rtpdec: Clean up FMTP parsing code in H.264 RTP depacketizer
[libav.git] / libavformat / rtpdec_h264.c
1 /*
2 * RTP H264 Protocol (RFC3984)
3 * Copyright (c) 2006 Ryan Martell
4 *
5 * This file is part of FFmpeg.
6 *
7 * FFmpeg is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2.1 of the License, or (at your option) any later version.
11 *
12 * FFmpeg is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
16 *
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with FFmpeg; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20 */
21
22 /**
23 * @file
24 * @brief H.264 / RTP Code (RFC3984)
25 * @author Ryan Martell <rdm4@martellventures.com>
26 *
27 * @note Notes:
28 * Notes:
29 * This currently supports packetization mode:
30 * Single Nal Unit Mode (0), or
31 * Non-Interleaved Mode (1). It currently does not support
32 * Interleaved Mode (2). (This requires implementing STAP-B, MTAP16, MTAP24, FU-B packet types)
33 *
34 * @note TODO:
35 * 1) RTCP sender reports for udp streams are required..
36 *
37 */
38
39 #include "libavutil/base64.h"
40 #include "libavutil/avstring.h"
41 #include "libavcodec/get_bits.h"
42 #include "avformat.h"
43 #include "mpegts.h"
44
45 #include <unistd.h>
46 #include "network.h"
47 #include <assert.h>
48
49 #include "rtpdec.h"
50 #include "rtpdec_h264.h"
51
52 /**
53 RTP/H264 specific private data.
54 */
55 struct PayloadContext {
56 unsigned long cookie; ///< sanity check, to make sure we get the pointer we're expecting.
57
58 //sdp setup parameters
59 uint8_t profile_idc; ///< from the sdp setup parameters.
60 uint8_t profile_iop; ///< from the sdp setup parameters.
61 uint8_t level_idc; ///< from the sdp setup parameters.
62 int packetization_mode; ///< from the sdp setup parameters.
63 #ifdef DEBUG
64 int packet_types_received[32];
65 #endif
66 };
67
68 #define MAGIC_COOKIE (0xdeadbeef) ///< Cookie for the extradata; to verify we are what we think we are, and that we haven't been freed.
69 #define DEAD_COOKIE (0xdeaddead) ///< Cookie for the extradata; once it is freed.
70
71 /* ---------------- private code */
72 static int sdp_parse_fmtp_config_h264(AVStream * stream,
73 PayloadContext * h264_data,
74 char *attr, char *value)
75 {
76 AVCodecContext *codec = stream->codec;
77 assert(codec->codec_id == CODEC_ID_H264);
78 assert(h264_data != NULL);
79
80 if (!strcmp(attr, "packetization-mode")) {
81 av_log(codec, AV_LOG_DEBUG, "RTP Packetization Mode: %d\n", atoi(value));
82 h264_data->packetization_mode = atoi(value);
83 /*
84 Packetization Mode:
85 0 or not present: Single NAL mode (Only nals from 1-23 are allowed)
86 1: Non-interleaved Mode: 1-23, 24 (STAP-A), 28 (FU-A) are allowed.
87 2: Interleaved Mode: 25 (STAP-B), 26 (MTAP16), 27 (MTAP24), 28 (FU-A), and 29 (FU-B) are allowed.
88 */
89 if (h264_data->packetization_mode > 1)
90 av_log(codec, AV_LOG_ERROR,
91 "Interleaved RTP mode is not supported yet.");
92 } else if (!strcmp(attr, "profile-level-id")) {
93 if (strlen(value) == 6) {
94 char buffer[3];
95 // 6 characters=3 bytes, in hex.
96 uint8_t profile_idc;
97 uint8_t profile_iop;
98 uint8_t level_idc;
99
100 buffer[0] = value[0]; buffer[1] = value[1]; buffer[2] = '\0';
101 profile_idc = strtol(buffer, NULL, 16);
102 buffer[0] = value[2]; buffer[1] = value[3];
103 profile_iop = strtol(buffer, NULL, 16);
104 buffer[0] = value[4]; buffer[1] = value[5];
105 level_idc = strtol(buffer, NULL, 16);
106
107 // set the parameters...
108 av_log(codec, AV_LOG_DEBUG,
109 "RTP Profile IDC: %x Profile IOP: %x Level: %x\n",
110 profile_idc, profile_iop, level_idc);
111 h264_data->profile_idc = profile_idc;
112 h264_data->profile_iop = profile_iop;
113 h264_data->level_idc = level_idc;
114 }
115 } else if (!strcmp(attr, "sprop-parameter-sets")) {
116 uint8_t start_sequence[]= { 0, 0, 1 };
117 codec->extradata_size= 0;
118 codec->extradata= NULL;
119
120 while (*value) {
121 char base64packet[1024];
122 uint8_t decoded_packet[1024];
123 uint32_t packet_size;
124 char *dst = base64packet;
125
126 while (*value && *value != ','
127 && (dst - base64packet) < sizeof(base64packet) - 1) {
128 *dst++ = *value++;
129 }
130 *dst++ = '\0';
131
132 if (*value == ',')
133 value++;
134
135 packet_size= av_base64_decode(decoded_packet, base64packet, sizeof(decoded_packet));
136 if (packet_size) {
137 uint8_t *dest = av_malloc(packet_size + sizeof(start_sequence) +
138 codec->extradata_size +
139 FF_INPUT_BUFFER_PADDING_SIZE);
140 if(dest)
141 {
142 if(codec->extradata_size)
143 {
144 // av_realloc?
145 memcpy(dest, codec->extradata, codec->extradata_size);
146 av_free(codec->extradata);
147 }
148
149 memcpy(dest+codec->extradata_size, start_sequence, sizeof(start_sequence));
150 memcpy(dest+codec->extradata_size+sizeof(start_sequence), decoded_packet, packet_size);
151 memset(dest+codec->extradata_size+sizeof(start_sequence)+
152 packet_size, 0, FF_INPUT_BUFFER_PADDING_SIZE);
153
154 codec->extradata= dest;
155 codec->extradata_size+= sizeof(start_sequence)+packet_size;
156 } else {
157 av_log(codec, AV_LOG_ERROR, "Unable to allocate memory for extradata!");
158 return AVERROR(ENOMEM);
159 }
160 }
161 }
162 av_log(codec, AV_LOG_DEBUG, "Extradata set to %p (size: %d)!", codec->extradata, codec->extradata_size);
163 }
164 return 0;
165 }
166
167 // return 0 on packet, no more left, 1 on packet, 1 on partial packet...
168 static int h264_handle_packet(AVFormatContext *ctx,
169 PayloadContext *data,
170 AVStream *st,
171 AVPacket * pkt,
172 uint32_t * timestamp,
173 const uint8_t * buf,
174 int len, int flags)
175 {
176 uint8_t nal = buf[0];
177 uint8_t type = (nal & 0x1f);
178 int result= 0;
179 uint8_t start_sequence[]= {0, 0, 1};
180
181 #ifdef DEBUG
182 assert(data);
183 assert(data->cookie == MAGIC_COOKIE);
184 #endif
185 assert(buf);
186
187 if (type >= 1 && type <= 23)
188 type = 1; // simplify the case. (these are all the nal types used internally by the h264 codec)
189 switch (type) {
190 case 0: // undefined;
191 result= -1;
192 break;
193
194 case 1:
195 av_new_packet(pkt, len+sizeof(start_sequence));
196 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
197 memcpy(pkt->data+sizeof(start_sequence), buf, len);
198 #ifdef DEBUG
199 data->packet_types_received[nal & 0x1f]++;
200 #endif
201 break;
202
203 case 24: // STAP-A (one packet, multiple nals)
204 // consume the STAP-A NAL
205 buf++;
206 len--;
207 // first we are going to figure out the total size....
208 {
209 int pass= 0;
210 int total_length= 0;
211 uint8_t *dst= NULL;
212
213 for(pass= 0; pass<2; pass++) {
214 const uint8_t *src= buf;
215 int src_len= len;
216
217 do {
218 uint16_t nal_size = AV_RB16(src); // this going to be a problem if unaligned (can it be?)
219
220 // consume the length of the aggregate...
221 src += 2;
222 src_len -= 2;
223
224 if (nal_size <= src_len) {
225 if(pass==0) {
226 // counting...
227 total_length+= sizeof(start_sequence)+nal_size;
228 } else {
229 // copying
230 assert(dst);
231 memcpy(dst, start_sequence, sizeof(start_sequence));
232 dst+= sizeof(start_sequence);
233 memcpy(dst, src, nal_size);
234 #ifdef DEBUG
235 data->packet_types_received[*src & 0x1f]++;
236 #endif
237 dst+= nal_size;
238 }
239 } else {
240 av_log(ctx, AV_LOG_ERROR,
241 "nal size exceeds length: %d %d\n", nal_size, src_len);
242 }
243
244 // eat what we handled...
245 src += nal_size;
246 src_len -= nal_size;
247
248 if (src_len < 0)
249 av_log(ctx, AV_LOG_ERROR,
250 "Consumed more bytes than we got! (%d)\n", src_len);
251 } while (src_len > 2); // because there could be rtp padding..
252
253 if(pass==0) {
254 // now we know the total size of the packet (with the start sequences added)
255 av_new_packet(pkt, total_length);
256 dst= pkt->data;
257 } else {
258 assert(dst-pkt->data==total_length);
259 }
260 }
261 }
262 break;
263
264 case 25: // STAP-B
265 case 26: // MTAP-16
266 case 27: // MTAP-24
267 case 29: // FU-B
268 av_log(ctx, AV_LOG_ERROR,
269 "Unhandled type (%d) (See RFC for implementation details\n",
270 type);
271 result= -1;
272 break;
273
274 case 28: // FU-A (fragmented nal)
275 buf++;
276 len--; // skip the fu_indicator
277 {
278 // these are the same as above, we just redo them here for clarity...
279 uint8_t fu_indicator = nal;
280 uint8_t fu_header = *buf; // read the fu_header.
281 uint8_t start_bit = fu_header >> 7;
282 // uint8_t end_bit = (fu_header & 0x40) >> 6;
283 uint8_t nal_type = (fu_header & 0x1f);
284 uint8_t reconstructed_nal;
285
286 // reconstruct this packet's true nal; only the data follows..
287 reconstructed_nal = fu_indicator & (0xe0); // the original nal forbidden bit and NRI are stored in this packet's nal;
288 reconstructed_nal |= nal_type;
289
290 // skip the fu_header...
291 buf++;
292 len--;
293
294 #ifdef DEBUG
295 if (start_bit)
296 data->packet_types_received[nal_type]++;
297 #endif
298 if(start_bit) {
299 // copy in the start sequence, and the reconstructed nal....
300 av_new_packet(pkt, sizeof(start_sequence)+sizeof(nal)+len);
301 memcpy(pkt->data, start_sequence, sizeof(start_sequence));
302 pkt->data[sizeof(start_sequence)]= reconstructed_nal;
303 memcpy(pkt->data+sizeof(start_sequence)+sizeof(nal), buf, len);
304 } else {
305 av_new_packet(pkt, len);
306 memcpy(pkt->data, buf, len);
307 }
308 }
309 break;
310
311 case 30: // undefined
312 case 31: // undefined
313 default:
314 av_log(ctx, AV_LOG_ERROR, "Undefined type (%d)", type);
315 result= -1;
316 break;
317 }
318
319 pkt->stream_index = st->index;
320
321 return result;
322 }
323
324 /* ---------------- public code */
325 static PayloadContext *h264_new_context(void)
326 {
327 PayloadContext *data =
328 av_mallocz(sizeof(PayloadContext) +
329 FF_INPUT_BUFFER_PADDING_SIZE);
330
331 if (data) {
332 data->cookie = MAGIC_COOKIE;
333 }
334
335 return data;
336 }
337
338 static void h264_free_context(PayloadContext *data)
339 {
340 #ifdef DEBUG
341 int ii;
342
343 for (ii = 0; ii < 32; ii++) {
344 if (data->packet_types_received[ii])
345 av_log(NULL, AV_LOG_DEBUG, "Received %d packets of type %d\n",
346 data->packet_types_received[ii], ii);
347 }
348 #endif
349
350 assert(data);
351 assert(data->cookie == MAGIC_COOKIE);
352
353 // avoid stale pointers (assert)
354 data->cookie = DEAD_COOKIE;
355
356 // and clear out this...
357 av_free(data);
358 }
359
360 static int parse_h264_sdp_line(AVFormatContext *s, int st_index,
361 PayloadContext *h264_data, const char *line)
362 {
363 AVStream *stream = s->streams[st_index];
364 AVCodecContext *codec = stream->codec;
365 const char *p = line;
366
367 assert(h264_data->cookie == MAGIC_COOKIE);
368
369 if (av_strstart(p, "framesize:", &p)) {
370 char buf1[50];
371 char *dst = buf1;
372
373 // remove the protocol identifier..
374 while (*p && *p == ' ') p++; // strip spaces.
375 while (*p && *p != ' ') p++; // eat protocol identifier
376 while (*p && *p == ' ') p++; // strip trailing spaces.
377 while (*p && *p != '-' && (dst - buf1) < sizeof(buf1) - 1) {
378 *dst++ = *p++;
379 }
380 *dst = '\0';
381
382 // a='framesize:96 320-240'
383 // set our parameters..
384 codec->width = atoi(buf1);
385 codec->height = atoi(p + 1); // skip the -
386 codec->pix_fmt = PIX_FMT_YUV420P;
387 } else if (av_strstart(p, "fmtp:", &p)) {
388 return ff_parse_fmtp(stream, h264_data, p, sdp_parse_fmtp_config_h264);
389 } else if (av_strstart(p, "cliprect:", &p)) {
390 // could use this if we wanted.
391 }
392
393 return 0; // keep processing it the normal way...
394 }
395
396 /**
397 This is the structure for expanding on the dynamic rtp protocols (makes everything static. yay!)
398 */
399 RTPDynamicProtocolHandler ff_h264_dynamic_handler = {
400 .enc_name = "H264",
401 .codec_type = AVMEDIA_TYPE_VIDEO,
402 .codec_id = CODEC_ID_H264,
403 .parse_sdp_a_line = parse_h264_sdp_line,
404 .open = h264_new_context,
405 .close = h264_free_context,
406 .parse_packet = h264_handle_packet
407 };