mirrors/pipewire
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/pipewire/pipewire.git synced 2025-11-17 07:00:03 -05:00
Code Issues Projects Releases Packages Wiki Activity Actions
pipewire/spa/include/spa/buffer/buffer.h

113 lines
3.3 KiB
C
Raw Normal View History

treewide: use SPDX tags to specify copyright information SPDX tags make the licensing information easy to understand and clear, and they are machine parseable. See https://spdx.dev for more information.
2023-02-08 18:12:00 +01:00
/* Simple Plugin API */
/* SPDX-FileCopyrightText: Copyright © 2018 Wim Taymans */
/* SPDX-License-Identifier: MIT */
more hacking Move plugin API to separate directory for now Add ringbuffer and way to get ringbuffer from a port
2016-06-28 12:21:56 +02:00
Don't use __ in defines or declarations, it's reserved SPA_TYPE -> SPA_TYPE_INFO for type info strings improve includes Fixes #115
2019-01-14 12:58:23 +01:00
#ifndef SPA_BUFFER_H
#define SPA_BUFFER_H
more hacking Move plugin API to separate directory for now Add ringbuffer and way to get ringbuffer from a port
2016-06-28 12:21:56 +02:00
#ifdef __cplusplus
extern "C" {
#endif
Reorganise SPA tree Reorganise the SPA includes to make it more extensible later Simplify the naming of the buffer and meta params
2017-11-10 13:36:14 +01:00
#include <spa/utils/defs.h>
#include <spa/buffer/meta.h>
Work on sink 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.
2016-08-24 16:26:58 +02:00
doc: reorganize Doxygen groups/modules to make more sense The Doxygen "Modules" page is not very illuminative, as different parts of the API are mixed together and not all parts are included. Try to address this: Put all parts of the public API to some Doxygen group, usually one group per header file. Use short, systematic names. Make these groups sub-groups of a few top-level groups, roughly corresponding to the different logical parts of the API (core, impl, stream, filter, spa, utilities).
2021-10-02 20:55:53 +03:00
/** \defgroup spa_buffer Buffers
Make done event on node Replace the AsyncDone event with an explicit done callback in the node to signal completion of an async operation. Pass read and write fd together with the transport
2017-06-09 17:24:18 +02:00
*
* Buffers describe the data and metadata that is exchanged between
* ports of a node.
*/
More hacking Add spa based v4l2 pinos module Add allocation params to port_alloc_buffers Let the app do allocation for handles.
2016-07-13 18:29:55 +02:00
spa: group the spa headers into doxygen modules Subdirectories buffer, control, debug, monitor, pod, support and utils, others are still missing. Headers are grouped either per subdirectory (e.g. buffer/ gets added to group spa_buffer) or per-file (e.g. spa_json is a separate group), whatever seemed like the most sensible approach.
2021-05-21 14:03:07 +10:00
/**
* \addtogroup spa_buffer
* \{
*/
Remove dynamic types 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.
2018-08-23 17:47:57 +02:00
enum spa_data_type {
Make media type and subtype properties This is easier to debug and we only want properties in objects in the future.
2018-08-29 18:08:52 +02:00
SPA_DATA_Invalid,
More type cleanups SPA_ID_* -> SPA_TYPE_*. We use 'type' for all registered types. ID refers to either a registered type or an enum. Improve introspection of object ids. Make 0 an invalid enum and use it to look up the object id type. Move some type-info files together. Swap type and id of the object, we first specify the type and then the id because the meaning of the id depends on the object type.
2018-08-27 15:03:11 +02:00
SPA_DATA_MemPtr, /**< pointer to memory, the data field in
* struct spa_data is set. */
SPA_DATA_MemFd, /**< generic fd, mmap to get to memory */
SPA_DATA_DmaBuf, /**< fd to dmabuf memory */
mem: use MemId to pass buffer memory Track memory on buffers in the server and simplify some memory tracking on the client. Don't send internal memblock flags to client. Add read/write flags to memblock.
2019-07-24 10:33:26 +02:00
SPA_DATA_MemId, /**< memory is identified with an id */
add more tests
2019-01-07 18:00:23 +01:00
spa: prefix SPA_*_LAST enum values with an underscore These are not part of the ABI, prefixing them makes this a bit more obvious that they shouldn't be used.
2021-05-06 10:55:40 +10:00
_SPA_DATA_LAST, /**< not part of ABI */
spa: don't use typedef for struct and enum
2017-05-25 13:28:15 +02:00
};
buffer: make data and meta types dynamic Use dynamic types for the data and meta types. Move the meta into a separate header. Remove port state, we can keep track of that ourselves. Add size to MetaDataEnable params
2017-04-26 18:42:50 +02:00
mem: Add tag to memmap Add a tag field when creating a memmap so that we can do lookup on it. This makes it easier to implement the tracking of mappings for io areas. Remove custom io memory tracking and use the tags. Add flags to spa_chunk to make data corrupted. The flags on the buffer stay constant for the life of the buffer. Add flags to mark memory readable and writable. Mark memory readonly in v4l2-source. Pass the daemon activation area to the client in the transport event. This never changes and need to be handled differently from the other activation areas. Use the right flags when importing memory. Add the (desired) memory type to mempool_alloc. improve some debug.
2019-07-25 12:10:05 +02:00
/** Chunk of memory, can change for each buffer */
spa: don't use typedef for struct and enum
2017-05-25 13:28:15 +02:00
struct spa_chunk {
buffer: don't use ringbuffer in chunk 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.
2017-11-21 12:30:15 +01:00
uint32_t offset; /**< offset of valid data. Should be taken
* modulo the data maxsize to get the offset
* in the data memory. */
uint32_t size; /**< size of valid data. Should be clamped to
* maxsize. */
int32_t stride; /**< stride of valid data */
mem: Add tag to memmap Add a tag field when creating a memmap so that we can do lookup on it. This makes it easier to implement the tracking of mappings for io areas. Remove custom io memory tracking and use the tags. Add flags to spa_chunk to make data corrupted. The flags on the buffer stay constant for the life of the buffer. Add flags to mark memory readable and writable. Mark memory readonly in v4l2-source. Pass the daemon activation area to the client in the transport event. This never changes and need to be handled differently from the other activation areas. Use the right flags when importing memory. Add the (desired) memory type to mempool_alloc. improve some debug.
2019-07-25 12:10:05 +02:00
#define SPA_CHUNK_FLAG_NONE 0
#define SPA_CHUNK_FLAG_CORRUPTED (1u<<0) /**< chunk data is corrupted in some way */
buffer: add EMPTY chunk flag Add an EMPTY chunk flag to mark a piece of memory as 'empty'. For audio this means silence. Use the empty flag to avoid mixing 0 samples. Set the empty flag in output buffers on audioconvert.
2022-06-24 11:09:01 +02:00
#define SPA_CHUNK_FLAG_EMPTY (1u<<1) /**< chunk data is empty with media specific
* neutral data such as silence or black. This
* could be used to optimize processing. */
mem: Add tag to memmap Add a tag field when creating a memmap so that we can do lookup on it. This makes it easier to implement the tracking of mappings for io areas. Remove custom io memory tracking and use the tags. Add flags to spa_chunk to make data corrupted. The flags on the buffer stay constant for the life of the buffer. Add flags to mark memory readable and writable. Mark memory readonly in v4l2-source. Pass the daemon activation area to the client in the transport event. This never changes and need to be handled differently from the other activation areas. Use the right flags when importing memory. Add the (desired) memory type to mempool_alloc. improve some debug.
2019-07-25 12:10:05 +02:00
int32_t flags; /**< chunk flags */
spa: don't use typedef for struct and enum
2017-05-25 13:28:15 +02:00
};
Rework buffer memory layout 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.
2016-12-15 14:57:34 +01:00
mem: Add tag to memmap Add a tag field when creating a memmap so that we can do lookup on it. This makes it easier to implement the tracking of mappings for io areas. Remove custom io memory tracking and use the tags. Add flags to spa_chunk to make data corrupted. The flags on the buffer stay constant for the life of the buffer. Add flags to mark memory readable and writable. Mark memory readonly in v4l2-source. Pass the daemon activation area to the client in the transport event. This never changes and need to be handled differently from the other activation areas. Use the right flags when importing memory. Add the (desired) memory type to mempool_alloc. improve some debug.
2019-07-25 12:10:05 +02:00
/** Data for a buffer this stays constant for a buffer */
spa: don't use typedef for struct and enum
2017-05-25 13:28:15 +02:00
struct spa_data {
buffer: document memory type when allocating When allocating buffers the type will contain a bitmask of allowed types.
2020-03-02 17:21:16 +01:00
uint32_t type; /**< memory type, one of enum spa_data_type, when
* allocating memory, the type contains a bitmask
buffer: clarify data type mask
2020-10-09 17:39:32 +02:00
* of allowed types. SPA_ID_INVALID is a special
* value for the allocator to indicate that the
* other side did not explicitly specify any
* supported data types. It should probably use
* a memory type that does not require special
* handling in addition to simple mmap/munmap. */
More type cleanups SPA_ID_* -> SPA_TYPE_*. We use 'type' for all registered types. ID refers to either a registered type or an enum. Improve introspection of object ids. Make 0 an invalid enum and use it to look up the object id type. Move some type-info files together. Swap type and id of the object, we first specify the type and then the id because the meaning of the id depends on the object type.
2018-08-27 15:03:11 +02:00
#define SPA_DATA_FLAG_NONE 0
mem: Add tag to memmap Add a tag field when creating a memmap so that we can do lookup on it. This makes it easier to implement the tracking of mappings for io areas. Remove custom io memory tracking and use the tags. Add flags to spa_chunk to make data corrupted. The flags on the buffer stay constant for the life of the buffer. Add flags to mark memory readable and writable. Mark memory readonly in v4l2-source. Pass the daemon activation area to the client in the transport event. This never changes and need to be handled differently from the other activation areas. Use the right flags when importing memory. Add the (desired) memory type to mempool_alloc. improve some debug.
2019-07-25 12:10:05 +02:00
#define SPA_DATA_FLAG_READABLE (1u<<0) /**< data is readable */
#define SPA_DATA_FLAG_WRITABLE (1u<<1) /**< data is writable */
#define SPA_DATA_FLAG_DYNAMIC (1u<<2) /**< data pointer can be changed */
#define SPA_DATA_FLAG_READWRITE (SPA_DATA_FLAG_READABLE|SPA_DATA_FLAG_WRITABLE)
Make done event on node Replace the AsyncDone event with an explicit done callback in the node to signal completion of an async operation. Pass read and write fd together with the transport
2017-06-09 17:24:18 +02:00
uint32_t flags; /**< data flags */
spa: make fd in spa_data an int64_t So that we can store other things in it later as well.
2019-02-07 12:22:51 +01:00
int64_t fd; /**< optional fd for data */
Make done event on node Replace the AsyncDone event with an explicit done callback in the node to signal completion of an async operation. Pass read and write fd together with the transport
2017-06-09 17:24:18 +02:00
uint32_t mapoffset; /**< offset to map fd at */
uint32_t maxsize; /**< max size of data */
void *data; /**< optional data pointer */
struct spa_chunk *chunk; /**< valid chunk of memory */
spa: don't use typedef for struct and enum
2017-05-25 13:28:15 +02:00
};
more hacking Move plugin API to separate directory for now Add ringbuffer and way to get ringbuffer from a port
2016-06-28 12:21:56 +02:00
Make done event on node Replace the AsyncDone event with an explicit done callback in the node to signal completion of an async operation. Pass read and write fd together with the transport
2017-06-09 17:24:18 +02:00
/** A Buffer */
spa: don't use typedef for struct and enum
2017-05-25 13:28:15 +02:00
struct spa_buffer {
cleanups Follow the data/size argument order convention to make the api more consistent.
2017-11-20 15:26:44 +01:00
uint32_t n_metas; /**< number of metadata */
uint32_t n_datas; /**< number of data members */
buffer: remove buffer id The buffer id is not needed/used, we use the index of the buffer in the buffer array when configuring the port.
2019-01-07 17:57:03 +01:00
struct spa_meta *metas; /**< array of metadata */
struct spa_data *datas; /**< array of data members */
more hacking Move plugin API to separate directory for now Add ringbuffer and way to get ringbuffer from a port
2016-06-28 12:21:56 +02:00
};
Make done event on node Replace the AsyncDone event with an explicit done callback in the node to signal completion of an async operation. Pass read and write fd together with the transport
2017-06-09 17:24:18 +02:00
/** Find metadata in a buffer */
meta: make safer find_meta_data function
2018-07-09 12:04:23 +02:00
static inline struct spa_meta *spa_buffer_find_meta(const struct spa_buffer *b, uint32_t type)
Add fd to the data 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
2016-10-04 19:37:44 +02:00
{
indent
2017-05-26 08:05:01 +02:00
uint32_t i;
Add fd to the data 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
2016-10-04 19:37:44 +02:00
indent
2017-05-26 08:05:01 +02:00
for (i = 0; i < b->n_metas; i++)
if (b->metas[i].type == type)
meta: make safer find_meta_data function
2018-07-09 12:04:23 +02:00
return &b->metas[i];
Make buffer structure sharable 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
2016-08-03 15:59:17 +02:00
indent
2017-05-26 08:05:01 +02:00
return NULL;
Reorganize serialization code a bit Move the proxy plugin to the client-node Move serialization code to pinos because its specific to pinos Move some functions to the .h files Make the mapper dynamic
2016-10-17 12:20:49 +02:00
}
Work on sink 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.
2016-08-24 16:26:58 +02:00
meta: make safer find_meta_data function
2018-07-09 12:04:23 +02:00
static inline void *spa_buffer_find_meta_data(const struct spa_buffer *b, uint32_t type, size_t size)
{
struct spa_meta *m;
if ((m = spa_buffer_find_meta(b, type)) && m->size >= size)
return m->data;
return NULL;
}
spa: group the spa headers into doxygen modules Subdirectories buffer, control, debug, monitor, pod, support and utils, others are still missing. Headers are grouped either per subdirectory (e.g. buffer/ gets added to group spa_buffer) or per-file (e.g. spa_json is a separate group), whatever seemed like the most sensible approach.
2021-05-21 14:03:07 +10:00
/**
* \}
*/
more hacking Move plugin API to separate directory for now Add ringbuffer and way to get ringbuffer from a port
2016-06-28 12:21:56 +02:00
#ifdef __cplusplus
} /* extern "C" */
#endif
Don't use __ in defines or declarations, it's reserved SPA_TYPE -> SPA_TYPE_INFO for type info strings improve includes Fixes #115
2019-01-14 12:58:23 +01:00
#endif /* SPA_BUFFER_H */
Reference in a new issue Copy permalink