Skip to content

iox::popo::ChunkDistributor🔗

The ChunkDistributor is the low layer building block to send SharedChunks to a dynamic number of ChunkQueus. Together with the ChunkQueuePusher, the ChunkDistributor builds the infrastructure to exchange memory chunks between different data producers and consumers that could be located in different processes. Besides a modifiable container of ChunkQueues to which a SharedChunk can be deliverd, it holds a configurable history of last sent chunks. This allows to provide a newly added queue a number of last chunks to start from. This is needed for functionality known as latched topic in ROS or field in ara::com. A ChunkDistributor is used to build elements of higher abstraction layers that also do memory managemet and provide an API towards the real user. More...

#include <chunk_distributor.hpp>

Public Types🔗

Name
using ChunkDistributorDataType MemberType_t
using typename ChunkDistributorDataType::ChunkQueueData_t ChunkQueueData_t
using typename ChunkDistributorDataType::ChunkQueuePusher_t ChunkQueuePusher_t

Public Functions🔗

Name
ChunkDistributor(cxx::not_null< MemberType_t *const > chunkDistrubutorDataPtr)
ChunkDistributor(const ChunkDistributor & other)
ChunkDistributor & operator=(const ChunkDistributor & )
ChunkDistributor(ChunkDistributor && rhs) =default
ChunkDistributor & operator=(ChunkDistributor && rhs) =default
virtual ~ChunkDistributor() =default
cxx::expected< ChunkDistributorError > tryAddQueue(cxx::not_null< ChunkQueueData_t *const > queueToAdd, const uint64_t requestedHistory =0u)
Add a queue to the internal list of chunk queues to which chunks are delivered when calling deliverToAllStoredQueues.
cxx::expected< ChunkDistributorError > tryRemoveQueue(cxx::not_null< ChunkQueueData_t *const > queueToRemove)
Remove a queue from the internal list of chunk queues.
void removeAllQueues()
Delete all the stored chunk queues.
bool hasStoredQueues() const
Get the information whether there are any stored chunk queues.
void deliverToAllStoredQueues(mepoo::SharedChunk chunk)
Deliver the provided shared chunk to all the stored chunk queues. The chunk will be added to the chunk history.
bool deliverToQueue(cxx::not_null< ChunkQueueData_t *const > queue, mepoo::SharedChunk chunk)
Deliver the provided shared chunk to the provided chunk queue. The chunk will NOT be added to the chunk history.
void addToHistoryWithoutDelivery(mepoo::SharedChunk chunk)
Update the chunk history but do not deliver the chunk to any chunk queue. E.g. use case is to to update a non offered field in ara.
uint64_t getHistorySize()
Get the current size of the chunk history.
uint64_t getHistoryCapacity() const
Get the capacity of the chunk history.
void clearHistory()
Clears the chunk history.
void cleanup()
cleanup the used shrared memory chunks

Protected Functions🔗

Name
const MemberType_t * getMembers() const
MemberType_t * getMembers()

Detailed Description🔗

template <typename ChunkDistributorDataType >
class iox::popo::ChunkDistributor;

The ChunkDistributor is the low layer building block to send SharedChunks to a dynamic number of ChunkQueus. Together with the ChunkQueuePusher, the ChunkDistributor builds the infrastructure to exchange memory chunks between different data producers and consumers that could be located in different processes. Besides a modifiable container of ChunkQueues to which a SharedChunk can be deliverd, it holds a configurable history of last sent chunks. This allows to provide a newly added queue a number of last chunks to start from. This is needed for functionality known as latched topic in ROS or field in ara::com. A ChunkDistributor is used to build elements of higher abstraction layers that also do memory managemet and provide an API towards the real user.

Todo: There are currently some challenge: For the stored queues and the history, containers are used which are not thread safe. Therefore we use an inter-process mutex. But this can lead to deadlocks if a user process gets terminated while one of its threads is in the ChunkDistributor and holds a lock. An easier setup would be if changing the queues by a middleware thread and sending chunks by the user process would not interleave. I.e. there is no concurrent access to the containers. Then a memory synchronization would be sufficient. The cleanup() call is the biggest challenge. This is used to free chunks that are still held by a not properly terminated user application. Even if access from middleware and user threads do not overlap, the history container to cleanup could be in an inconsistent state as the application was hard terminated while changing it. We would need a container like the UsedChunkList to have one that is robust against such inconsistencies.... A perfect job for our future selves

About Concurrency: This ChunkDistributor can be used with different LockingPolicies for different scenarios When different threads operate on it (e.g. application sends chunks and RouDi adds and removes queues), a locking policy must be used that ensures consistent data in the ChunkDistributorData.

Public Types Documentation🔗

using MemberType_t🔗

using iox::popo::ChunkDistributor< ChunkDistributorDataType >::MemberType_t =  ChunkDistributorDataType;

using ChunkQueueData_t🔗

using iox::popo::ChunkDistributor< ChunkDistributorDataType >::ChunkQueueData_t =  typename ChunkDistributorDataType::ChunkQueueData_t;

using ChunkQueuePusher_t🔗

using iox::popo::ChunkDistributor< ChunkDistributorDataType >::ChunkQueuePusher_t =  typename ChunkDistributorDataType::ChunkQueuePusher_t;

Public Functions Documentation🔗

function ChunkDistributor🔗

inline explicit ChunkDistributor(
    cxx::not_null< MemberType_t *const > chunkDistrubutorDataPtr
)

function ChunkDistributor🔗

ChunkDistributor(
    const ChunkDistributor & other
)

function operator=🔗

ChunkDistributor & operator=(
    const ChunkDistributor & 
)

function ChunkDistributor🔗

ChunkDistributor(
    ChunkDistributor && rhs
) =default

function operator=🔗

ChunkDistributor & operator=(
    ChunkDistributor && rhs
) =default

function ~ChunkDistributor🔗

virtual ~ChunkDistributor() =default

function tryAddQueue🔗

inline cxx::expected< ChunkDistributorError > tryAddQueue(
    cxx::not_null< ChunkQueueData_t *const > queueToAdd,
    const uint64_t requestedHistory =0u
)

Add a queue to the internal list of chunk queues to which chunks are delivered when calling deliverToAllStoredQueues.

Parameters:

  • queueToAdd chunk queue to add to the list
  • requestedHistory number of last chunks from history to send if available. If history size is smaller then the available history size chunks are provided

Return: if the queue could be added it returns success, otherwiese a ChunkDistributor error

function tryRemoveQueue🔗

inline cxx::expected< ChunkDistributorError > tryRemoveQueue(
    cxx::not_null< ChunkQueueData_t *const > queueToRemove
)

Remove a queue from the internal list of chunk queues.

Parameters:

  • chunk queue to remove from the list

Return: if the queue could be removed it returns success, otherwiese a ChunkDistributor error

function removeAllQueues🔗

inline void removeAllQueues()

Delete all the stored chunk queues.

function hasStoredQueues🔗

inline bool hasStoredQueues() const

Get the information whether there are any stored chunk queues.

Return: true if there are stored chunk queues, false if not

function deliverToAllStoredQueues🔗

inline void deliverToAllStoredQueues(
    mepoo::SharedChunk chunk
)

Deliver the provided shared chunk to all the stored chunk queues. The chunk will be added to the chunk history.

Parameters:

  • shared chunk to be delivered

function deliverToQueue🔗

inline bool deliverToQueue(
    cxx::not_null< ChunkQueueData_t *const > queue,
    mepoo::SharedChunk chunk
)

Deliver the provided shared chunk to the provided chunk queue. The chunk will NOT be added to the chunk history.

Parameters:

  • chunk queue to which this chunk shall be delivered
  • shared chunk to be delivered

Return: false if a queue overflow occured, otherwise true

function addToHistoryWithoutDelivery🔗

inline void addToHistoryWithoutDelivery(
    mepoo::SharedChunk chunk
)

Update the chunk history but do not deliver the chunk to any chunk queue. E.g. use case is to to update a non offered field in ara.

Parameters:

  • shared chunk add to the chunk history

function getHistorySize🔗

inline uint64_t getHistorySize()

Get the current size of the chunk history.

Return: chunk history size

function getHistoryCapacity🔗

inline uint64_t getHistoryCapacity() const

Get the capacity of the chunk history.

Return: chunk history capacity

function clearHistory🔗

inline void clearHistory()

Clears the chunk history.

function cleanup🔗

inline void cleanup()

cleanup the used shrared memory chunks

Todocurrently we have a deadlock / mutex destroy vulnerability if the ThreadSafePolicy is used and a sending application dies when having the lock for sending. If the RouDi daemon wants to cleanup or does discovery changes we have a deadlock or an exception when destroying the mutex As long as we don't have a multi-threaded lock-free ChunkDistributor or another concept we die here

Protected Functions Documentation🔗

function getMembers🔗

inline const MemberType_t * getMembers() const

function getMembers🔗

MemberType_t * getMembers()

Updated on 26 April 2021 at 15:31:02 CEST