process_output is now to finish processing of the output in async nodes,
which means we need to fill the io area before sending the HAVE_OUTPUT
event. This simplifies some things and improves performance because we
don't need to deal with queues and additional checks.
Only send a buffer when we have received a NEED_DATA message.
Add a signal to pulla buffer from the sink. Restructure the sink to use
a queue like the source and only push a buffer when we can.
Improve SpaData. Offset and size should be between 0 and maxsize, make
sure we clamp correctly when needed.
node_process_output completes the processing of the output after
receiving HAVE_OUTPUT for async elements. It instructs the node that
it now can produce more output.
Move offset, size and stride to separate chunk structure that we
can keep in shared memory.
Add shared metadata type to hold the memory block holding the
metadata.
Add connection message for PORT_COMMAND
Add rtkit support to ask for realtime priority
work on stream states and improve negotiation
Rework of port linking works, keep separate state for realtime threads
and use message passing to update the state.
Don't try to link nodes that are removed.
Open the device in the ALSA monitor to detect source or sink
Implement send_command as async methods on the plugins, use async
replies to sync start and stop.
Work on alsa sink.
Implement async PAUSE/START on v4l2 src. move the STREAMON/OFF calls to
the mainloop because they have high latency, add the poll descriptors
from the data loop.
Keep lock while doing recycle_buffer
Pass context to main loop
Recycle WorkItems in a free list
Push and pull from the node data loop
Do async start/pause in alsa
Add alsa fds right after open.
Make a main-loop object with associated helper functions to handle async
methods.
rtloop -> data_loop
Handle async results a lot better.
Remove REMOVE_MEM command. We don't need it.
Handle stream memory updates better.
Add the fd to the SpaData so that we can noth use the fd and the mmapped
memory when we want instead of needing to mmap again in the plugin.
various cleanups
Add explicit DMABUF memory because we can't use it like an fd in v4l2.
mmap random fd memory and use userptr in v4l2 to use it.
Allocate buffer data memory in memfd when no node can allocate.
Improve debug some more
Enable the header metadata so that we have timestamps again.
Rework allocation in the link a little. Reuse previously allocated
buffers when possible. Keep track of where the buffers was allocated.
Make some more convenience functions to initialize buffer headers. Make
it so that we move allocations to the app.
Make some helpers to make memfd memory in pinos now that we will be able
to do all allocation there.
Remove the memory stuff from the spa API, we can do thing more simple
and efficiently if we always allocate buffers outside of the plugins and
only implement an alloc function on the node when it can do something
clever.
Move serialize code to the props/format/buffer code
Make it possible to copy a format and properties.
Add an async result code and an event to signal the completion.
Use async return values to signal completion of a method and potential
state change.
Add selected format to port update message.
Make it possible to parse into a custom copy of the command memory.
Remove state change events from the elements, we now just update the
state.
Implement async results in the proxy element
Add support for removing buffers in the client.
Fix up pinossink
Deal with async return in the links.
Reserve 0,0 for stack allocated mem that we never free
Improve allocation in the link
Parse format and properties in-line, there is no need to allocate
memory.
Improve buffer cleanup in alsa
Add timestamps and offset to alsa
Add a link state
Add error quark
Track the state of node we create and error when it is in error.
Handle stream error states when negotiating
Make the node error when a link is in error
Use NULL filter when the port format enumeration returns nothing.
Send clock update even when there is no clock.
Don't send data to a port when the node is not streaming
Add latency support to the clock update
Copy the unset_mask when copying the formats
Make sure we always send a clock update before sending a start command.
Fix memory size.
Follow the state of the node to do allocation
Use the pinos node to control the state
When doing STREAMOFF, all buffers will be dequeued. requeue them buffers
for when we go to playing again.
Add a clock interface to get raw clocking information from an element
Make events and commands to request and configure clock updates.
Add support for passing events and commands in control
Set the size of the buffers in pinossrc
Remove _remove from properties, we can do the same with set of a NULL
value.
Add signals to the stream API to manage the buffers. Wrap those buffers
in a GstBuffer in the pinossrc and pinossink elements and pool them in a
bufferpool.
Remove SPA_EVENT_TYPE_PULL_INPUT, we can do the same with NEED_INPUT and
by using a ringbuffer.
Do more complete allocation of buffers in the link. Use the buffer
allocator if none of the nodes can allocate.
Follow the node state to trigger negotiation and allocation.
Remove offset and size when refering to buffers, we want to always deal
with the complete buffer and use a ringbuffer for ranges or change the
offset/size in the buffer data when needed.
Serialize port_info structures as part of the port_update
Print both the enum number and the name when debuging properties or
formats.
Add helpers to convert between pinos and gstreamer formats.
Use pinos formats in the API.
Work on removing some hardcoded stuff and instead use the real format
from the pinos server.
Use memfd and sealing.
Make buffer data point to a registered memory block by its mem_id.
Add some more helpers to allocate memfd backed memory.
Allocate buffers in memfd so that we easily share them between client
and server.
Update pts and seq in v4l2 now that this change will actually be visible
at the client.
Use stucture offsets instead of pointers so that we can share the buffer
structure between client and server.
We can then pass an fd wrapping the memory of the buffer to the client.
Add beginnings of a memory pool
Rework things so that we negotiate buffer pools beforehand and only pass
buffer ids around We can then remove the refcount of buffers, events and
commands.
More work on buffer reuse
Use the node state changes to trigger the next step in the configuration
sequence.
Move most of the client-node to a plugin
Do buffer allocation in the port link.
Remove the old PinosBuffer object and replace it with SpaControl, this
communication protocol is designed to make it possible to implement
remote nodes and so it is moved to Spa.
Move SpaBuffer into to API
Work on easier API to make formats, implement enumeration and support
for all formats in v4l2.
Improve format output in -inspect
Add more buffer types to add and remove memory shared memory between the
server and client. We would like to send buffers only once and then
simply reference them by index.
Do format negotiation and stream start with a START message.
Bring back the channel object. Making a node and port on the client side
was rather awkward because of the async nature of many methods. It feels
better to have a specific communication channel object to interface with
a server side port.
Use port activate/deactivate to start/stop streams
Remove links from the ports. We let other objects install a callback on
the port to receive and route buffers.
Add autoconnect property to the ports
Remove the upload-node, clients that want to upload media just create a
node with an (unconnected) output port. Other clients can then connect
to this new output port.
Copy formats on newly linked ports.
Add refcounting to buffers.
Use static memory on ports for buffer input and output memory.
Improve negotiation, let format and properties be negotiated on ports as
the data passes.
Improve some debug info.
Add buffer flags. The idea is to make it possible to easily check when a
buffer contains control information that we need to parse to update the
port fields.
Make the client create remote nodes and ports and set up proxies for
them.
Make a port base class implementing most of the logic to pass buffers
locally and remotely.
Remove most code from stream.c, it's now in the port.
Make a portsink and portsrc that can write and read to/from any port. We
use these in the server to send and receive data.
Rework format negotiation. The final format is now sent in-line before
the data. The server will select a format on output ports.
Make it possible to pass custom memory to the builder so that it can
construct packets directly into specified memory and avoid allocs.
Remove GError and GSocketControlMessage in the buffer API to make it
possible to use other (rt-safe) API later.
A port is an input or output on a Node.
Channels are created from the ports and inherit the direction of the
port.
do automatic port selection based on the direction and caps and
node/port name.
Simplify stream_connect by passing the direction.
Fix pinossink to connect in setcaps so that we know the format and can
select a good sink to connect to.
Make an object for a processing node.
Implement a sink node. Make it possible to implement Sink and Source
interfaces to provide input/output from the node.
Improve pinosdepay to track fds and handle format changes.
Rename the source-output object to channel because it is used for both
input and output.
Start the beginnings of sink support. This will make it possible to make
pinos consume data as well as provide data.
Only go to READY in setup_pipeline, this is enough to get the caps. We
don't want to go to PAUSED yet because we don't want to negotiate a
format.
Mark the source busy after we configured the capsfilter so that we can
let the source negotiate correctly.
Send the base-time to the client so that it can calibrarte the clock.
Wait until we are completely negotiated before completing the state
change to paused. This makes it possible to provide a clock and decide
if we are live or not.
Add a new refresh message to request a keyframe from the pinos server.
pinospay: pass the refresh-request message upstream
pinossink: turn refresh-request messages into events
pinossrc: turn a keyframe event into a refresh-request message
