ff_shrink22(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height);
88
void ff_shrink44(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height);
89
void ff_shrink88(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height);
91
void ff_gmc_c(uint8_t *dst, uint8_t *src, int stride, int h, int ox, int oy,
92
int dxx, int dxy, int dyx, int dyy, int shift, int r, int width, int height);
58
94
/* minimum alignment rules ;)
59
95
if u notice errors in the align stuff, need more alignment for some asm code for some cpu
76
112
/* add and put pixel (decoding) */
77
113
// blocksizes for op_pixels_func are 8x4,8x8 16x8 16x16
114
//h for op_pixels_func is limited to {width/2, width} but never larger than 16 and never smaller then 4
78
115
typedef void (*op_pixels_func)(uint8_t *block/*align width (8 or 16)*/, const uint8_t *pixels/*align 1*/, int line_size, int h);
79
116
typedef void (*tpel_mc_func)(uint8_t *block/*align width (8 or 16)*/, const uint8_t *pixels/*align 1*/, int line_size, int w, int h);
80
117
typedef void (*qpel_mc_func)(uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, int stride);
81
118
typedef void (*h264_chroma_mc_func)(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int srcStride, int h, int x, int y);
119
typedef void (*h264_weight_func)(uint8_t *block, int stride, int log2_denom, int weight, int offset);
120
typedef void (*h264_biweight_func)(uint8_t *dst, uint8_t *src, int stride, int log2_denom, int weightd, int weights, int offset);
83
122
#define DEF_OLD_QPEL(name)\
84
123
void ff_put_ ## name (uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, int stride);\
128
172
* global motion compensation.
130
174
void (*gmc )(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int ox, int oy,
131
int dxx, int dxy, int dyx, int dyy, int shift, int r, int width, int height);
175
int dxx, int dxy, int dyx, int dyy, int shift, int r, int width, int height);
132
176
void (*clear_blocks)(DCTELEM *blocks/*align 16*/);
133
177
int (*pix_sum)(uint8_t * pix, int line_size);
134
178
int (*pix_norm1)(uint8_t * pix, int line_size);
135
me_cmp_func sad[2]; /* identical to pix_absAxA except additional void * */
137
me_cmp_func hadamard8_diff[2];
138
me_cmp_func dct_sad[2];
139
me_cmp_func quant_psnr[2];
142
int (*hadamard8_abs )(uint8_t *src, int stride, int mean);
144
me_cmp_func me_pre_cmp[11];
145
me_cmp_func me_cmp[11];
146
me_cmp_func me_sub_cmp[11];
147
me_cmp_func mb_cmp[11];
149
/* maybe create an array for 16/8/4/2 functions */
179
// 16x16 8x8 4x4 2x2 16x8 8x4 4x2 8x16 4x8 2x4
181
me_cmp_func sad[5]; /* identical to pix_absAxA except additional void * */
183
me_cmp_func hadamard8_diff[5];
184
me_cmp_func dct_sad[5];
185
me_cmp_func quant_psnr[5];
193
me_cmp_func dct_max[5];
194
me_cmp_func dct264_sad[5];
196
me_cmp_func me_pre_cmp[5];
197
me_cmp_func me_cmp[5];
198
me_cmp_func me_sub_cmp[5];
199
me_cmp_func mb_cmp[5];
200
me_cmp_func ildct_cmp[5]; //only width 16 used
201
me_cmp_func frame_skip_cmp[5]; //only width 8 used
151
204
* Halfpel motion compensation with rounding (a+b+1)>>1.
152
* this is an array[4][4] of motion compensation funcions for 4
153
* horizontal blocksizes (2,4,8,16) and the 4 halfpel positions<br>
205
* this is an array[4][4] of motion compensation funcions for 4
206
* horizontal blocksizes (8,16) and the 4 halfpel positions<br>
154
207
* *pixels_tab[ 0->16xH 1->8xH ][ xhalfpel + 2*yhalfpel ]
155
208
* @param block destination where the result is stored
156
209
* @param pixels source
212
267
qpel_mc_func put_no_rnd_qpel_pixels_tab[2][16];
213
268
qpel_mc_func avg_no_rnd_qpel_pixels_tab[2][16];
214
269
qpel_mc_func put_mspel_pixels_tab[8];
219
274
h264_chroma_mc_func put_h264_chroma_pixels_tab[3];
275
/* This is really one func used in VC-1 decoding */
276
h264_chroma_mc_func put_no_rnd_h264_chroma_pixels_tab[3];
220
277
h264_chroma_mc_func avg_h264_chroma_pixels_tab[3];
222
qpel_mc_func put_h264_qpel_pixels_tab[3][16];
223
qpel_mc_func avg_h264_qpel_pixels_tab[3][16];
225
op_pixels_abs_func pix_abs16x16;
226
op_pixels_abs_func pix_abs16x16_x2;
227
op_pixels_abs_func pix_abs16x16_y2;
228
op_pixels_abs_func pix_abs16x16_xy2;
229
op_pixels_abs_func pix_abs8x8;
230
op_pixels_abs_func pix_abs8x8_x2;
231
op_pixels_abs_func pix_abs8x8_y2;
232
op_pixels_abs_func pix_abs8x8_xy2;
279
qpel_mc_func put_h264_qpel_pixels_tab[4][16];
280
qpel_mc_func avg_h264_qpel_pixels_tab[4][16];
282
qpel_mc_func put_2tap_qpel_pixels_tab[4][16];
283
qpel_mc_func avg_2tap_qpel_pixels_tab[4][16];
285
h264_weight_func weight_h264_pixels_tab[10];
286
h264_biweight_func biweight_h264_pixels_tab[10];
289
qpel_mc_func put_cavs_qpel_pixels_tab[2][16];
290
qpel_mc_func avg_cavs_qpel_pixels_tab[2][16];
291
void (*cavs_filter_lv)(uint8_t *pix, int stride, int alpha, int beta, int tc, int bs1, int bs2);
292
void (*cavs_filter_lh)(uint8_t *pix, int stride, int alpha, int beta, int tc, int bs1, int bs2);
293
void (*cavs_filter_cv)(uint8_t *pix, int stride, int alpha, int beta, int tc, int bs1, int bs2);
294
void (*cavs_filter_ch)(uint8_t *pix, int stride, int alpha, int beta, int tc, int bs1, int bs2);
295
void (*cavs_idct8_add)(uint8_t *dst, DCTELEM *block, int stride);
297
me_cmp_func pix_abs[2][4];
234
299
/* huffyuv specific */
235
300
void (*add_bytes)(uint8_t *dst/*align 16*/, uint8_t *src/*align 16*/, int w);
236
301
void (*diff_bytes)(uint8_t *dst/*align 16*/, uint8_t *src1/*align 16*/, uint8_t *src2/*align 1*/,int w);
303
* subtract huffyuv's variant of median prediction
304
* note, this might read from src1[-1], src2[-1]
306
void (*sub_hfyu_median_prediction)(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w, int *left, int *left_top);
237
307
void (*bswap_buf)(uint32_t *dst, uint32_t *src, int w);
309
void (*h264_v_loop_filter_luma)(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0);
310
void (*h264_h_loop_filter_luma)(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0);
311
void (*h264_v_loop_filter_chroma)(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0);
312
void (*h264_h_loop_filter_chroma)(uint8_t *pix, int stride, int alpha, int beta, int8_t *tc0);
313
void (*h264_v_loop_filter_chroma_intra)(uint8_t *pix, int stride, int alpha, int beta);
314
void (*h264_h_loop_filter_chroma_intra)(uint8_t *pix, int stride, int alpha, int beta);
315
// h264_loop_filter_strength: simd only. the C version is inlined in h264.c
316
void (*h264_loop_filter_strength)(int16_t bS[2][4][4], uint8_t nnz[40], int8_t ref[2][40], int16_t mv[2][40][2],
317
int bidir, int edges, int step, int mask_mv0, int mask_mv1);
319
void (*h263_v_loop_filter)(uint8_t *src, int stride, int qscale);
320
void (*h263_h_loop_filter)(uint8_t *src, int stride, int qscale);
322
void (*h261_loop_filter)(uint8_t *src, int stride);
324
/* assume len is a multiple of 4, and arrays are 16-byte aligned */
325
void (*vorbis_inverse_coupling)(float *mag, float *ang, int blocksize);
326
/* assume len is a multiple of 8, and arrays are 16-byte aligned */
327
void (*vector_fmul)(float *dst, const float *src, int len);
328
void (*vector_fmul_reverse)(float *dst, const float *src0, const float *src1, int len);
329
/* assume len is a multiple of 8, and src arrays are 16-byte aligned */
330
void (*vector_fmul_add_add)(float *dst, const float *src0, const float *src1, const float *src2, int src3, int len, int step);
332
/* C version: convert floats from the range [384.0,386.0] to ints in [-32768,32767]
333
* simd versions: convert floats from [-32768.0,32767.0] without rescaling and arrays are 16byte aligned */
334
void (*float_to_int16)(int16_t *dst, const float *src, int len);
240
337
void (*fdct)(DCTELEM *block/* align 16*/);
338
void (*fdct248)(DCTELEM *block/* align 16*/);
243
341
void (*idct)(DCTELEM *block/* align 16*/);
246
344
* block -> idct -> clip to unsigned 8 bit -> dest.
247
345
* (-1392, 0, 0, ...) -> idct -> (-174, -174, ...) -> put -> (0, 0, ...)
248
346
* @param line_size size in bytes of a horizotal line of dest
250
348
void (*idct_put)(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/);
253
351
* block -> idct -> add dest -> clip to unsigned 8 bit -> dest.
254
352
* @param line_size size in bytes of a horizotal line of dest
256
354
void (*idct_add)(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/);
259
357
* idct input permutation.
260
358
* several optimized IDCTs need a permutated input (relative to the normal order of the reference
273
371
#define FF_LIBMPEG2_IDCT_PERM 2
274
372
#define FF_SIMPLE_IDCT_PERM 3
275
373
#define FF_TRANSPOSE_IDCT_PERM 4
374
#define FF_PARTTRANS_IDCT_PERM 5
376
int (*try_8x8basis)(int16_t rem[64], int16_t weight[64], int16_t basis[64], int scale);
377
void (*add_8x8basis)(int16_t rem[64], int16_t basis[64], int scale);
378
#define BASIS_SHIFT 16
379
#define RECON_SHIFT 6
382
void (*h264_idct_add)(uint8_t *dst, DCTELEM *block, int stride);
383
void (*h264_idct8_add)(uint8_t *dst, DCTELEM *block, int stride);
384
void (*h264_idct_dc_add)(uint8_t *dst, DCTELEM *block, int stride);
385
void (*h264_idct8_dc_add)(uint8_t *dst, DCTELEM *block, int stride);
386
void (*h264_dct)(DCTELEM block[4][4]);
389
void (*vertical_compose97i)(DWTELEM *b0, DWTELEM *b1, DWTELEM *b2, DWTELEM *b3, DWTELEM *b4, DWTELEM *b5, int width);
390
void (*horizontal_compose97i)(DWTELEM *b, int width);
391
void (*inner_add_yblock)(const uint8_t *obmc, const int obmc_stride, uint8_t * * block, int b_w, int b_h, int src_x, int src_y, int src_stride, slice_buffer * sb, int add, uint8_t * dst8);
393
void (*prefetch)(void *mem, int stride, int h);
395
void (*shrink[4])(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height);
398
void (*vc1_inv_trans_8x8)(DCTELEM *b);
399
void (*vc1_inv_trans_8x4)(DCTELEM *b, int n);
400
void (*vc1_inv_trans_4x8)(DCTELEM *b, int n);
401
void (*vc1_inv_trans_4x4)(DCTELEM *b, int n);
402
void (*vc1_v_overlap)(uint8_t* src, int stride);
403
void (*vc1_h_overlap)(uint8_t* src, int stride);
404
/* put 8x8 block with bicubic interpolation and quarterpel precision
405
* last argument is actually round value instead of height
407
op_pixels_func put_vc1_mspel_pixels_tab[16];
279
410
void dsputil_static_init(void);
280
411
void dsputil_init(DSPContext* p, AVCodecContext *avctx);
413
int ff_check_alignment(void);
283
416
* permute block according to permuatation.
284
417
* @param last last non zero element in scantable order
286
419
void ff_block_permute(DCTELEM *block, uint8_t *permutation, const uint8_t *scantable, int last);
288
#define BYTE_VEC32(c) ((c)*0x01010101UL)
421
void ff_set_cmp(DSPContext* c, me_cmp_func *cmp, int type);
423
#define BYTE_VEC32(c) ((c)*0x01010101UL)
290
425
static inline uint32_t rnd_avg32(uint32_t a, uint32_t b)
343
518
#elif defined(ARCH_ARMV4L)
345
520
/* This is to use 4 bytes read to the IDCT pointers for some 'zero'
347
#define __align8 __attribute__ ((aligned (4)))
521
line optimizations */
522
#define DECLARE_ALIGNED_8(t,v) t v __attribute__ ((aligned (4)))
523
#define STRIDE_ALIGN 4
525
#define MM_IWMMXT 0x0100 /* XScale IWMMXT */
349
529
void dsputil_init_armv4l(DSPContext* c, AVCodecContext *avctx);
351
531
#elif defined(HAVE_MLIB)
353
533
/* SPARC/VIS IDCT needs 8-byte aligned DCT blocks */
354
#define __align8 __attribute__ ((aligned (8)))
534
#define DECLARE_ALIGNED_8(t,v) t v __attribute__ ((aligned (8)))
535
#define STRIDE_ALIGN 8
356
537
void dsputil_init_mlib(DSPContext* c, AVCodecContext *avctx);
539
#elif defined(ARCH_SPARC)
541
/* SPARC/VIS IDCT needs 8-byte aligned DCT blocks */
542
#define DECLARE_ALIGNED_8(t,v) t v __attribute__ ((aligned (8)))
543
#define STRIDE_ALIGN 8
544
void dsputil_init_vis(DSPContext* c, AVCodecContext *avctx);
358
546
#elif defined(ARCH_ALPHA)
360
#define __align8 __attribute__ ((aligned (8)))
548
#define DECLARE_ALIGNED_8(t,v) t v __attribute__ ((aligned (8)))
549
#define STRIDE_ALIGN 8
362
551
void dsputil_init_alpha(DSPContext* c, AVCodecContext *avctx);
368
557
extern int mm_flags;
370
559
#if defined(HAVE_ALTIVEC) && !defined(CONFIG_DARWIN)
560
#define pixel altivec_pixel
371
561
#include <altivec.h>
374
#define __align8 __attribute__ ((aligned (16)))
565
#define DECLARE_ALIGNED_8(t,v) t v __attribute__ ((aligned (16)))
566
#define STRIDE_ALIGN 16
376
568
void dsputil_init_ppc(DSPContext* c, AVCodecContext *avctx);
378
570
#elif defined(HAVE_MMI)
380
#define __align8 __attribute__ ((aligned (16)))
572
#define DECLARE_ALIGNED_8(t,v) t v __attribute__ ((aligned (16)))
573
#define STRIDE_ALIGN 16
382
575
void dsputil_init_mmi(DSPContext* c, AVCodecContext *avctx);
384
577
#elif defined(ARCH_SH4)
386
#define __align8 __attribute__ ((aligned (8)))
579
#define DECLARE_ALIGNED_8(t,v) t v __attribute__ ((aligned (8)))
580
#define STRIDE_ALIGN 8
388
582
void dsputil_init_sh4(DSPContext* c, AVCodecContext *avctx);
584
#elif defined(ARCH_BFIN)
586
#define DECLARE_ALIGNED_8(t,v) t v __attribute__ ((aligned (8)))
587
#define STRIDE_ALIGN 8
589
void dsputil_init_bfin(DSPContext* c, AVCodecContext *avctx);
593
#define DECLARE_ALIGNED_8(t,v) t v __attribute__ ((aligned (8)))
594
#define STRIDE_ALIGN 8
398
struct unaligned_64 { uint64_t l; } __attribute__((packed));
399
struct unaligned_32 { uint32_t l; } __attribute__((packed));
400
struct unaligned_16 { uint16_t l; } __attribute__((packed));
402
#define LD16(a) (((const struct unaligned_16 *) (a))->l)
403
#define LD32(a) (((const struct unaligned_32 *) (a))->l)
404
#define LD64(a) (((const struct unaligned_64 *) (a))->l)
406
#define ST32(a, b) (((struct unaligned_32 *) (a))->l) = (b)
410
#define LD16(a) (*((uint16_t*)(a)))
411
#define LD32(a) (*((uint32_t*)(a)))
412
#define LD64(a) (*((uint64_t*)(a)))
414
#define ST32(a, b) *((uint32_t*)(a)) = (b)
416
#endif /* !__GNUC__ */
419
599
void get_psnr(uint8_t *orig_image[3], uint8_t *coded_image[3],
420
600
int orig_linesize[3], int coded_linesize,
437
619
FFTComplex *exptab;
438
620
FFTComplex *exptab1; /* only used by SSE code */
439
621
void (*fft_calc)(struct FFTContext *s, FFTComplex *z);
622
void (*imdct_calc)(struct MDCTContext *s, FFTSample *output,
623
const FFTSample *input, FFTSample *tmp);
442
int fft_init(FFTContext *s, int nbits, int inverse);
443
void fft_permute(FFTContext *s, FFTComplex *z);
444
void fft_calc_c(FFTContext *s, FFTComplex *z);
445
void fft_calc_sse(FFTContext *s, FFTComplex *z);
446
void fft_calc_altivec(FFTContext *s, FFTComplex *z);
626
int ff_fft_init(FFTContext *s, int nbits, int inverse);
627
void ff_fft_permute(FFTContext *s, FFTComplex *z);
628
void ff_fft_calc_c(FFTContext *s, FFTComplex *z);
629
void ff_fft_calc_sse(FFTContext *s, FFTComplex *z);
630
void ff_fft_calc_3dn(FFTContext *s, FFTComplex *z);
631
void ff_fft_calc_3dn2(FFTContext *s, FFTComplex *z);
632
void ff_fft_calc_altivec(FFTContext *s, FFTComplex *z);
448
static inline void fft_calc(FFTContext *s, FFTComplex *z)
634
static inline void ff_fft_calc(FFTContext *s, FFTComplex *z)
450
636
s->fft_calc(s, z);
452
void fft_end(FFTContext *s);
638
void ff_fft_end(FFTContext *s);
454
640
/* MDCT computation */
465
651
int ff_mdct_init(MDCTContext *s, int nbits, int inverse);
466
652
void ff_imdct_calc(MDCTContext *s, FFTSample *output,
467
653
const FFTSample *input, FFTSample *tmp);
654
void ff_imdct_calc_3dn2(MDCTContext *s, FFTSample *output,
655
const FFTSample *input, FFTSample *tmp);
656
void ff_imdct_calc_sse(MDCTContext *s, FFTSample *output,
657
const FFTSample *input, FFTSample *tmp);
468
658
void ff_mdct_calc(MDCTContext *s, FFTSample *out,
469
659
const FFTSample *input, FFTSample *tmp);
470
660
void ff_mdct_end(MDCTContext *s);
472
#define WARPER88_1616(name8, name16)\
473
static int name16(void /*MpegEncContext*/ *s, uint8_t *dst, uint8_t *src, int stride){\
474
return name8(s, dst , src , stride)\
475
+name8(s, dst+8 , src+8 , stride)\
476
+name8(s, dst +8*stride, src +8*stride, stride)\
477
+name8(s, dst+8+8*stride, src+8+8*stride, stride);\
481
/* XXX: add ISOC specific test to avoid specific BSD testing. */
482
/* better than nothing implementation. */
483
/* btw, rintf() is existing on fbsd too -- alex */
484
static inline long int lrintf(float x)
487
/* XXX: incorrect, but make it compile */
490
return (int)(rint(x));
495
#if defined(CONFIG_OS2) || defined(CONFIG_SUNOS)
496
static inline float floorf(float f) {
662
#define WARPER8_16(name8, name16)\
663
static int name16(void /*MpegEncContext*/ *s, uint8_t *dst, uint8_t *src, int stride, int h){\
664
return name8(s, dst , src , stride, h)\
665
+name8(s, dst+8 , src+8 , stride, h);\
668
#define WARPER8_16_SQ(name8, name16)\
669
static int name16(void /*MpegEncContext*/ *s, uint8_t *dst, uint8_t *src, int stride, int h){\
671
score +=name8(s, dst , src , stride, 8);\
672
score +=name8(s, dst+8 , src+8 , stride, 8);\
676
score +=name8(s, dst , src , stride, 8);\
677
score +=name8(s, dst+8 , src+8 , stride, 8);\
683
static inline void copy_block2(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
688
ST16(dst , LD16(src ));
694
static inline void copy_block4(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
699
ST32(dst , LD32(src ));
705
static inline void copy_block8(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
710
ST32(dst , LD32(src ));
711
ST32(dst+4 , LD32(src+4 ));
717
static inline void copy_block9(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
722
ST32(dst , LD32(src ));
723
ST32(dst+4 , LD32(src+4 ));
730
static inline void copy_block16(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
735
ST32(dst , LD32(src ));
736
ST32(dst+4 , LD32(src+4 ));
737
ST32(dst+8 , LD32(src+8 ));
738
ST32(dst+12, LD32(src+12));
744
static inline void copy_block17(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
749
ST32(dst , LD32(src ));
750
ST32(dst+4 , LD32(src+4 ));
751
ST32(dst+8 , LD32(src+8 ));
752
ST32(dst+12, LD32(src+12));