1 /*
2  * Copyright (c) 2009 David Conrad <lessen42@gmail.com>
3  *
4  * This file is part of FFmpeg.
5  *
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.
10  *
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.
15  *
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
19  */
20 
21 #include <stdint.h>
22 
23 #include "libavutil/attributes.h"
24 #include "libavutil/cpu.h"
25 #include "libavutil/x86/cpu.h"
26 #include "libavutil/x86/asm.h"
27 #include "libavcodec/avcodec.h"
28 #include "libavcodec/vp3dsp.h"
29 
30 #include "config.h"
31 
32 void ff_vp3_idct_put_mmx(uint8_t *dest, int line_size, int16_t *block);
33 void ff_vp3_idct_add_mmx(uint8_t *dest, int line_size, int16_t *block);
34 
35 void ff_vp3_idct_put_sse2(uint8_t *dest, int line_size, int16_t *block);
36 void ff_vp3_idct_add_sse2(uint8_t *dest, int line_size, int16_t *block);
37 
38 void ff_vp3_idct_dc_add_mmxext(uint8_t *dest, int line_size,
39                                int16_t *block);
40 
41 void ff_vp3_v_loop_filter_mmxext(uint8_t *src, int stride,
42                                  int *bounding_values);
43 void ff_vp3_h_loop_filter_mmxext(uint8_t *src, int stride,
44                                  int *bounding_values);
45 
46 #if HAVE_MMX_INLINE
47 
48 #define MOVQ_BFE(regd)                                  \
49     __asm__ volatile (                                  \
50         "pcmpeqd %%"#regd", %%"#regd"   \n\t"           \
51         "paddb   %%"#regd", %%"#regd"   \n\t" ::)
52 
53 #define PAVGBP_MMX_NO_RND(rega, regb, regr,  regc, regd, regp)   \
54     "movq  "#rega", "#regr"             \n\t"                    \
55     "movq  "#regc", "#regp"             \n\t"                    \
56     "pand  "#regb", "#regr"             \n\t"                    \
57     "pand  "#regd", "#regp"             \n\t"                    \
58     "pxor  "#rega", "#regb"             \n\t"                    \
59     "pxor  "#regc", "#regd"             \n\t"                    \
60     "pand    %%mm6, "#regb"             \n\t"                    \
61     "pand    %%mm6, "#regd"             \n\t"                    \
62     "psrlq      $1, "#regb"             \n\t"                    \
63     "psrlq      $1, "#regd"             \n\t"                    \
64     "paddb "#regb", "#regr"             \n\t"                    \
65     "paddb "#regd", "#regp"             \n\t"
66 
67 #if HAVE_6REGS
put_vp_no_rnd_pixels8_l2_mmx(uint8_t * dst,const uint8_t * a,const uint8_t * b,ptrdiff_t stride,int h)68 static void put_vp_no_rnd_pixels8_l2_mmx(uint8_t *dst, const uint8_t *a, const uint8_t *b, ptrdiff_t stride, int h)
69 {
70 //    START_TIMER
71     MOVQ_BFE(mm6);
72     __asm__ volatile(
73         "1:                             \n\t"
74         "movq   (%1), %%mm0             \n\t"
75         "movq   (%2), %%mm1             \n\t"
76         "movq   (%1,%4), %%mm2          \n\t"
77         "movq   (%2,%4), %%mm3          \n\t"
78         PAVGBP_MMX_NO_RND(%%mm0, %%mm1, %%mm4,   %%mm2, %%mm3, %%mm5)
79         "movq   %%mm4, (%3)             \n\t"
80         "movq   %%mm5, (%3,%4)          \n\t"
81 
82         "movq   (%1,%4,2), %%mm0        \n\t"
83         "movq   (%2,%4,2), %%mm1        \n\t"
84         "movq   (%1,%5), %%mm2          \n\t"
85         "movq   (%2,%5), %%mm3          \n\t"
86         "lea    (%1,%4,4), %1           \n\t"
87         "lea    (%2,%4,4), %2           \n\t"
88         PAVGBP_MMX_NO_RND(%%mm0, %%mm1, %%mm4,   %%mm2, %%mm3, %%mm5)
89         "movq   %%mm4, (%3,%4,2)        \n\t"
90         "movq   %%mm5, (%3,%5)          \n\t"
91         "lea    (%3,%4,4), %3           \n\t"
92         "subl   $4, %0                  \n\t"
93         "jnz    1b                      \n\t"
94         :"+r"(h), "+r"(a), "+r"(b), "+r"(dst)
95         :"r"((x86_reg)stride), "r"((x86_reg)3L*stride)
96         :"memory");
97 //    STOP_TIMER("put_vp_no_rnd_pixels8_l2_mmx")
98 }
99 #endif /*HAVE_6REGS */
100 #endif /* HAVE_MMX_INLINE */
101 
ff_vp3dsp_init_x86(VP3DSPContext * c,int flags)102 av_cold void ff_vp3dsp_init_x86(VP3DSPContext *c, int flags)
103 {
104     int cpu_flags = av_get_cpu_flags();
105 
106 #if HAVE_6REGS && HAVE_MMX_INLINE
107     c->put_no_rnd_pixels_l2 = put_vp_no_rnd_pixels8_l2_mmx;
108 #endif /* HAVE_6REGS && HAVE_MMX_INLINE */
109 
110 #if ARCH_X86_32
111 #if (HAVE_MMX_EXTERNAL == 1)
112     if (EXTERNAL_MMX(cpu_flags)) {
113         c->idct_put  = ff_vp3_idct_put_mmx;
114         c->idct_add  = ff_vp3_idct_add_mmx;
115     }
116 #endif
117 #endif
118 
119 #if (HAVE_MMXEXT_EXTERNAL == 1)
120     if (EXTERNAL_MMXEXT(cpu_flags)) {
121         c->idct_dc_add = ff_vp3_idct_dc_add_mmxext;
122 
123         if (!(flags & CODEC_FLAG_BITEXACT)) {
124             c->v_loop_filter = ff_vp3_v_loop_filter_mmxext;
125             c->h_loop_filter = ff_vp3_h_loop_filter_mmxext;
126         }
127     }
128 #endif
129 
130 #if (HAVE_SSE2_EXTERNAL == 1)
131     if (EXTERNAL_SSE2(cpu_flags)) {
132         c->idct_put  = ff_vp3_idct_put_sse2;
133         c->idct_add  = ff_vp3_idct_add_sse2;
134     }
135 #endif
136 }
137