Do not use dynamic types anymore. The reason is that it's difficult:
- to maintain a shared type database over a network.
- the extra overhead when translating between processes and for
maintaining the translation tables.
- race conditions in translating in RT-threads, this is a problem
because we want to make event streams.
We now have simple enums with types and extension points for all
types. This is also nicer to use in general.
We don't need the mapper anymore or pass strings around as types.
There is a parallel type info system to get more info about ids and
enums and their hierarchy. It can also be used for debugging.
Make a method to find the memory block for the given ptr. We can
use this to find the memfd of the memory if there is any. We can
then remove the Shared metadata on buffers.
We can't use a ringbuffer on the chunk because it implies the
consumer would write to it to update the read position, which we
can't do because the chunk is read-only and might even be shared.
Go back to offset/size pairs, which can sortof do the same thing
if we want later when we keep a non-shared read pointer in the
consumer.
Keep alsa timestamp around and filled state for future.
mmap the input port meta/data/chunk as read-only.
Only do clock update requests when asked.
We can't allocate the shared meta in shared mem because then clients can
damage it for other clients. Place it instead right after the buffer
metadata array. Filter out the shared metadata for a client, we send it
as part of the client_buffer structure.
Remove pointer metadata, it's not so useful.
Document the layout of the allocated buffers and the shared memory.
Work on metadata to define control parameters
ringbuffer: remove size and mask from the ringbuffer, we have that
elsewhere in the user of the ringbuffer.
Remove the buffer data offset and size fields and replace with a
ringbuffer. We then have a ringbuffer in all buffer data, which
simplifies things.
We can now remove the ringbuffer metadata.
