Instead add the necessary blocks directly into Mpeg12SliceContext.
This allows to completely remove MPVContext.block.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Add a GetBitContext to H263DecContext instead. This is in preparation
for removing MpegEncContext.gb.
Also move last_resync_gb to H263DecContext.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Given that the MPEG-1/2 decoders unquantize the blocks
on their own, block_last_index is only used to signal
to ff_mpv_reconstruct_mb() whether a block is skipped
or not; but this is only checked for inter macroblocks,
so it is unnecessary to set block_last_index for intra
macroblocks.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
These have been added in 29644cb504
in 2007 at a time when the MPEG-1/2 parser just set
the AVCodecContext's dimensions when encountering a sequence header,
so that the checks for the coded dimensions didn't trigger.
Yet this is no more and so we can simply remove these redundant checks.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Otherwise the MpegEncContext would be unnecessarily reinitialized
once (this does not affect the output for an intra-only variant
like VCR2) in mpeg_decode_postinit().
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The spec indeed does not allow the aspect ratio to change
within a video sequence, but this does not imply that we
must reinitialize the decoder if only the aspect ratio changes,
as this value does not affect the decoding process at all.
Moreover, our reinitialization is a bit buggy: If there is
frame reordering, then the not yet output next P-frame is
simply discarded upon reinit. With this patch this no longer
happens when only the aspect ratio changes.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The MPEG-1 decoder has an update_thread_context function set,
yet it was never enabled (missing AV_CODEC_CAP_FRAME_THREADS
flag). So remove it and also the ff_thread_finish_setup() call
as well as the progress reporting.
(Simply setting the flag would not be enough. The main problems
are:
a) The update_thread_context function relies on memcpy'ing
the whole context instead of only copying the necessary fields.
This leads to data races which is undefined behaviour.
(Btw: The check for whether the non-MpegEncContext fields
of Mpeg1Context should be copied has been broken in
7f0efe2324.)
b) Even an AVBufferRef* is simply copied, without creating a new
reference.
c) Copying the whole context happens only during init; when parameters
change lateron, the change is not propagated to the next thread.
This affects at least the quant matrix (for MPEG-1).
d) The MPEG-1/2 decoders are made to decode both MPEG-1 and MPEG-2.
When MPEG-2 is decoded, another complication arises: In case of
coded fields, both fields can be in one AVPacket or they can be
in separate AVPackets. One would need to parse enough of the data
to be able to determine whether the next thread needs to start
a new frame or decode the second field of the current frame;
for this one would need to either postpone calling
ff_thread_finish_setup() for coded slices altogether or implement
some form of ff_h2645_packet_split().
One would also need a second ThreadProgress to signal progress
of each field.
e) One would need to reenable waiting in ff_mpv_reconstruct_mb()
for MPEG-1/2 (and H.261).
f) Probably lots of other stuff for invalid input that I am currently
not thinking about.
If this were done, nothing from the current update_thread_context
would remain, so one can just nuke it altogether.)
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This commit adds a reference to the buffer as an argument to
start_frame, and adapts all existing code.
This allows for asynchronous hardware accelerators to skip
copying packet data by referencing it.
Nothing in this decoder would break if the generic code were to be
changed to allow slice "threading" with only one thread.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Fixes undefined integer overflows. The overflows could always
happen, yet before 4d8b706b1d
it was not undefined because the code implicitly used atomic
types, for which signed integer overflow is defined.
Reported-by: Kacper Michajlow <kasper93@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Compilers can not perform this optimization on their own
given that they don't know that the different thread_context
pointers don't alias.
Also avoid using sequentially consistent operations with
atomics when only a single thread accesses them.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
When used with the "mpegvideo" or "mpeg1video" decoder (which do not
have a private class) this would lead to the log callback receiving
a non-NULL object with a NULL AVClass pointer. The default log callback
handles this gracefully, yet it is probably an API violation.
Reviewed-by: Ramiro Polla <ramiro.polla@gmail.com>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The MPEG-1/2 decoders don't need ScanTable.raster_end
(as the coefficients are unquantized as they are parsed).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This decoder unquantizes while parsing blocks
and does not use dct_unquantize_mpeg1_intra
(which uses *_dc_scale) at all.
repeat_field was unused since
e0a3d744a0.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Only the MPEG-1/2, MSMPEG4V1, MPEG-4 and RV.10 decoders use last_dc
at all. Of these, RV.10 only uses it for intra frames; it does not
need these predictors reset in ff_mpv_reconstruct_mb(). MSMPEG4V1
has h263_pred set, so that last_dc is already not reset in
ff_mpv_reconstruct_mb() (instead it is reset at the beginning
of every line). MPEG-4 also has h263_pred set (and uses last_dc only
for the intra-only studio profile and needs them reset to sligthly
different values anyway).
So only the MPEG-1/2 decoders need these values reset. So move
resetting them there. This avoids resetting them unnecessarily
for FLV1, H.261, H.263I, RV.10, RV.20 and H.263(+)
(for the latter it depends upon whether h263_aic is in use).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
The format is used by at least Dish Network, but is not
defined in any DVB standard, so remove references to DVB.
This is a simple rename, no functional change.
Signed-off-by: Marth64 <marth64@proxyid.net>
In mpeg_decode_a53_cc() only the A/53 part 4 CC data ("GA94") is saved between
frames. The other formats incorrectly created a larger buffer than they use
since a705bcd763 because they did not append to
the previous data.
The a53_buf_ref is added to the frame in mpeg_field_start() which will only be
called in decode_chunks() if not all of the picture data slices are skipped.
For these formats, utilize the data added to the buffer in case frames are skipped
(concatenating the CC data until a frame can be exported), in a similar fashion to
the A/53 part 4 logic.
Reviewed-by: Marth64 <marth64@proxyid.net>
Signed-off-by: Marth64 <marth64@proxyid.net>
After support was added for DVB 0502 Closed Caption coding,
the cc_format option's range was never updated so user
cannot select this coding as a forced formatting choice.
Fix the range of the option by ending it with the new coding type.
Signed-off-by: Marth64 <marth64@proxyid.net>
These properties are unreliable because they depend on the frames decoded so
far, users should check directly the presence of the decoded AVFrame side data
or AVFrame flags.
Signed-off-by: Marton Balint <cus@passwd.hu>
When Closed Captions are discovered, this method is always invoked.
Therefore, use it to set the property instead of repeating the statement.
Signed-off-by: Marth64 <marth64@proxyid.net>
These values are not read anywhere. Furthermore, since commit
fe6037fd04 the linesize values
of the MPVWorkPictures were wrong for subsequent fields
in a chain of B-pictures (as they are always doubled and no longer
based upon the frame-linesizes) which can eventually lead to overflow.
Finally, it makes no real sense to ever double the linesize
of the reference pictures at all: Even when the current picture
is a field, it can still reference both fields of reference
pictures and therefore the linesize should allow to address
both fields (for the same reason, data is not offset for
reference pictures).
libavcodec/mpeg12dec.c:1304:41: runtime error: signed integer overflow: 4611686018427387904 * 2 cannot be represented in type 'long'
issue: 69732/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_MPEGVIDEO_fuzzer-5123551179374592
Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Possible by using MB_TYPE_CODEC_SPECIFIC for MB_TYPE_ZERO_MV
and due to the MB_TYPE_*_MV flags fitting into an int16_t.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
MB_TYPE_L[01] is based upon H.264 terminology (it stands for
list); yet the mpegvideo based decoders don't have lists
of reference frames, they have at most one forward and one
backward reference. So use terminology based upon this.
This also has a second advantage: MB_TYPE_L[01] is actually
an OR of two flags (which are set independently for H.264,
but aren't for mpegvideo). Switching to different flags
makes the flags fit into an int16_t, which will be useful
in future commits.
The only downside to this is a very small amount of code
in error_resilience.c and mpegutils.c (the only code shared
between the H.264 decoder and mpegvideo).
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
It has been added in a579db0c4f
due to XvMC, but today it is only used to swap U and V
for VCR2, a MPEG-2 variant with U and V swapped.
This can be done in a simpler fashion, namely by simply
swapping the U and V pointers of the corresponding
MPVWorkPictures.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
MPEG-1/2 only needs one scantable and therefore all code
already uses the intra one. So stop initializing
the inter one altogether.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Up until now, an initialized MpegEncContext had an array of
MPVPictures (way more than were ever needed) and the MPVPicture*
contained in the MPVWorkPictures as well as the input_picture
and reordered_input_picture arrays (for the encoder) pointed
into this array. Several of the pointers could point to the
same slot and because there was no reference counting involved,
one had to check for aliasing before unreferencing.
Furthermore, given that these pointers were not ownership pointers
the pointers were often simply reset without unreferencing
the slot (happened e.g. for the RV30 and RV40 decoders) or
there were moved without resetting the src pointer (happened
for the encoders where the entries in the input_picture
and reordered_input_picture arrays were not reset).
Instead actually releasing these pictures was performed by looping
over the whole array and checking which one of the entries needed
to be kept. Given that the array had way too many slots (36),
this meant that more than 30 MPVPictures have been unnecessarily
unreferenced in every ff_mpv_frame_start(); something similar
happened for the encoder.
This commit changes this by making the MPVPictures refcounted
via the RefStruct API. The MPVPictures itself are part of a pool
so that this does not entail constant allocations; instead,
the amount of allocations actually goes down, because the
earlier code used such a large array of MPVPictures (36 entries) and
allocated an AVFrame for every one of these on every
ff_mpv_common_init(). In fact, the pool is only freed when closing
the codec, so that reinitializations don't lead to new allocations
(this avoids having to sync the pool in update_thread_context).
Making MPVPictures refcounted also has another key benefit:
It makes it possible to directly share them across threads
(when using frame-threaded decoding), eliminating ugly code
with underlying av_frame_ref()'s; sharing these pictures
can't fail any more.
The pool is allocated in ff_mpv_decode_init() for decoders,
which therefore can fail now. This and the fact that the pool
is not unreferenced in ff_mpv_common_end() also necessitated
to mark several mpegvideo-decoders with the FF_CODEC_CAP_INIT_CLEANUP
flag.
*: This also means that there is no good reason any more for
ff_mpv_common_frame_size_change() to exist.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
Currently identical to the H.261 and H.263 close functions
(which it replaces). It will be extended in future commits.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
This is all that is used. This is in preparation for further
commits that will extend ff_mpv_decode_init() in a way
that will make it possible to fail and require cleanup.
Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
