REBASE_PICTURE (more specifically, this half of it) takes a Picture
pointer that points into one larger struct, finds the offset of
that Picture within the struct and finds the corresponding field
within another instance of a similar struct.
The pointer difference "pic - (Picture*)old_ctx" is a value given
in sizeof(Picture) units, and when applied back on
(Picture*)new_ctx gets multiplied back with sizeof(Picture). Many
compilers seem to optimize out this division/multiplication, but
not all do.
GCC 4.2 on OS X doesn't seem to remove the division/multiplication,
therefore the new pointer didn't turn out to point to exactly
the right place in the new struct since it only had sizeof(Picture)
granularity (and the Picture is not aligned on a sizeof(Picture)
boundary within the encompassing struct). This bug has been present
before 47318953d as well - with H264, pointers to h->ref_list[0][0]
pointed to 88 bytes before h->ref_list[0][0] after the rebase. After
shrinking Picture, the difference ended up even larger, making
writes via such a Picture pointer overwrite other fields at random
in H264Context, ending up in crashes later.
This fixes H264 multithreaded decoding on OS X with GCC 4.2.
Signed-off-by: Martin Storsjö <martin@martin.st>
This makes the plain-armv6 version use the same registers as the
armv6t2 version above.
This fixes fate-vp8 on plain-armv6 devices.
Signed-off-by: Martin Storsjö <martin@martin.st>
In ff_rtp_get_payload_type, the AVFormatContext is used for checking
whether the payload_type or rtpflags options are set. In rtpenc_chain,
the rtpctx struct is a newly initialized struct where no options have
been set yet, so no options can be fetched from there.
All muxers that internally chain rtp muxers have the "rtpflags" field
that allows passing such options on (which is how this worked before
8034130e06), so this works just as intended.
This makes it possible to produce H263 in RFC2190 format with chained
RTP muxers.
CC: libav-stable@libav.org
Signed-off-by: Martin Storsjö <martin@martin.st>
Not sure if this actually happens, but we do the same check when
checking payload_type further above in the function, so it might
be needed.
Signed-off-by: Martin Storsjö <martin@martin.st>
This fixes encoding where the idct setting originally was set to
FF_IDCT_AUTO and dsputil chose a default idct with a non-null
permutation - even if the permutation tables were updated,
dct_quantize in x86/mpegvideoenc_template.c also checked the
value of this type variable.
Signed-off-by: Martin Storsjö <martin@martin.st>
This makes sure that the restrict keyword is mapped to whatever
keyword the compiler prefers/supports. This fixes building on MSVC
(and possibly on GCC 2.x as well).
Signed-off-by: Martin Storsjö <martin@martin.st>
Expose the current sequence number via an AVOption - this can
be used both for setting the initial sequence number, or for
querying the current number.
Signed-off-by: Martin Storsjö <martin@martin.st>
The theoretical minimum for a (not totally well formed) RTCP packet
is 8 bytes, so we shouldn't require 12 bytes as minimum input.
Also return AVERROR_INVALIDDATA instead of 0 if something that is
not a proper packet is given.
Signed-off-by: Martin Storsjö <martin@martin.st>
This clarifies where the limit number comes from, and only
requires exactly as much padding space as will be needed.
Signed-off-by: Martin Storsjö <martin@martin.st>
The main difference to the existing suites from RFC 4568 is
that the version with a 32 bit HMAC still uses 80 bit HMAC
for RTCP packets.
Signed-off-by: Martin Storsjö <martin@martin.st>
The lavf-internal parser functions are used when receiving
mpegts over RTP. This fixes memory leaks in this setup.
The normal mpegts demuxer close function was updated in ec7d0d2e in
2004 to fix leaks, but the parsing function used for RTP wasn't
updated and has been leaking ever since.
Signed-off-by: Martin Storsjö <martin@martin.st>
This makes the returned data valid to stream copy into other
containers as well, not only for decoding straight away.
Signed-off-by: Martin Storsjö <martin@martin.st>
This fixes build failures on ppc machines with a compiler that
supports -Werror=implicit-function-declaration.
Signed-off-by: Martin Storsjö <martin@martin.st>
This provides a fallback when building with Yasm enabled, but neither
inline assembly, nor the _mm_empty intrinsic are available or enabled.
Signed-off-by: Diego Biurrun <diego@biurrun.de>
This fixes a regression since d9cf5f51 with theora over RTP
(possibly with other variants of theora as well).
In theora over RTP, the second of the 3 headers turns out to be
0 bytes long, which prior to d9cf5f51 worked just fine. After
d9cf5f51, reading from the bitstream reader fails (since the reader
wasn't initialized but returned an error if initialized with 0 bits).
CC: libav-stable@libav.org
Signed-off-by: Martin Storsjö <martin@martin.st>
This is mostly useful for encryption together with the RTP muxer,
but could also be set up as IO towards the peer with the SDP
demuxer with custom IO.
Signed-off-by: Martin Storsjö <martin@martin.st>
This only takes care of decrypting incoming packets; the outgoing
RTCP packets are not encrypted. This is enough for some use cases,
and signalling crypto keys for use with outgoing RTCP packets
doesn't fit as simply into the API. If the SDP demuxer is hooked
up with custom IO, the return packets can be encrypted e.g. via the
SRTP protocol.
If the SRTP keys aren't available within the SDP, the decryption
can be handled externally as well (when using custom IO).
Signed-off-by: Martin Storsjö <martin@martin.st>
This supports the AES_CM_128_HMAC_SHA1_80 and
AES_CM_128_HMAC_SHA1_32 cipher suites (from RFC 4568) at the
moment. The main missing features are replay protection (which can be
added later without changing the internal API), and the F8 and null
ciphers.
Signed-off-by: Martin Storsjö <martin@martin.st>
This makes the behaviour defined when they wrap around. The value
assigned to expected_prior was a uint32_t already.
Signed-off-by: Martin Storsjö <martin@martin.st>
Without this, we'd signal a huge loss rate (due to unsigned
wraparound) if we had received one packet more than expected (that
is, one seq number sent twice). The code has a check for lost_interval
<= 0, but that doesn't do what was intended as long as the variable is
unsigned.
Signed-off-by: Martin Storsjö <martin@martin.st>
The code below the comment does not at all relate to statistics,
and even if moved to the right place, the comment adds little
value.
Signed-off-by: Martin Storsjö <martin@martin.st>
Previously, we always signalled a zero time since the last RTCP
SR, which is dubious.
The code also suggested that this would be the difference in
RTP NTP time units (32.32 fixed point), while it actually is
in in 1/65536 second units. (RFC 3550 section 6.4.1)
Signed-off-by: Martin Storsjö <martin@martin.st>
This brings back some code that was added originally in 4a6cc061
but never was used, and was removed as unused in 4cc843fa. The
code is updated to actually work and is tested to return sane
values.
Signed-off-by: Martin Storsjö <martin@martin.st>
The base_seq variable is set to first_seq - 1 (in
rtp_init_sequence), so no + 1 is needed here.
This avoids reporting 1 lost packet from the start.
Signed-off-by: Martin Storsjö <martin@martin.st>
The question can be answered: No, we do not know the initial sequence
number from the SDP. In certain cases, it can be known from the
RTP-Info response header in RTSP though. (In that case, we use it as
timestamp origin, but not for rtp receiver statistics.)
Signed-off-by: Martin Storsjö <martin@martin.st>
It is unclear what the bug exactly was and if it ever was fixed,
and we don't even support decoding via faad any longer. The
comment has been present since d0deedcb in 2006.
Signed-off-by: Martin Storsjö <martin@martin.st>
One of them is renamed now, but mentioning it by name serves
no purpose here. The other table mentioned ceased to exist
under that name in 4934884a1 in 2006.
Signed-off-by: Martin Storsjö <martin@martin.st>
Previously, for broken frames, we only returned the first partition
of the frame (we would append all the received packets to the packet
buffer, then set pkt->size to the size of the first partition, since
the rest of the frame could have lost data inbetween) - now instead
return the full buffered data we have, but don't append anything more
to the buffer after the lost packet discontinuity. Decoding the
truncated packet should hopefully get better quality than trimming out
everything after the first partition.
Signed-off-by: Martin Storsjö <martin@martin.st>
This is required by RFC 3550 (section 6.5):
The list of items in each chunk MUST be terminated by one or more
null octets, the first of which is interpreted as an item type of
zero to denote the end of the list.
This was implicitly added as padding before, unless the host name
length matched up so no padding was added.
This makes wireshark parse the packets properly if other RTCP items
are appended to the same packet.
Signed-off-by: Martin Storsjö <martin@martin.st>
This allows the caller to either include them (and get more packets
decoded, but possibly some nonperfect frames), or discard them (by
setting fflags=discardcorrupt).
Signed-off-by: Martin Storsjö <martin@martin.st>
This sends NACK for missed packets and PLI (picture loss indication)
if a depacketizer indicates that it needs a new keyframe, according
to RFC 4585.
This is only enabled if the SDP indicated that feedback is supported
(via the AVPF or SAVPF profile names).
The feedback packets are throttled to a certain maximum interval
(currently 250 ms) to make sure the feedback packets don't eat up
too much bandwidth (which might be counterproductive). The RFC
specifies a more elaborate feedback packet scheduling.
The feedback packets are currently sent independently from normal
RTCP RR packets, which is not totally spec compliant, but works
fine in the environments I've tested it in. (RFC 5506 allows this,
but requires a SDP attribute for enabling it.)
Signed-off-by: Martin Storsjö <martin@martin.st>
The warning is a false positive, but I prefer actually initializing
it over masking it with av_uninit, since the code is not performance
critical.
Signed-off-by: Martin Storsjö <martin@martin.st>
This is a bug from c7d4de3d73 - if the previous frame wasn't
returned yet (due to missing the final packets), but we have
enough data of it to return the first partition, we write that into
pkt and set returned_old_frame. That commit forgot returning 0 for
the case where this current packet didn't have the end_packet flag
set.
Signed-off-by: Martin Storsjö <martin@martin.st>
If we timed out and consumed a packet from the reordering queue,
but didn't return a packet to the caller, recheck the queue status.
Otherwise, we could end up in an infinite loop, trying to consume
a queued packet that has already been consumed.
CC: libav-stable@libav.org
Signed-off-by: Martin Storsjö <martin@martin.st>
To use this, set sdpflags=custom_io to the sdp demuxer. During
the avformat_open_input call, the SDP is read from the AVFormatContext
AVIOContext (ctx->pb) - after the avformat_open_input call,
during the av_read_frame() calls, the same ctx->pb is used for reading
packets (and sending back RTCP RR packets).
Normally, one would use this with a read-only AVIOContext for the
SDP during the avformat_open_input call, then close that one and
replace it with a read-write one for the packets after the
avformat_open_input call has returned.
This allows using the RTP depacketizers as "pure" demuxers, without
having them tied to the libavformat network IO.
Signed-off-by: Martin Storsjö <martin@martin.st>
So far, aviocontexts are used either in pure-read or pure-write
mode - full read/write mode doesn't work well (and implementing it
is a much larger, not totally trivial change).
This patch allows using avio_read and ffio_read_partial on
read/write aviocontexts, where the read operations are passed
through directly unbuffered, while writes are buffered as usual.
This is enough to support the operations needed by packet based
data transfer like in udp/rtp, where aviocontext is the only
public API for hooking up custom IO.
Signed-off-by: Martin Storsjö <martin@martin.st>
Limelight is a not too uncommon CDN. The authentication scheme is
pretty similar to the adobe authentication, but is even closer to
normal http digest authentication (but not close enough to warrant
sharing code) than the adobe version.
Signed-off-by: Martin Storsjö <martin@martin.st>
This is mostly used to authenticate the client when publishing.
Tested with wowza and akamai.
Some but not all servers support resending a new connect invoke
within the same connection, so always reconnect for sending a new
connection attempt. This matches what other applications do as well.
The authentication scheme is structurally pretty similar to http
digest authentication, but uses base64 instead of hex strings.
Signed-off-by: Martin Storsjö <martin@martin.st>
The description as updated in 6e5c44b0f made it sound like it is
handled exactly like the pts, which it is not. Clarify that it
is the packet that triggered returning the frame instead of just
the last avpacket (in case of frame threading, it won't be the
dts of the last packet).
Signed-off-by: Martin Storsjö <martin@martin.st>
This is built on the assumption that the first partition of each
VP8 packet is essential for decoding any later packet - if this
partition is broken/missed, the arithmetic coder gets out of sync
and decoding the bitstream in further packet ends up with total
garbage. If packets of a frame are lost, make sure the first
partition is intact (return only this part of the packet, nothing
else), otherwise stop returning data until the next keyframe is
received.
Alternatively, one would simply not return any packets at all
until the next keyframe, if packet loss is detected.
Signed-off-by: Martin Storsjö <martin@martin.st>
This makes all users of rtpenc_chain (rtsp muxer, sapenc, mov
rtp hinting) work again, broken since 8034130e0.
Signed-off-by: Martin Storsjö <martin@martin.st>
This makes it easier for receivers to decide what to do if data
is lost.
Refactor calculating the max payload size, to avoid hardcoding the
header size in too many places, reducing the number of lines that
have to be touched if the header is adjusted further.
Signed-off-by: Martin Storsjö <martin@martin.st>
This comment was added in e309128f, in 2002, and has been brought
along since then more or less unmodified.
The first point of the todo was implemented in dbf30963 in 2006,
the second one is not relevant to rtpdec.c (brought along from
rtp.c in 8eb793c4 in 2008) but would be more relevant to the
rtp muxer, although it isn't a good idea anyway.
Signed-off-by: Martin Storsjö <martin@martin.st>
When called from the v4l2 input device, pix_fmt can be
AV_PIX_FMT_NONE (for jpeg formats). Before 50ba57e0, this wasn't
an issue for avpicture_get_size, but after that commit, this
lead to crashes.
Signed-off-by: Martin Storsjö <martin@martin.st>
The def files are used for generating import libraries for
other toolchains (in particular, for generating import libraries
for MSVC for DLLs built with mingw).
The def files produced by mingw/gcc contains ordinals for each
exported function. When MSVC tools generate import libraries
from such a def file, MSVC links to the DLL by the ordinals
instead of linking by name.
Since the def files aren't maintained by hand, the ordinal
numbers are assigned (more or less) randomly and any caller
linking to the libs by ordinals will break as soon as the libraries
export more/fewer functions.
Therefore, strip out the ordinals from the generated def files,
to make users link to the libraries by name.
Callers linking to the DLLs using the gcc provided import library
link by name as they should.
Signed-off-by: Martin Storsjö <martin@martin.st>
When targeting the metro API subset, this function still exists in
the link libraries, but is excluded from the headers. This makes
sure w32threads is automatically disabled when targeting this API
subset (since not all the necessary functions for it are available).
Signed-off-by: Martin Storsjö <martin@martin.st>
Not all versions of windows have the console color functions,
while io.h might be needed for isatty (which can be found in
unistd.h or io.h).
Signed-off-by: Martin Storsjö <martin@martin.st>
The existence of MapViewOfFile isn't linked to the existence of
io.h.
Not all versions of windows have MapViewOfFile (in particular,
Windows Phone 8 and the "metro" windows 8 API subset don't),
while they still have io.h (and need it for open/read/close).
Signed-off-by: Martin Storsjö <martin@martin.st>
This simplifies the condition to avoid hardcoding the systems
where the function exists. This also simplifies support for
newer Windows API subsets where this function doesn't exist,
such as Windows Phone 8 and the "metro" API subset of Windows 8.
Signed-off-by: Martin Storsjö <martin@martin.st>
Martin Storsjö
