With large quantum and many channels, the silence buffer can become
quite large (many MB), which can cause a stack overflow.
Use a fixed size silence buffer instead and write in smaller chunks.
If the "debug" log level is not enabled for the "spa.alsa" log topic,
then there is no point in going into the loop and splitting the data
into lines, so skip that.
Do not use `strcspn()` because it assumes a null terminated string,
but the `fopencookie()` write callback receives a (ptr, length) pair.
So use `memchr()` instead to find the `\n`.
The `fopencookie()` write callback should return the number of consumed
bytes, but it currently only ever returns 0, which signals an error
condition according to the documentation.
Fix that by not overwriting `size`.
Fixes: 73073eb33f ("alsa: redirect alsa output to log file")
The programs in `doc/examples` are not compiled currently, so let's
compile them if the `docs` and `examples` options don't disallow it.
`tutorial4.c` needs a small modification to avoid `-Wfloat-conversion`.
Additionally, install them if `installed_tests` is not disabled.
When the graph rate changes it is possible that the follower node can
renegotiate to the new suggested audioconvert rate without requiring
resampling and so the extra check for the disabled resampler is not
required.
Fixes#4933
Move the source offs, stride, data and size calculations out of the
destination loop. We only need to clamp the size to copy to the maxsize
of the destination buffer.
aacDecoder_DecodeFrame expects the number of destination INT_PCM samples,
not bytes. Since INT_PCM is int16_t (2 bytes), passing dst_size in bytes
tells the decoder the buffer is 2x larger than reality.
Note that we don't need to care about the number of channels in this
size, the decoder will do that for us.
parse_sdp_a_rtpmap used c += strlen(c) + 1 to skip past the MIME type to the
rate/channels part, but if the a=rtpmap: line had no / separator, strcspn
returned the full string length and the +1 advanced past the null terminator.
Fix this by checking if / was actually found, returning -EINVAL if not.
JACK2 only sends -1 as the frames, meaning we should take the value from
the negotiated period as the frames to process.
We however send the actual number of frames and use the sync value to
decide how many frames to process. We need to be careful because a value
of 0 will cause a division by 0 so treat <= 0 frames the negotiated period
size as well.
Check that the params don't include more than MAX_CHANNELS of audio or
else we overflow the position array.
Adapt to the compiled value of SPA_AUDIO_MAX_CHANNELS but allow at least
128 channels.
It's a terrible idea, doesn't work so well (locks up the data-loop when
read is blocked) and a security mightmare. If you really need to pipe
samples through some program, do that somewhere else, like from the
command line with pw-cat and pw-record.
Normally, when loading a plugin feature, often a library.name property
is given as well. If the feature to load is not explicitly listed in
context.spa-libs, the library.name is used a fallback library.
Add an option to ignore this library.name and only use the
context.spa-libs entries. This makes it possible to only load explicitly
listed features in the config file and makes it possible to lock down
what plugins can be loaded.
Set the option to true by default for now, which keeps the existing
behaviour of using the fallback library. Add some more entries to the
context.spa-libs in case the option is switched off to make things
work.
Set the option to false for the minimal.conf.
If an adapter's removal is processed before the pending `RegisterApplication()`
dbus calls return, then those pending calls are not cancelled, and when the
(error) replies arrive, the callbacks will run into use-after-free issues
since they reference the removed adapter.
See #5096
Add a special 'blocked' spa-libs value that returns EPERM when trying to
load the factory.
Only allow loading the LADSPA filter.graph nodes for the LADSPA sink and
source. The most problematic part is the pipe filter, that allows it to
spawn arbirary programs as part of the filter.graph.
You can add a filter-graph to any stream with stream_props.
Use spa_steal_ptr to transfer props ownership when we can.
This fixes a problem in the upload stream where the props would be freed
twice when buffer allocation failed, once with properties_free and
then with stream_free.
This handles overflow and errors correctly, unlike snprintf which might
return -1 or the size that would have been written if truncated, causing
overwrite later.
