1 /*
2 * V210 decoder
3 *
4 * Copyright (C) 2009 Michael Niedermayer <michaelni@gmx.at>
5 * Copyright (c) 2009 Baptiste Coudurier <baptiste dot coudurier at gmail dot com>
6 *
7 * This file is part of FFmpeg.
8 *
9 * FFmpeg is free software; you can redistribute it and/or
10 * modify it under the terms of the GNU Lesser General Public
11 * License as published by the Free Software Foundation; either
12 * version 2.1 of the License, or (at your option) any later version.
13 *
14 * FFmpeg is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
17 * Lesser General Public License for more details.
18 *
19 * You should have received a copy of the GNU Lesser General Public
20 * License along with FFmpeg; if not, write to the Free Software
21 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
22 */
23
24 #include "avcodec.h"
25 #include "internal.h"
26 #include "v210dec.h"
27 #include "libavutil/bswap.h"
28 #include "libavutil/internal.h"
29 #include "libavutil/mem.h"
30 #include "libavutil/intreadwrite.h"
31 #include "thread.h"
32
33 #define READ_PIXELS(a, b, c) \
34 do { \
35 val = av_le2ne32(*src++); \
36 *a++ = val & 0x3FF; \
37 *b++ = (val >> 10) & 0x3FF; \
38 *c++ = (val >> 20) & 0x3FF; \
39 } while (0)
40
41 typedef struct ThreadData {
42 AVFrame *frame;
43 uint8_t *buf;
44 int stride;
45 } ThreadData;
46
v210_planar_unpack_c(const uint32_t * src,uint16_t * y,uint16_t * u,uint16_t * v,int width)47 static void v210_planar_unpack_c(const uint32_t *src, uint16_t *y, uint16_t *u, uint16_t *v, int width)
48 {
49 uint32_t val;
50 int i;
51
52 for( i = 0; i < width-5; i += 6 ){
53 READ_PIXELS(u, y, v);
54 READ_PIXELS(y, u, y);
55 READ_PIXELS(v, y, u);
56 READ_PIXELS(y, v, y);
57 }
58 }
59
ff_v210dec_init(V210DecContext * s)60 av_cold void ff_v210dec_init(V210DecContext *s)
61 {
62 s->unpack_frame = v210_planar_unpack_c;
63 if (ARCH_X86)
64 ff_v210_x86_init(s);
65 }
66
decode_init(AVCodecContext * avctx)67 static av_cold int decode_init(AVCodecContext *avctx)
68 {
69 V210DecContext *s = avctx->priv_data;
70
71 avctx->pix_fmt = AV_PIX_FMT_YUV422P10;
72 avctx->bits_per_raw_sample = 10;
73
74 s->thread_count = av_clip(avctx->thread_count, 1, avctx->height/4);
75 s->aligned_input = 0;
76 ff_v210dec_init(s);
77
78 return 0;
79 }
80
v210_decode_slice(AVCodecContext * avctx,void * arg,int jobnr,int threadnr)81 static int v210_decode_slice(AVCodecContext *avctx, void *arg, int jobnr, int threadnr)
82 {
83 V210DecContext *s = avctx->priv_data;
84 int h, w;
85 ThreadData *td = arg;
86 AVFrame *frame = td->frame;
87 int stride = td->stride;
88 int slice_start = (avctx->height * jobnr) / s->thread_count;
89 int slice_end = (avctx->height * (jobnr+1)) / s->thread_count;
90 uint8_t *psrc = td->buf + stride * slice_start;
91 uint16_t *y, *u, *v;
92
93 y = (uint16_t*)frame->data[0] + slice_start * frame->linesize[0] / 2;
94 u = (uint16_t*)frame->data[1] + slice_start * frame->linesize[1] / 2;
95 v = (uint16_t*)frame->data[2] + slice_start * frame->linesize[2] / 2;
96 for (h = slice_start; h < slice_end; h++) {
97 const uint32_t *src = (const uint32_t*)psrc;
98 uint32_t val;
99
100 w = (avctx->width / 12) * 12;
101 s->unpack_frame(src, y, u, v, w);
102
103 y += w;
104 u += w >> 1;
105 v += w >> 1;
106 src += (w << 1) / 3;
107
108 if (w < avctx->width - 5) {
109 READ_PIXELS(u, y, v);
110 READ_PIXELS(y, u, y);
111 READ_PIXELS(v, y, u);
112 READ_PIXELS(y, v, y);
113 w += 6;
114 }
115
116 if (w < avctx->width - 1) {
117 READ_PIXELS(u, y, v);
118
119 val = av_le2ne32(*src++);
120 *y++ = val & 0x3FF;
121 if (w < avctx->width - 3) {
122 *u++ = (val >> 10) & 0x3FF;
123 *y++ = (val >> 20) & 0x3FF;
124
125 val = av_le2ne32(*src++);
126 *v++ = val & 0x3FF;
127 *y++ = (val >> 10) & 0x3FF;
128 }
129 }
130
131 psrc += stride;
132 y += frame->linesize[0] / 2 - avctx->width + (avctx->width & 1);
133 u += frame->linesize[1] / 2 - avctx->width / 2;
134 v += frame->linesize[2] / 2 - avctx->width / 2;
135 }
136
137 return 0;
138 }
139
decode_frame(AVCodecContext * avctx,void * data,int * got_frame,AVPacket * avpkt)140 static int decode_frame(AVCodecContext *avctx, void *data, int *got_frame,
141 AVPacket *avpkt)
142 {
143 V210DecContext *s = avctx->priv_data;
144 ThreadData td;
145 int ret, stride, aligned_input;
146 ThreadFrame frame = { .f = data };
147 AVFrame *pic = data;
148 const uint8_t *psrc = avpkt->data;
149
150 if (s->custom_stride )
151 stride = s->custom_stride;
152 else {
153 int aligned_width = ((avctx->width + 47) / 48) * 48;
154 stride = aligned_width * 8 / 3;
155 }
156
157 if (avpkt->size < stride * avctx->height) {
158 if ((((avctx->width + 23) / 24) * 24 * 8) / 3 * avctx->height == avpkt->size) {
159 stride = avpkt->size / avctx->height;
160 if (!s->stride_warning_shown)
161 av_log(avctx, AV_LOG_WARNING, "Broken v210 with too small padding (64 byte) detected\n");
162 s->stride_warning_shown = 1;
163 } else {
164 av_log(avctx, AV_LOG_ERROR, "packet too small\n");
165 return AVERROR_INVALIDDATA;
166 }
167 }
168 td.stride = stride;
169 if ( avctx->codec_tag == MKTAG('C', '2', '1', '0')
170 && avpkt->size > 64
171 && AV_RN32(psrc) == AV_RN32("INFO")
172 && avpkt->size - 64 >= stride * avctx->height)
173 psrc += 64;
174
175 aligned_input = !((uintptr_t)psrc & 0x1f) && !(stride & 0x1f);
176 if (aligned_input != s->aligned_input) {
177 s->aligned_input = aligned_input;
178 ff_v210dec_init(s);
179 }
180
181 if ((ret = ff_thread_get_buffer(avctx, &frame, 0)) < 0)
182 return ret;
183
184 pic->pict_type = AV_PICTURE_TYPE_I;
185 pic->key_frame = 1;
186
187 td.buf = (uint8_t*)psrc;
188 td.frame = pic;
189 avctx->execute2(avctx, v210_decode_slice, &td, NULL, s->thread_count);
190
191 if (avctx->field_order > AV_FIELD_PROGRESSIVE) {
192 /* we have interlaced material flagged in container */
193 pic->interlaced_frame = 1;
194 if (avctx->field_order == AV_FIELD_TT || avctx->field_order == AV_FIELD_TB)
195 pic->top_field_first = 1;
196 }
197
198 *got_frame = 1;
199
200 return avpkt->size;
201 }
202
203 #define V210DEC_FLAGS AV_OPT_FLAG_DECODING_PARAM | AV_OPT_FLAG_VIDEO_PARAM
204 static const AVOption v210dec_options[] = {
205 {"custom_stride", "Custom V210 stride", offsetof(V210DecContext, custom_stride), AV_OPT_TYPE_INT,
206 {.i64 = 0}, INT_MIN, INT_MAX, V210DEC_FLAGS},
207 {NULL}
208 };
209
210 static const AVClass v210dec_class = {
211 .class_name = "V210 Decoder",
212 .item_name = av_default_item_name,
213 .option = v210dec_options,
214 .version = LIBAVUTIL_VERSION_INT,
215 };
216
217 AVCodec ff_v210_decoder = {
218 .name = "v210",
219 .long_name = NULL_IF_CONFIG_SMALL("Uncompressed 4:2:2 10-bit"),
220 .type = AVMEDIA_TYPE_VIDEO,
221 .id = AV_CODEC_ID_V210,
222 .priv_data_size = sizeof(V210DecContext),
223 .init = decode_init,
224 .decode = decode_frame,
225 .capabilities = AV_CODEC_CAP_DR1 |
226 AV_CODEC_CAP_SLICE_THREADS |
227 AV_CODEC_CAP_FRAME_THREADS,
228 .priv_class = &v210dec_class,
229 };
230