| 1 | /* |
| 2 | * Xiph RTP Protocols |
| 3 | * Copyright (c) 2009 Colin McQuillian |
| 4 | * Copyright (c) 2010 Josh Allmann |
| 5 | * |
| 6 | * This file is part of FFmpeg. |
| 7 | * |
| 8 | * FFmpeg is free software; you can redistribute it and/or |
| 9 | * modify it under the terms of the GNU Lesser General Public |
| 10 | * License as published by the Free Software Foundation; either |
| 11 | * version 2.1 of the License, or (at your option) any later version. |
| 12 | * |
| 13 | * FFmpeg is distributed in the hope that it will be useful, |
| 14 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 15 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 16 | * Lesser General Public License for more details. |
| 17 | * |
| 18 | * You should have received a copy of the GNU Lesser General Public |
| 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 |
| 21 | */ |
| 22 | |
| 23 | /** |
| 24 | * @file |
| 25 | * @brief Xiph / RTP Code |
| 26 | * @author Colin McQuillan <m.niloc@gmail.com> |
| 27 | * @author Josh Allmann <joshua.allmann@gmail.com> |
| 28 | */ |
| 29 | |
| 30 | #include "libavutil/attributes.h" |
| 31 | #include "libavutil/avassert.h" |
| 32 | #include "libavutil/avstring.h" |
| 33 | #include "libavutil/base64.h" |
| 34 | #include "libavcodec/bytestream.h" |
| 35 | |
| 36 | #include "internal.h" |
| 37 | #include "rtpdec.h" |
| 38 | #include "rtpdec_formats.h" |
| 39 | |
| 40 | /** |
| 41 | * RTP/Xiph specific private data. |
| 42 | */ |
| 43 | struct PayloadContext { |
| 44 | unsigned ident; ///< 24-bit stream configuration identifier |
| 45 | uint32_t timestamp; |
| 46 | AVIOContext* fragment; ///< buffer for split payloads |
| 47 | uint8_t *split_buf; |
| 48 | int split_pos, split_buf_len, split_buf_size; |
| 49 | int split_pkts; |
| 50 | }; |
| 51 | |
| 52 | static PayloadContext *xiph_new_context(void) |
| 53 | { |
| 54 | return av_mallocz(sizeof(PayloadContext)); |
| 55 | } |
| 56 | |
| 57 | static inline void free_fragment_if_needed(PayloadContext * data) |
| 58 | { |
| 59 | if (data->fragment) { |
| 60 | uint8_t* p; |
| 61 | avio_close_dyn_buf(data->fragment, &p); |
| 62 | av_free(p); |
| 63 | data->fragment = NULL; |
| 64 | } |
| 65 | } |
| 66 | |
| 67 | static void xiph_free_context(PayloadContext * data) |
| 68 | { |
| 69 | free_fragment_if_needed(data); |
| 70 | av_free(data->split_buf); |
| 71 | av_free(data); |
| 72 | } |
| 73 | |
| 74 | static av_cold int xiph_vorbis_init(AVFormatContext *ctx, int st_index, |
| 75 | PayloadContext *data) |
| 76 | { |
| 77 | if (st_index < 0) |
| 78 | return 0; |
| 79 | ctx->streams[st_index]->need_parsing = AVSTREAM_PARSE_HEADERS; |
| 80 | return 0; |
| 81 | } |
| 82 | |
| 83 | |
| 84 | static int xiph_handle_packet(AVFormatContext *ctx, PayloadContext *data, |
| 85 | AVStream *st, AVPacket *pkt, uint32_t *timestamp, |
| 86 | const uint8_t *buf, int len, uint16_t seq, |
| 87 | int flags) |
| 88 | { |
| 89 | |
| 90 | int ident, fragmented, tdt, num_pkts, pkt_len; |
| 91 | |
| 92 | if (!buf) { |
| 93 | if (!data->split_buf || data->split_pos + 2 > data->split_buf_len || |
| 94 | data->split_pkts <= 0) { |
| 95 | av_log(ctx, AV_LOG_ERROR, "No more data to return\n"); |
| 96 | return AVERROR_INVALIDDATA; |
| 97 | } |
| 98 | pkt_len = AV_RB16(data->split_buf + data->split_pos); |
| 99 | data->split_pos += 2; |
| 100 | if (data->split_pos + pkt_len > data->split_buf_len) { |
| 101 | av_log(ctx, AV_LOG_ERROR, "Not enough data to return\n"); |
| 102 | return AVERROR_INVALIDDATA; |
| 103 | } |
| 104 | if (av_new_packet(pkt, pkt_len)) { |
| 105 | av_log(ctx, AV_LOG_ERROR, "Out of memory.\n"); |
| 106 | return AVERROR(ENOMEM); |
| 107 | } |
| 108 | pkt->stream_index = st->index; |
| 109 | memcpy(pkt->data, data->split_buf + data->split_pos, pkt_len); |
| 110 | data->split_pos += pkt_len; |
| 111 | data->split_pkts--; |
| 112 | return data->split_pkts > 0; |
| 113 | } |
| 114 | |
| 115 | if (len < 6) { |
| 116 | av_log(ctx, AV_LOG_ERROR, "Invalid %d byte packet\n", len); |
| 117 | return AVERROR_INVALIDDATA; |
| 118 | } |
| 119 | |
| 120 | // read xiph rtp headers |
| 121 | ident = AV_RB24(buf); |
| 122 | fragmented = buf[3] >> 6; |
| 123 | tdt = (buf[3] >> 4) & 3; |
| 124 | num_pkts = buf[3] & 0xf; |
| 125 | pkt_len = AV_RB16(buf + 4); |
| 126 | |
| 127 | if (pkt_len > len - 6) { |
| 128 | av_log(ctx, AV_LOG_ERROR, |
| 129 | "Invalid packet length %d in %d byte packet\n", pkt_len, |
| 130 | len); |
| 131 | return AVERROR_INVALIDDATA; |
| 132 | } |
| 133 | |
| 134 | if (ident != data->ident) { |
| 135 | av_log(ctx, AV_LOG_ERROR, |
| 136 | "Unimplemented Xiph SDP configuration change detected\n"); |
| 137 | return AVERROR_PATCHWELCOME; |
| 138 | } |
| 139 | |
| 140 | if (tdt) { |
| 141 | av_log(ctx, AV_LOG_ERROR, |
| 142 | "Unimplemented RTP Xiph packet settings (%d,%d,%d)\n", |
| 143 | fragmented, tdt, num_pkts); |
| 144 | return AVERROR_PATCHWELCOME; |
| 145 | } |
| 146 | |
| 147 | buf += 6; // move past header bits |
| 148 | len -= 6; |
| 149 | |
| 150 | if (fragmented == 0) { |
| 151 | if (av_new_packet(pkt, pkt_len)) { |
| 152 | av_log(ctx, AV_LOG_ERROR, "Out of memory.\n"); |
| 153 | return AVERROR(ENOMEM); |
| 154 | } |
| 155 | pkt->stream_index = st->index; |
| 156 | memcpy(pkt->data, buf, pkt_len); |
| 157 | buf += pkt_len; |
| 158 | len -= pkt_len; |
| 159 | num_pkts--; |
| 160 | |
| 161 | if (num_pkts > 0) { |
| 162 | if (len > data->split_buf_size || !data->split_buf) { |
| 163 | av_freep(&data->split_buf); |
| 164 | data->split_buf_size = 2 * len; |
| 165 | data->split_buf = av_malloc(data->split_buf_size); |
| 166 | if (!data->split_buf) { |
| 167 | av_log(ctx, AV_LOG_ERROR, "Out of memory.\n"); |
| 168 | av_free_packet(pkt); |
| 169 | return AVERROR(ENOMEM); |
| 170 | } |
| 171 | } |
| 172 | memcpy(data->split_buf, buf, len); |
| 173 | data->split_buf_len = len; |
| 174 | data->split_pos = 0; |
| 175 | data->split_pkts = num_pkts; |
| 176 | return 1; |
| 177 | } |
| 178 | |
| 179 | return 0; |
| 180 | |
| 181 | } else if (fragmented == 1) { |
| 182 | // start of xiph data fragment |
| 183 | int res; |
| 184 | |
| 185 | // end packet has been lost somewhere, so drop buffered data |
| 186 | free_fragment_if_needed(data); |
| 187 | |
| 188 | if((res = avio_open_dyn_buf(&data->fragment)) < 0) |
| 189 | return res; |
| 190 | |
| 191 | avio_write(data->fragment, buf, pkt_len); |
| 192 | data->timestamp = *timestamp; |
| 193 | |
| 194 | } else { |
| 195 | av_assert1(fragmented < 4); |
| 196 | if (data->timestamp != *timestamp) { |
| 197 | // skip if fragmented timestamp is incorrect; |
| 198 | // a start packet has been lost somewhere |
| 199 | free_fragment_if_needed(data); |
| 200 | av_log(ctx, AV_LOG_ERROR, "RTP timestamps don't match!\n"); |
| 201 | return AVERROR_INVALIDDATA; |
| 202 | } |
| 203 | if (!data->fragment) { |
| 204 | av_log(ctx, AV_LOG_WARNING, |
| 205 | "Received packet without a start fragment; dropping.\n"); |
| 206 | return AVERROR(EAGAIN); |
| 207 | } |
| 208 | |
| 209 | // copy data to fragment buffer |
| 210 | avio_write(data->fragment, buf, pkt_len); |
| 211 | |
| 212 | if (fragmented == 3) { |
| 213 | // end of xiph data packet |
| 214 | int ret = ff_rtp_finalize_packet(pkt, &data->fragment, st->index); |
| 215 | if (ret < 0) { |
| 216 | av_log(ctx, AV_LOG_ERROR, |
| 217 | "Error occurred when getting fragment buffer."); |
| 218 | return ret; |
| 219 | } |
| 220 | |
| 221 | return 0; |
| 222 | } |
| 223 | } |
| 224 | |
| 225 | return AVERROR(EAGAIN); |
| 226 | } |
| 227 | |
| 228 | /** |
| 229 | * Length encoding described in RFC5215 section 3.1.1. |
| 230 | */ |
| 231 | static int get_base128(const uint8_t ** buf, const uint8_t * buf_end) |
| 232 | { |
| 233 | int n = 0; |
| 234 | for (; *buf < buf_end; ++*buf) { |
| 235 | n <<= 7; |
| 236 | n += **buf & 0x7f; |
| 237 | if (!(**buf & 0x80)) { |
| 238 | ++*buf; |
| 239 | return n; |
| 240 | } |
| 241 | } |
| 242 | return 0; |
| 243 | } |
| 244 | |
| 245 | /** |
| 246 | * Based off parse_packed_headers in Vorbis RTP |
| 247 | */ |
| 248 | static int |
| 249 | parse_packed_headers(const uint8_t * packed_headers, |
| 250 | const uint8_t * packed_headers_end, |
| 251 | AVCodecContext * codec, PayloadContext * xiph_data) |
| 252 | { |
| 253 | |
| 254 | unsigned num_packed, num_headers, length, length1, length2, extradata_alloc; |
| 255 | uint8_t *ptr; |
| 256 | |
| 257 | if (packed_headers_end - packed_headers < 9) { |
| 258 | av_log(codec, AV_LOG_ERROR, |
| 259 | "Invalid %"PTRDIFF_SPECIFIER" byte packed header.", |
| 260 | packed_headers_end - packed_headers); |
| 261 | return AVERROR_INVALIDDATA; |
| 262 | } |
| 263 | |
| 264 | num_packed = bytestream_get_be32(&packed_headers); |
| 265 | xiph_data->ident = bytestream_get_be24(&packed_headers); |
| 266 | length = bytestream_get_be16(&packed_headers); |
| 267 | num_headers = get_base128(&packed_headers, packed_headers_end); |
| 268 | length1 = get_base128(&packed_headers, packed_headers_end); |
| 269 | length2 = get_base128(&packed_headers, packed_headers_end); |
| 270 | |
| 271 | if (num_packed != 1 || num_headers > 3) { |
| 272 | av_log(codec, AV_LOG_ERROR, |
| 273 | "Unimplemented number of headers: %d packed headers, %d headers\n", |
| 274 | num_packed, num_headers); |
| 275 | return AVERROR_PATCHWELCOME; |
| 276 | } |
| 277 | |
| 278 | if (packed_headers_end - packed_headers != length || |
| 279 | length1 > length || length2 > length - length1) { |
| 280 | av_log(codec, AV_LOG_ERROR, |
| 281 | "Bad packed header lengths (%d,%d,%"PTRDIFF_SPECIFIER",%d)\n", length1, |
| 282 | length2, packed_headers_end - packed_headers, length); |
| 283 | return AVERROR_INVALIDDATA; |
| 284 | } |
| 285 | |
| 286 | /* allocate extra space: |
| 287 | * -- length/255 +2 for xiphlacing |
| 288 | * -- one for the '2' marker |
| 289 | * -- FF_INPUT_BUFFER_PADDING_SIZE required */ |
| 290 | extradata_alloc = length + length/255 + 3 + FF_INPUT_BUFFER_PADDING_SIZE; |
| 291 | |
| 292 | if (ff_alloc_extradata(codec, extradata_alloc)) { |
| 293 | av_log(codec, AV_LOG_ERROR, "Out of memory\n"); |
| 294 | return AVERROR(ENOMEM); |
| 295 | } |
| 296 | ptr = codec->extradata; |
| 297 | *ptr++ = 2; |
| 298 | ptr += av_xiphlacing(ptr, length1); |
| 299 | ptr += av_xiphlacing(ptr, length2); |
| 300 | memcpy(ptr, packed_headers, length); |
| 301 | ptr += length; |
| 302 | codec->extradata_size = ptr - codec->extradata; |
| 303 | // clear out remaining parts of the buffer |
| 304 | memset(ptr, 0, extradata_alloc - codec->extradata_size); |
| 305 | |
| 306 | return 0; |
| 307 | } |
| 308 | |
| 309 | static int xiph_parse_fmtp_pair(AVFormatContext *s, |
| 310 | AVStream* stream, |
| 311 | PayloadContext *xiph_data, |
| 312 | char *attr, char *value) |
| 313 | { |
| 314 | AVCodecContext *codec = stream->codec; |
| 315 | int result = 0; |
| 316 | |
| 317 | if (!strcmp(attr, "sampling")) { |
| 318 | if (!strcmp(value, "YCbCr-4:2:0")) { |
| 319 | codec->pix_fmt = AV_PIX_FMT_YUV420P; |
| 320 | } else if (!strcmp(value, "YCbCr-4:4:2")) { |
| 321 | codec->pix_fmt = AV_PIX_FMT_YUV422P; |
| 322 | } else if (!strcmp(value, "YCbCr-4:4:4")) { |
| 323 | codec->pix_fmt = AV_PIX_FMT_YUV444P; |
| 324 | } else { |
| 325 | av_log(s, AV_LOG_ERROR, |
| 326 | "Unsupported pixel format %s\n", attr); |
| 327 | return AVERROR_INVALIDDATA; |
| 328 | } |
| 329 | } else if (!strcmp(attr, "width")) { |
| 330 | /* This is an integer between 1 and 1048561 |
| 331 | * and MUST be in multiples of 16. */ |
| 332 | codec->width = atoi(value); |
| 333 | return 0; |
| 334 | } else if (!strcmp(attr, "height")) { |
| 335 | /* This is an integer between 1 and 1048561 |
| 336 | * and MUST be in multiples of 16. */ |
| 337 | codec->height = atoi(value); |
| 338 | return 0; |
| 339 | } else if (!strcmp(attr, "delivery-method")) { |
| 340 | /* Possible values are: inline, in_band, out_band/specific_name. */ |
| 341 | return AVERROR_PATCHWELCOME; |
| 342 | } else if (!strcmp(attr, "configuration-uri")) { |
| 343 | /* NOTE: configuration-uri is supported only under 2 conditions: |
| 344 | *--after the delivery-method tag |
| 345 | * --with a delivery-method value of out_band */ |
| 346 | return AVERROR_PATCHWELCOME; |
| 347 | } else if (!strcmp(attr, "configuration")) { |
| 348 | /* NOTE: configuration is supported only AFTER the delivery-method tag |
| 349 | * The configuration value is a base64 encoded packed header */ |
| 350 | uint8_t *decoded_packet = NULL; |
| 351 | int packet_size; |
| 352 | size_t decoded_alloc = strlen(value) / 4 * 3 + 4; |
| 353 | |
| 354 | if (decoded_alloc <= INT_MAX) { |
| 355 | decoded_packet = av_malloc(decoded_alloc); |
| 356 | if (decoded_packet) { |
| 357 | packet_size = |
| 358 | av_base64_decode(decoded_packet, value, decoded_alloc); |
| 359 | |
| 360 | result = parse_packed_headers |
| 361 | (decoded_packet, decoded_packet + packet_size, codec, |
| 362 | xiph_data); |
| 363 | } else { |
| 364 | av_log(s, AV_LOG_ERROR, |
| 365 | "Out of memory while decoding SDP configuration.\n"); |
| 366 | result = AVERROR(ENOMEM); |
| 367 | } |
| 368 | } else { |
| 369 | av_log(s, AV_LOG_ERROR, "Packet too large\n"); |
| 370 | result = AVERROR_INVALIDDATA; |
| 371 | } |
| 372 | av_free(decoded_packet); |
| 373 | } |
| 374 | return result; |
| 375 | } |
| 376 | |
| 377 | static int xiph_parse_sdp_line(AVFormatContext *s, int st_index, |
| 378 | PayloadContext *data, const char *line) |
| 379 | { |
| 380 | const char *p; |
| 381 | |
| 382 | if (st_index < 0) |
| 383 | return 0; |
| 384 | |
| 385 | if (av_strstart(line, "fmtp:", &p)) { |
| 386 | return ff_parse_fmtp(s, s->streams[st_index], data, p, |
| 387 | xiph_parse_fmtp_pair); |
| 388 | } |
| 389 | |
| 390 | return 0; |
| 391 | } |
| 392 | |
| 393 | RTPDynamicProtocolHandler ff_theora_dynamic_handler = { |
| 394 | .enc_name = "theora", |
| 395 | .codec_type = AVMEDIA_TYPE_VIDEO, |
| 396 | .codec_id = AV_CODEC_ID_THEORA, |
| 397 | .parse_sdp_a_line = xiph_parse_sdp_line, |
| 398 | .alloc = xiph_new_context, |
| 399 | .free = xiph_free_context, |
| 400 | .parse_packet = xiph_handle_packet |
| 401 | }; |
| 402 | |
| 403 | RTPDynamicProtocolHandler ff_vorbis_dynamic_handler = { |
| 404 | .enc_name = "vorbis", |
| 405 | .codec_type = AVMEDIA_TYPE_AUDIO, |
| 406 | .codec_id = AV_CODEC_ID_VORBIS, |
| 407 | .init = xiph_vorbis_init, |
| 408 | .parse_sdp_a_line = xiph_parse_sdp_line, |
| 409 | .alloc = xiph_new_context, |
| 410 | .free = xiph_free_context, |
| 411 | .parse_packet = xiph_handle_packet |
| 412 | }; |