Up until now setting the input and output devices lists is guarded
by a mutex. This prevents data races emanating from multiple concurrent
calls to avpriv_register_devices() (triggered by multiple concurrent
calls to avdevice_register_all()). Yet reading the lists pointers was
done without any lock and with nonatomic variables. This means that
there are data races in case of concurrent calls to
av_(de)muxer_iterate() and avdevice_register_all() (but only if the
iteration in av_(de)muxer_iterate exhausts the non-device (de)muxers).
This commit fixes this by putting said pointers into atomic objects.
Due to the unavailability of _Atomic the object is an atomic_uintptr,
leading to ugly casts. Switching to atomics also allowed to remove
the mutex currently used in avpriv_register_devices().
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Fixes: signed integer overflow: 9223372036854775807 + 8 cannot be represented in type 'long'
Fixes: 38787/clusterfuzz-testcase-minimized-ffmpeg_dem_AVI_fuzzer-4859845799444480
Found-by: continuous fuzzing process https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>
This brings the FLAC demuxer in line with all the other demuxers.
Furthermore, if it is not done and the FLAC decoder is disabled,
the FLAC parser will overwrite the channel layout with the standard
channel layout for that number of channels.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
In most contexts, arrays are automatically converted to a pointer
to their first element; taking the address of the array just yields
a pointer to an array of fixed-size arrays, which is not intended here.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This information is coded in a standard MP4 KindBox and utilizes the
scheme and values as per the DASH role scheme defined in MPEG-DASH.
Other schemes are technically allowed, but where multiple schemes
define the same concepts, the DASH scheme should be utilized.
Such flagging is additionally utilized by the DASH-IF CMAF ingest
specification, enabling an encoder to inform the following component
of the roles of the incoming media streams.
A test is added for this functionality in a similar manner to the
matroska test.
Signed-off-by: Jan Ekström <jan.ekstrom@24i.com>
This is especially useful when reading things such as null-terminated
strings from MOV/MP4-likes, where the size of the box is known, but
not the exact size of the string.
Signed-off-by: Jan Ekström <jan.ekstrom@24i.com>
For now, same as ff_read_line_to_bprint_overwrite, but reads until
the end of a null-terminated string.
Signed-off-by: Jan Ekström <jan.ekstrom@24i.com>
Additionally:
* rename it to read_string_to_bprint
* split most of ff_read_line_to_bprint_overwrite into an internal
function which can then be utilized to implement other
functionality without duplicating code.
Signed-off-by: Jan Ekström <jan.ekstrom@24i.com>
Given that the packet sent to av_bsf_send_packet() is always
already refcounted, it is doubtful whether the error can even
be triggered currently.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
In this case it means replacing a packet in the muxer's context by
a pointer to an AVPacket, namely AVFormatInternal.pkt.
Because this packet is freed generically, one can remove the muxer's
deinit function.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
It has never been done in b08569a23948db107e5e6175cd4c695427d5339d,
30b8f3e7dcd0318b91e205dcbf774ef92fb6193c.
After this change, this muxer does no longer use sizeof(AVPacket).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The documentation of said packet ("Every user has to ensure that
this packet is blank after using it") perfectly fits how we use said
packet in the generic muxing code. Better than the documentation of pkt.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Currently the interleave_packet functions use a packet for
a new packet to be interleaved (may be NULL if there is none) and
a packet for output; said packet is always a stack packet in
interleaved_write_packet(). But all the interleave_packet functions
in use first move the packet to the packet list and then check whether
a packet can be returned, i.e. the effective lifetime of the new packet
ends before the packet for output is touched.
So one can use one packet both for input and output by adding a new
parameter that indicates whether there is a packet to add to the packet
list; there is just one complication: In case the muxer is flushed,
there is no packet available. This can be solved by reusing one of
the packets from AVFormatInternal. They are currently unused when
flushing in av_interleaved_write_frame().
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The earlier documentation claimed that av_interleaved_write_frame()
always orders by dts, which is not necessarily true when using muxers
with custom interleavement functions or the audio_preload option.
Furthermore, the documentation stated that libavformat takes ownership
of the reference of the provided packet (if it is refcounted) and that
the caller may not access the data through this reference after the
function returns. This suggests that the returned packet is not blank,
but instead still contains some set, but invalid fields, which implies
that it would be dangerous to unreference this packet again.
But this is not true: av_interleaved_write_frame()'s actual behaviour
is to always output blank packet (even on error). This commit documents
this fact so that callers know that they can directly reuse this packet.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The BSF API treats such packets as signalling EOF and therefore
such a packet might corrupt the BSF state. In such a case,
the guarantee that av_interleaved_write_frame() always frees
the packet is not upheld.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
It claims to always zero-terminate its buffer like snprintf(),
yet it does it not on EOF. Because of this the mcc demuxer
used uninitialized values when reading an empty input file.
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
ff_subtitles_queue_insert() does not require its events to be
zero-terminated as it has a parameter for the length.
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The av_sscanf() will filter lines like "Scenarist_SCC V1.0" out.
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Up until now, the scc demuxer not only read the line that it intends
to process, but also the next line, in order to be able to calculate
the duration of the current line. This approach leads to unnecessary
complexity and also to bugs: For the last line, the timing of the
next subtitle is not only logically indeterminate, but also
uninitialized and the same applies to the duration of the last packet
derived from it.* Worse yet, in case of e.g. an empty file, it is not
only the duration that is uninitialized, but the whole timing as well
as the line buffer itself.** The latter is used in av_strtok(), which
could lead to crashes. Furthermore, the current code always outputs
at least one packet, even for empty files.
This commit fixes all of this: It stops using two lines at a time;
instead only the current line is dealt with and in case there is
a packet after that, the duration of the last packet is fixed up
after having already parsed it; consequently the duration of the
last packet is left in its default state (meaning "unknown/up until
the next subtitle"). If no further line could be read, processing
is stopped; in particular, no packet is output for an empty file.
*: Due to stack reuse it seems to be zero quite often; for the same
reason Valgrind does not report any errors for a normal input file.
**: While ff_subtitles_read_line() claims to always zero-terminate
the buffer like snprintf(), it doesn't do so if it didn't read anything.
And even if it did, it would not necessarily help here: The current
code jumps over 12 bytes that it deems to have read even when it
hasn't.
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
