microbit-dal/source/MicroBitMessageBus.cpp

506 lines
14 KiB
C++
Raw Normal View History

/**
* Class definition for a MicroBitMessageBus.
*
* The MicroBitMessageBus handles all messages passed between components.
*/
#include "MicroBit.h"
/**
* Constructor.
* Create a new Message Bus.
*/
MicroBitMessageBus::MicroBitMessageBus()
{
this->listeners = NULL;
this->evt_queue_head = NULL;
this->evt_queue_tail = NULL;
}
/**
* Invokes a callback on a given MicroBitListener
*
* Internal wrapper function, used to enable
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
* parameterised callbacks through the fiber scheduler.
*/
void async_callback(void *param)
{
MicroBitListener *listener = (MicroBitListener *)param;
microbit: Added configurable concurrency modes for MicroBitMessageBus handlers. MessageBus handlers can now have one of four concurrency modes for the eventuality of an event being raised whilst a previous event is still being processed. An additional (optional) parameter is provided to the listen() functions to allow this to be selected on a per event handler basis. The permissable options are: MESSAGE_BUS_LISTENER_REENTRANT: The event handler is fired with the new event, regardless of whether or not a previous event is still be processed by that handler. MESSAGE_BUS_LISTENER_QUEUE_IF_BUSY: The new event is queued until such a time as the previous event has completed execution. The new event is then processed. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_DROP_IF_BUSY: The new event is dropped, and will never be processed the the event handler. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_NONBLOCKING: The event handler is self-declaring that it never blocks. This flag is used purely for optimisation, as it permits direct execution of the event hadnelr without inducing any overhead from the scheduler. In addition, the following minor revisions were made in this release: * Cleanup of the #include dependencies contained in the microbit-dal .h files * Bugfix to the scheduler block on event matching code. * Introduced a MICROBIT_ID_ALERT MessageBus channel, for general purpose eventing using nonces.
2015-09-11 15:39:38 +00:00
// OK, now we need to decide how to behave depending on our configuration.
// If this a fiber f already active within this listener then check our
// configuration to determine the correct course of action.
//
if (listener->flags & MESSAGE_BUS_LISTENER_BUSY)
{
// Drop this event, if that's how we've been configured.
if (listener->flags & MESSAGE_BUS_LISTENER_DROP_IF_BUSY)
return;
// Queue this event up for later, if that's how we've been configured.
if (listener->flags & MESSAGE_BUS_LISTENER_QUEUE_IF_BUSY)
{
listener->queue(listener->evt);
return;
}
}
// Determine the calling convention for the callback, and invoke...
// C++ is really bad at this! Especially as the ARM compiler is yet to support C++ 11 :-/
microbit: Added configurable concurrency modes for MicroBitMessageBus handlers. MessageBus handlers can now have one of four concurrency modes for the eventuality of an event being raised whilst a previous event is still being processed. An additional (optional) parameter is provided to the listen() functions to allow this to be selected on a per event handler basis. The permissable options are: MESSAGE_BUS_LISTENER_REENTRANT: The event handler is fired with the new event, regardless of whether or not a previous event is still be processed by that handler. MESSAGE_BUS_LISTENER_QUEUE_IF_BUSY: The new event is queued until such a time as the previous event has completed execution. The new event is then processed. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_DROP_IF_BUSY: The new event is dropped, and will never be processed the the event handler. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_NONBLOCKING: The event handler is self-declaring that it never blocks. This flag is used purely for optimisation, as it permits direct execution of the event hadnelr without inducing any overhead from the scheduler. In addition, the following minor revisions were made in this release: * Cleanup of the #include dependencies contained in the microbit-dal .h files * Bugfix to the scheduler block on event matching code. * Introduced a MICROBIT_ID_ALERT MessageBus channel, for general purpose eventing using nonces.
2015-09-11 15:39:38 +00:00
// Record that we have a fiber going into this listener...
listener->flags |= MESSAGE_BUS_LISTENER_BUSY;
microbit: Added configurable concurrency modes for MicroBitMessageBus handlers. MessageBus handlers can now have one of four concurrency modes for the eventuality of an event being raised whilst a previous event is still being processed. An additional (optional) parameter is provided to the listen() functions to allow this to be selected on a per event handler basis. The permissable options are: MESSAGE_BUS_LISTENER_REENTRANT: The event handler is fired with the new event, regardless of whether or not a previous event is still be processed by that handler. MESSAGE_BUS_LISTENER_QUEUE_IF_BUSY: The new event is queued until such a time as the previous event has completed execution. The new event is then processed. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_DROP_IF_BUSY: The new event is dropped, and will never be processed the the event handler. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_NONBLOCKING: The event handler is self-declaring that it never blocks. This flag is used purely for optimisation, as it permits direct execution of the event hadnelr without inducing any overhead from the scheduler. In addition, the following minor revisions were made in this release: * Cleanup of the #include dependencies contained in the microbit-dal .h files * Bugfix to the scheduler block on event matching code. * Introduced a MICROBIT_ID_ALERT MessageBus channel, for general purpose eventing using nonces.
2015-09-11 15:39:38 +00:00
while (1)
{
// Firstly, check for a method callback into an object.
if (listener->flags & MESSAGE_BUS_LISTENER_METHOD)
listener->cb_method->fire(listener->evt);
// Now a parameterised C function
else if (listener->flags & MESSAGE_BUS_LISTENER_PARAMETERISED)
listener->cb_param(listener->evt, listener->cb_arg);
// We must have a plain C function
else
listener->cb(listener->evt);
// If there are more events to process, dequeue the next one and process it.
microbit: Added configurable concurrency modes for MicroBitMessageBus handlers. MessageBus handlers can now have one of four concurrency modes for the eventuality of an event being raised whilst a previous event is still being processed. An additional (optional) parameter is provided to the listen() functions to allow this to be selected on a per event handler basis. The permissable options are: MESSAGE_BUS_LISTENER_REENTRANT: The event handler is fired with the new event, regardless of whether or not a previous event is still be processed by that handler. MESSAGE_BUS_LISTENER_QUEUE_IF_BUSY: The new event is queued until such a time as the previous event has completed execution. The new event is then processed. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_DROP_IF_BUSY: The new event is dropped, and will never be processed the the event handler. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_NONBLOCKING: The event handler is self-declaring that it never blocks. This flag is used purely for optimisation, as it permits direct execution of the event hadnelr without inducing any overhead from the scheduler. In addition, the following minor revisions were made in this release: * Cleanup of the #include dependencies contained in the microbit-dal .h files * Bugfix to the scheduler block on event matching code. * Introduced a MICROBIT_ID_ALERT MessageBus channel, for general purpose eventing using nonces.
2015-09-11 15:39:38 +00:00
if ((listener->flags & MESSAGE_BUS_LISTENER_QUEUE_IF_BUSY) && listener->evt_queue)
{
MicroBitEventQueueItem *item = listener->evt_queue;
listener->evt = item->evt;
listener->evt_queue = listener->evt_queue->next;
delete item;
// We spin the scheduler here, to preven any particular event handler from continuously holding onto resources.
schedule();
microbit: Added configurable concurrency modes for MicroBitMessageBus handlers. MessageBus handlers can now have one of four concurrency modes for the eventuality of an event being raised whilst a previous event is still being processed. An additional (optional) parameter is provided to the listen() functions to allow this to be selected on a per event handler basis. The permissable options are: MESSAGE_BUS_LISTENER_REENTRANT: The event handler is fired with the new event, regardless of whether or not a previous event is still be processed by that handler. MESSAGE_BUS_LISTENER_QUEUE_IF_BUSY: The new event is queued until such a time as the previous event has completed execution. The new event is then processed. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_DROP_IF_BUSY: The new event is dropped, and will never be processed the the event handler. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_NONBLOCKING: The event handler is self-declaring that it never blocks. This flag is used purely for optimisation, as it permits direct execution of the event hadnelr without inducing any overhead from the scheduler. In addition, the following minor revisions were made in this release: * Cleanup of the #include dependencies contained in the microbit-dal .h files * Bugfix to the scheduler block on event matching code. * Introduced a MICROBIT_ID_ALERT MessageBus channel, for general purpose eventing using nonces.
2015-09-11 15:39:38 +00:00
}
else
break;
}
// The fiber of exiting... clear our state.
listener->flags &= ~MESSAGE_BUS_LISTENER_BUSY;
}
/**
* Queue the given event for processing at a later time.
* Add the given event at the tail of our queue.
*
* @param The event to queue.
*/
void MicroBitMessageBus::queueEvent(MicroBitEvent &evt)
{
int processingComplete;
// Firstly, process all handler regsitered as URGENT. These pre-empt the queue, and are useful for fast, high priority services.
processingComplete = this->process(evt, MESSAGE_BUS_LISTENER_URGENT);
if (!processingComplete)
{
// We need to queue this event for later processing...
MicroBitEventQueueItem *item = new MicroBitEventQueueItem(evt);
__disable_irq();
if (evt_queue_tail == NULL)
evt_queue_head = evt_queue_tail = item;
else
evt_queue_tail->next = item;
__enable_irq();
}
}
/**
* Extract the next event from the front of the event queue (if present).
* @return
*
* @param The event to queue.
*/
MicroBitEventQueueItem* MicroBitMessageBus::dequeueEvent()
{
MicroBitEventQueueItem *item = NULL;
__disable_irq();
if (evt_queue_head != NULL)
{
item = evt_queue_head;
evt_queue_head = item->next;
if (evt_queue_head == NULL)
evt_queue_tail = NULL;
}
__enable_irq();
return item;
}
/**
* Periodic callback from MicroBit.
* Process at least one event from the event queue, if it is not empty.
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
* We then continue processing events until something appears on the runqueue.
*/
void MicroBitMessageBus::idleTick()
{
MicroBitEventQueueItem *item = this->dequeueEvent();
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
// Whilst there are events to process and we have no useful other work to do, pull them off the queue and process them.
while (item)
{
// send the event to all standard event listeners.
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
this->process(item->evt);
// Free the queue item.
delete item;
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
// If we have created some useful work to do, we stop processing.
// This helps to minimise the number of blocked fibers we create at any point in time, therefore
// also reducing the RAM footprint.
if(!scheduler_runqueue_empty())
break;
// Pull the next event to process, if there is one.
item = this->dequeueEvent();
}
}
/**
* Indicates whether or not we have any background work to do.
* @ return 1 if there are any events waitingto be processed, 0 otherwise.
*/
int MicroBitMessageBus::isIdleCallbackNeeded()
{
return !(evt_queue_head == NULL);
}
/**
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
* Queues the given event to be sent to all registered recipients.
*
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
* @param The event to send.
*
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
* n.b. THIS IS NOW WRAPPED BY THE MicroBitEvent CLASS FOR CONVENIENCE...
*
* Example:
* @code
* MicroBitEvent evt(id,MICROBIT_BUTTON_EVT_DOWN,ticks,CREATE_ONLY);
* evt.fire();
*
* //OR YOU CAN DO THIS...
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
* MicroBitEvent evt(id,MICROBIT_BUTTON_EVT_DOWN);
* @endcode
*/
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
void MicroBitMessageBus::send(MicroBitEvent evt)
{
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
// We simply queue processing of the event until we're scheduled in normal thread context.
// We do this to avoid the possibility of executing event handler code in IRQ context, which may bring
// hidden race conditions to kids code. Queuing all events ensures causal ordering (total ordering in fact).
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
this->queueEvent(evt);
return;
}
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
/*
* Deliver the given event to all registered event handlers.
* Event handlers are called using the invoke() mechanism provided by the fier scheduler
* This will attempt to call the event handler directly, but spawn a fiber should that
* event handler attempt a blocking operation.
* @param evt The event to be delivered.
* @param mask The type of listeners to process (optional). Matches MicroBitListener flags. If not defined, all standard listeners will be processed.
* @return The 1 if all matching listeners were processed, 0 if further processing is required.
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
*/
int MicroBitMessageBus::process(MicroBitEvent &evt, uint32_t mask)
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
{
MicroBitListener *l;
int complete = 1;
microbit: Memory Optimisation Mega Update This release contains a widespread set of updates and optimisations to the micro:bit runtime, with a view to reducing the SRAM footprint of the whole system. This is to provide as much usable HEAP storage for application programs as possible. Specific updates and optimisations include: - Additional compilation flags to allow the core micro:bit runtime to be configured. These are defined in MicroBitConfig.h - A custom heap allocator. This is now included for two reasons: 1) To provide a simple mechanism to to utilise both the mbed heap space and other memory regions (such as unused memory in the SoftDevice region) as a single virtual heap. 2) To address some issues that have been noted that are attributable to heap fragmentation. The micro:bit heap allocator has a simple algorithm, but one that is chosen to respond well to the relativelt high 'heap churn' found in the micro:bit environment. All micro:bit components and user programs now use this heap allocator trasparently. - Updates to BLE services to remove persistent references to their GATT services. This consumes vast amounts SRAM, rather unecessarily. Instead only handles to the relevant GATT characteristics are now stored. This specifically includes: + MicroBitDFUService + MicroBitEventService + DeviceInformationService - Updates to the Fiber scheduler to save SRAM. More specifically: + Removed the need to hold an empty processor context to intialise fibers. + The IDLE fiber now runs without a stack + fiber stacks are now only created when a fiber is descheduled for the first time, thereby reducing heap churn. + the 'main' fiber is now recycled into the fiber_pool if it leaves app_main() + fibers created through invoke() now only maintains the necessary part of teh parent stack that is needed, thereby reducing the stack size of spawned fibers. - Updates to the Message Bus to reduce the overall memory footprint of processing events. More specifically: + Event handlers are now always called using invoke(), such that non-blocking event handlers no longer need a dedicated fiber to execute - thereby saving SRAM and processor time. + Processing of events from the event queue is now rate paced. Events only continue to be processed as long as there are no fibers on the run queue. i.e. event processing is no longer greedy, thereby reducing the number of fibers created on the runqueue. - Updates to BLUEZOENE code to bring up core BLE services even if they are not enabled by default. This allows programs that do not require BLE to operate to benefit from the full range of SRAM, whilst still allowing the device to be programmed over BLE. - Updates to the Soft Device initialisation configuration, reducing the size of the GATT table held in the top 1.8K of its 8K memory region to around 800 bytes. This is sufficient to run the default set of BLE services on the micro:bit so the additional memory is configured as HEAP storage by MicroBitHeapAllocator. - Minor changes to a range of components to integrate with the above changes. + rename of free() to release() in DynamicPWM to avoid namespace collision with MicroBitHeap free() + rename of fork_on_block to invoke() to enhance readbility. - Many code cleanups and updates to out of date comments.
2015-08-31 22:25:10 +00:00
l = listeners;
microbit: Added configurable concurrency modes for MicroBitMessageBus handlers. MessageBus handlers can now have one of four concurrency modes for the eventuality of an event being raised whilst a previous event is still being processed. An additional (optional) parameter is provided to the listen() functions to allow this to be selected on a per event handler basis. The permissable options are: MESSAGE_BUS_LISTENER_REENTRANT: The event handler is fired with the new event, regardless of whether or not a previous event is still be processed by that handler. MESSAGE_BUS_LISTENER_QUEUE_IF_BUSY: The new event is queued until such a time as the previous event has completed execution. The new event is then processed. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_DROP_IF_BUSY: The new event is dropped, and will never be processed the the event handler. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_NONBLOCKING: The event handler is self-declaring that it never blocks. This flag is used purely for optimisation, as it permits direct execution of the event hadnelr without inducing any overhead from the scheduler. In addition, the following minor revisions were made in this release: * Cleanup of the #include dependencies contained in the microbit-dal .h files * Bugfix to the scheduler block on event matching code. * Introduced a MICROBIT_ID_ALERT MessageBus channel, for general purpose eventing using nonces.
2015-09-11 15:39:38 +00:00
while (l != NULL)
{
if((l->id == evt.source || l->id == MICROBIT_ID_ANY) && (l->value == evt.value || l->value == MICROBIT_EVT_ANY))
{
if(l->flags & mask)
{
l->evt = evt;
// OK, if this handler has regisitered itself as non-blocking, we just execute it directly...
// This is normally only done for trusted system components.
// Otherwise, we invoke it in a 'fork on block' context, that will automatically create a fiber
// should the event handler attempt a blocking operation, but doesn't have the overhead
// of creating a fiber needlessly. (cool huh?)
if (l->flags & MESSAGE_BUS_LISTENER_NONBLOCKING)
async_callback(l);
else
invoke(async_callback, l);
}
microbit: Added configurable concurrency modes for MicroBitMessageBus handlers. MessageBus handlers can now have one of four concurrency modes for the eventuality of an event being raised whilst a previous event is still being processed. An additional (optional) parameter is provided to the listen() functions to allow this to be selected on a per event handler basis. The permissable options are: MESSAGE_BUS_LISTENER_REENTRANT: The event handler is fired with the new event, regardless of whether or not a previous event is still be processed by that handler. MESSAGE_BUS_LISTENER_QUEUE_IF_BUSY: The new event is queued until such a time as the previous event has completed execution. The new event is then processed. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_DROP_IF_BUSY: The new event is dropped, and will never be processed the the event handler. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_NONBLOCKING: The event handler is self-declaring that it never blocks. This flag is used purely for optimisation, as it permits direct execution of the event hadnelr without inducing any overhead from the scheduler. In addition, the following minor revisions were made in this release: * Cleanup of the #include dependencies contained in the microbit-dal .h files * Bugfix to the scheduler block on event matching code. * Introduced a MICROBIT_ID_ALERT MessageBus channel, for general purpose eventing using nonces.
2015-09-11 15:39:38 +00:00
else
{
complete = 0;
}
}
l = l->next;
}
return complete;
}
/**
* Register a listener function.
*
* @param id The source of messages to listen for. Events sent from any other IDs will be filtered.
* Use MICROBIT_ID_ANY to receive events from all components.
*
* @param value The value of messages to listen for. Events with any other values will be filtered.
* Use MICROBIT_VALUE_ANY to receive events of any value.
*
* @param hander The function to call when an event is received.
*
* Example:
* @code
* void onButtonBClick(MicroBitEvent evt)
* {
* //do something
* }
* uBit.MessageBus.listen(MICROBIT_ID_BUTTON_B, MICROBIT_BUTTON_EVT_CLICK, onButtonBClick); // call function when ever a click event is detected.
* @endcode
*/
microbit: Added configurable concurrency modes for MicroBitMessageBus handlers. MessageBus handlers can now have one of four concurrency modes for the eventuality of an event being raised whilst a previous event is still being processed. An additional (optional) parameter is provided to the listen() functions to allow this to be selected on a per event handler basis. The permissable options are: MESSAGE_BUS_LISTENER_REENTRANT: The event handler is fired with the new event, regardless of whether or not a previous event is still be processed by that handler. MESSAGE_BUS_LISTENER_QUEUE_IF_BUSY: The new event is queued until such a time as the previous event has completed execution. The new event is then processed. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_DROP_IF_BUSY: The new event is dropped, and will never be processed the the event handler. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_NONBLOCKING: The event handler is self-declaring that it never blocks. This flag is used purely for optimisation, as it permits direct execution of the event hadnelr without inducing any overhead from the scheduler. In addition, the following minor revisions were made in this release: * Cleanup of the #include dependencies contained in the microbit-dal .h files * Bugfix to the scheduler block on event matching code. * Introduced a MICROBIT_ID_ALERT MessageBus channel, for general purpose eventing using nonces.
2015-09-11 15:39:38 +00:00
void MicroBitMessageBus::listen(int id, int value, void (*handler)(MicroBitEvent), uint16_t flags)
{
if (handler == NULL)
return;
microbit: Added configurable concurrency modes for MicroBitMessageBus handlers. MessageBus handlers can now have one of four concurrency modes for the eventuality of an event being raised whilst a previous event is still being processed. An additional (optional) parameter is provided to the listen() functions to allow this to be selected on a per event handler basis. The permissable options are: MESSAGE_BUS_LISTENER_REENTRANT: The event handler is fired with the new event, regardless of whether or not a previous event is still be processed by that handler. MESSAGE_BUS_LISTENER_QUEUE_IF_BUSY: The new event is queued until such a time as the previous event has completed execution. The new event is then processed. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_DROP_IF_BUSY: The new event is dropped, and will never be processed the the event handler. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_NONBLOCKING: The event handler is self-declaring that it never blocks. This flag is used purely for optimisation, as it permits direct execution of the event hadnelr without inducing any overhead from the scheduler. In addition, the following minor revisions were made in this release: * Cleanup of the #include dependencies contained in the microbit-dal .h files * Bugfix to the scheduler block on event matching code. * Introduced a MICROBIT_ID_ALERT MessageBus channel, for general purpose eventing using nonces.
2015-09-11 15:39:38 +00:00
MicroBitListener *newListener = new MicroBitListener(id, value, handler, flags);
if(!add(newListener))
delete newListener;
}
microbit: Added configurable concurrency modes for MicroBitMessageBus handlers. MessageBus handlers can now have one of four concurrency modes for the eventuality of an event being raised whilst a previous event is still being processed. An additional (optional) parameter is provided to the listen() functions to allow this to be selected on a per event handler basis. The permissable options are: MESSAGE_BUS_LISTENER_REENTRANT: The event handler is fired with the new event, regardless of whether or not a previous event is still be processed by that handler. MESSAGE_BUS_LISTENER_QUEUE_IF_BUSY: The new event is queued until such a time as the previous event has completed execution. The new event is then processed. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_DROP_IF_BUSY: The new event is dropped, and will never be processed the the event handler. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_NONBLOCKING: The event handler is self-declaring that it never blocks. This flag is used purely for optimisation, as it permits direct execution of the event hadnelr without inducing any overhead from the scheduler. In addition, the following minor revisions were made in this release: * Cleanup of the #include dependencies contained in the microbit-dal .h files * Bugfix to the scheduler block on event matching code. * Introduced a MICROBIT_ID_ALERT MessageBus channel, for general purpose eventing using nonces.
2015-09-11 15:39:38 +00:00
void MicroBitMessageBus::listen(int id, int value, void (*handler)(MicroBitEvent, void*), void* arg, uint16_t flags)
{
if (handler == NULL)
return;
microbit: Added configurable concurrency modes for MicroBitMessageBus handlers. MessageBus handlers can now have one of four concurrency modes for the eventuality of an event being raised whilst a previous event is still being processed. An additional (optional) parameter is provided to the listen() functions to allow this to be selected on a per event handler basis. The permissable options are: MESSAGE_BUS_LISTENER_REENTRANT: The event handler is fired with the new event, regardless of whether or not a previous event is still be processed by that handler. MESSAGE_BUS_LISTENER_QUEUE_IF_BUSY: The new event is queued until such a time as the previous event has completed execution. The new event is then processed. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_DROP_IF_BUSY: The new event is dropped, and will never be processed the the event handler. This option does not preclude the processing of the new event by other event handlers. MESSAGE_BUS_LISTENER_NONBLOCKING: The event handler is self-declaring that it never blocks. This flag is used purely for optimisation, as it permits direct execution of the event hadnelr without inducing any overhead from the scheduler. In addition, the following minor revisions were made in this release: * Cleanup of the #include dependencies contained in the microbit-dal .h files * Bugfix to the scheduler block on event matching code. * Introduced a MICROBIT_ID_ALERT MessageBus channel, for general purpose eventing using nonces.
2015-09-11 15:39:38 +00:00
MicroBitListener *newListener = new MicroBitListener(id, value, handler, arg, flags);
if(!add(newListener))
delete newListener;
}
/**
* Unregister a listener function.
* Listners are identified by the Event ID, Event VALUE and handler registered using listen().
*
* @param id The Event ID used to register the listener.
* @param value The Event VALUE used to register the listener.
* @param handler The function used to register the listener.
*
*
* Example:
* @code
* void onButtonBClick()
* {
* //do something
* }
*
* uBit.MessageBus.ignore(MICROBIT_ID_BUTTON_B, MICROBIT_BUTTON_EVT_CLICK, onButtonBClick);
* @endcode
*/
void MicroBitMessageBus::ignore(int id, int value, void (*handler)(MicroBitEvent))
{
if (handler == NULL)
return;
MicroBitListener listener(id, value, handler);
remove(&listener);
}
/**
* Unregister a listener function.
* Listners are identified by the Event ID, Event VALUE and handler registered using listen().
*
* @param id The Event ID used to register the listener.
* @param value The Event VALUE used to register the listener.
* @param handler The function used to register the listener.
*
*
* Example:
* @code
* void onButtonBClick(void *arg)
* {
* //do something
* }
*
* uBit.MessageBus.ignore(MICROBIT_ID_BUTTON_B, MICROBIT_BUTTON_EVT_CLICK, onButtonBClick);
* @endcode
*/
void MicroBitMessageBus::ignore(int id, int value, void (*handler)(MicroBitEvent, void*))
{
if (handler == NULL)
return;
// The remove function is not comparing the [arg] anyhow.
MicroBitListener listener(id, value, handler, NULL);
remove(&listener);
}
/**
* Add the given MicroBitListener to the list of event handlers, unconditionally.
* @param listener The MicroBitListener to validate.
* @return 1 if the listener is valid, 0 otherwise.
*/
int MicroBitMessageBus::add(MicroBitListener *newListener)
{
MicroBitListener *l, *p;
int methodCallback;
//handler can't be NULL!
if (newListener == NULL)
return 0;
l = listeners;
// Firstly, we treat a listener as an idempotent operation. Ensure we don't already have this handler
// registered in a that will already capture these events. If we do, silently ignore.
// We always check the ID, VALUE and CB_METHOD fields.
// If we have a callback to a method, check the cb_method class. Otherwise, the cb function point is sufficient.
while (l != NULL)
{
methodCallback = (newListener->flags & MESSAGE_BUS_LISTENER_METHOD) && (l->flags & MESSAGE_BUS_LISTENER_METHOD);
if (l->id == newListener->id && l->value == newListener->value && (methodCallback ? *l->cb_method == *newListener->cb_method : l->cb == newListener->cb))
return 0;
l = l->next;
}
// We have a valid, new event handler. Add it to the list.
// if listeners is null - we can automatically add this listener to the list at the beginning...
if (listeners == NULL)
{
listeners = newListener;
return 1;
}
// We maintain an ordered list of listeners.
// The chain is held stictly in increasing order of ID (first level), then value code (second level).
// Find the correct point in the chain for this event.
// Adding a listener is a rare occurance, so we just walk the list...
p = listeners;
l = listeners;
while (l != NULL && l->id < newListener->id)
{
p = l;
l = l->next;
}
while (l != NULL && l->id == newListener->id && l->value < newListener->value)
{
p = l;
l = l->next;
}
//add at front of list
if (p == listeners && (newListener->id < p->id || (p->id == newListener->id && p->value > newListener->value)))
{
newListener->next = p;
//this new listener is now the front!
listeners = newListener;
}
//add after p
else
{
newListener->next = p->next;
p->next = newListener;
}
return 1;
}
/**
* Remove a MicroBitListener from the list that matches the given listener.
* @param listener The MicroBitListener to validate.
* @return The number of listeners removed from the list.
*/
int MicroBitMessageBus::remove(MicroBitListener *listener)
{
MicroBitListener *l, *p;
int removed = 0;
//handler can't be NULL!
if (listener == NULL)
return 0;
l = listeners;
p = NULL;
// Walk this list of event handlers. Delete any that match the given listener.
while (l != NULL)
{
if ((listener->flags & MESSAGE_BUS_LISTENER_METHOD) == (l->flags & MESSAGE_BUS_LISTENER_METHOD))
{
if(((listener->flags & MESSAGE_BUS_LISTENER_METHOD) && (*l->cb_method == *listener->cb_method)) ||
((!(listener->flags & MESSAGE_BUS_LISTENER_METHOD) && l->cb == listener->cb)))
{
if ((listener->id == MICROBIT_ID_ANY || listener->id == l->id) && (listener->value == MICROBIT_EVT_ANY || listener->value == l->value))
{
// Found a match. Remove from the list.
if (p == NULL)
listeners = l->next;
else
p->next = l->next;
// delete the listener.
MicroBitListener *t = l;
l = l->next;
delete t;
removed++;
continue;
}
}
}
p = l;
l = l->next;
}
return removed;
}
/**
* Returns the microBitListener with the given position in our list.
* @param n The position in the list to return.
* @return the MicroBitListener at postion n in the list, or NULL if the position is invalid.
*/
MicroBitListener* MicroBitMessageBus::elementAt(int n)
{
MicroBitListener *l = listeners;
while (n > 0)
{
if (l == NULL)
return NULL;
n--;
l = l->next;
}
return l;
}