Schoumi/microbit-dal
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity
microbit-dal/source/MicroBitHeapAllocator.cpp

404 lines
12 KiB
C++
Raw Normal View History

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
#include "MicroBit.h"
/**
* A simple 32 bit block based memory allocator. This allows one or more memory segments to
* be designated as heap storage, and is designed to run in a static memory area or inside the standard C
* heap for use by the micro:bit runtime. This is required for several reasons:
*
* 1) It reduces memory fragmentation due to the high churn sometime placed on the heap
* by ManagedTypes, fibers and user code. Underlying heap implentations are often have very simplistic
* allocation pilicies and suffer from fragmentation in prolonged use - which can cause programs to
* stop working after a period of time. The algorithm implemented here is simple, but highly tolerant to
* large amounts of churn.
*
* 2) It allows us to reuse the 8K of SRAM set aside for SoftDevice as additional heap storage
* when BLE is not in use.
*
* 3) It gives a simple example of how memory allocation works! :-)
*
* N.B. The need for this should be reviewed in the future, should a different memory allocator be
* made availiable in the mbed platform.
*
* P.S. This is a very simple allocator, therefore not without its weaknesses. Why don't you consider
* what these are, and consider the tradeoffs against simplicity...
*
* TODO: Consider caching recently freed blocks to improve allocation time.
*/
struct HeapDefinition
{
uint32_t *heap_start; // Physical address of the start of this heap.
uint32_t *heap_end; // Physical address of the end of this heap.
};
// Create the necessary heap definitions.
// We use two heaps by default: one for SoftDevice reuse, and one to run inside the mbed heap.
Silence GCC -Wall
2015-10-27 22:17:04 +00:00
HeapDefinition heap[MICROBIT_HEAP_COUNT] = { };
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
microbit: Conditional compilation on MicroBitConfig All options defined in MicroBitConfig are now set IFF they already do not have a value. This allows external configuration of these options much more cleanly. Additional macros also defined to provide clean implementation for boolean configuration switches. Also some minor bugfixes: - Correction of typo of "Microbit.h" => "MicroBit.h" - Fixed heap corruption bug on gcc as a result of a debug printf in an unsafe context.
2015-09-02 11:42:24 +00:00
// Scans the status of the heap definition table, and returns the number of INITIALISED heaps.
int microbit_active_heaps()
{
int heapCount = 0;
for (int i=0; i < MICROBIT_HEAP_COUNT; i++)
{
if(heap[i].heap_start != NULL)
heapCount++;
}
return heapCount;
}
#if CONFIG_ENABLED(MICROBIT_DBG) && CONFIG_ENABLED(MICROBIT_HEAP_DBG)
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
// Internal diagnostics function.
// Diplays a usage summary about a given heap...
void microbit_heap_print(HeapDefinition &heap)
{
uint32_t blockSize;
uint32_t *block;
int totalFreeBlock = 0;
int totalUsedBlock = 0;
int cols = 0;
if (heap.heap_start == NULL)
{
microbit: Removal of explicit Serial instance for debugging. Debugging now runs via the standrad MicroBit::serial interface.
2015-11-01 23:42:53 +00:00
uBit.serial.printf("--- HEAP NOT INITIALISED ---\n");
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
return;
}
microbit: Removal of explicit Serial instance for debugging. Debugging now runs via the standrad MicroBit::serial interface.
2015-11-01 23:42:53 +00:00
uBit.serial.printf("heap_start : %p\n", heap.heap_start);
uBit.serial.printf("heap_end : %p\n", heap.heap_end);
uBit.serial.printf("heap_size : %d\n", (int)heap.heap_end - (int)heap.heap_start);
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
microbit: Conditional compilation on MicroBitConfig All options defined in MicroBitConfig are now set IFF they already do not have a value. This allows external configuration of these options much more cleanly. Additional macros also defined to provide clean implementation for boolean configuration switches. Also some minor bugfixes: - Correction of typo of "Microbit.h" => "MicroBit.h" - Fixed heap corruption bug on gcc as a result of a debug printf in an unsafe context.
2015-09-02 11:42:24 +00:00
// Disable IRQ temporarily to ensure no race conditions!
__disable_irq();
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
block = heap.heap_start;
while (block < heap.heap_end)
{
blockSize = *block & ~MICROBIT_HEAP_BLOCK_FREE;
minor bugfixes in MICROBIT_HEAP_DEBUG
2015-11-17 16:58:42 +00:00
uBit.serial.printf("[%C:%d] ", *block & MICROBIT_HEAP_BLOCK_FREE ? 'F' : 'U', blockSize*4);
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 (cols++ == 20)
{
microbit: Removal of explicit Serial instance for debugging. Debugging now runs via the standrad MicroBit::serial interface.
2015-11-01 23:42:53 +00:00
uBit.serial.printf("\n");
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
cols = 0;
}
if (*block & MICROBIT_HEAP_BLOCK_FREE)
totalFreeBlock += blockSize;
else
totalUsedBlock += blockSize;
block += blockSize;
}
microbit: Conditional compilation on MicroBitConfig All options defined in MicroBitConfig are now set IFF they already do not have a value. This allows external configuration of these options much more cleanly. Additional macros also defined to provide clean implementation for boolean configuration switches. Also some minor bugfixes: - Correction of typo of "Microbit.h" => "MicroBit.h" - Fixed heap corruption bug on gcc as a result of a debug printf in an unsafe context.
2015-09-02 11:42:24 +00:00
// Enable Interrupts
__enable_irq();
microbit: Removal of explicit Serial instance for debugging. Debugging now runs via the standrad MicroBit::serial interface.
2015-11-01 23:42:53 +00:00
uBit.serial.printf("\n");
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
microbit: Removal of explicit Serial instance for debugging. Debugging now runs via the standrad MicroBit::serial interface.
2015-11-01 23:42:53 +00:00
uBit.serial.printf("mb_total_free : %d\n", totalFreeBlock*4);
uBit.serial.printf("mb_total_used : %d\n", totalUsedBlock*4);
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
}
// Internal diagnostics function.
// Diplays a usage summary about all known heaps...
void microbit_heap_print()
{
for (int i=0; i < MICROBIT_HEAP_COUNT; i++)
{
microbit: Removal of explicit Serial instance for debugging. Debugging now runs via the standrad MicroBit::serial interface.
2015-11-01 23:42:53 +00:00
uBit.serial.printf("\nHEAP %d: \n", i);
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
microbit_heap_print(heap[i]);
}
}
#endif
void microbit_initialise_heap(HeapDefinition &heap)
{
// Simply mark the entire heap as free.
*heap.heap_start = ((uint32_t) heap.heap_end - (uint32_t) heap.heap_start) / MICROBIT_HEAP_BLOCK_SIZE;
*heap.heap_start |= MICROBIT_HEAP_BLOCK_FREE;
}
int
microbit_create_sd_heap(HeapDefinition &heap)
{
microbot: Cleanup of #ifdef in MicroBitHeapAllocator Removed compiler warning if MICROBIT_HEAP_REUSE_SD is not defined.
2015-10-09 15:58:10 +00:00
#if CONFIG_ENABLED(MICROBIT_HEAP_REUSE_SD)
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
microbit: Conditional compilation on MicroBitConfig All options defined in MicroBitConfig are now set IFF they already do not have a value. This allows external configuration of these options much more cleanly. Additional macros also defined to provide clean implementation for boolean configuration switches. Also some minor bugfixes: - Correction of typo of "Microbit.h" => "MicroBit.h" - Fixed heap corruption bug on gcc as a result of a debug printf in an unsafe context.
2015-09-02 11:42:24 +00:00
#if CONFIG_ENABLED(MICROBIT_BLE_ENABLED)
Updates to DFU service to align with BLE standard BLE security - Removed flashcode base authorization - silenced unused parameter in MicroBitHeapAllocator - bugfixed BLUEZONE mode so that sensor data is still available via BLE
2015-11-20 11:32:42 +00:00
// Reclaim RAM from unusused areas on the BLE stack GATT table.
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
heap.heap_start = (uint32_t *)MICROBIT_HEAP_BASE_BLE_ENABLED;
heap.heap_end = (uint32_t *)MICROBIT_HEAP_SD_LIMIT;
#else
Updates to DFU service to align with BLE standard BLE security - Removed flashcode base authorization - silenced unused parameter in MicroBitHeapAllocator - bugfixed BLUEZONE mode so that sensor data is still available via BLE
2015-11-20 11:32:42 +00:00
// Reclaim all the RAM normally reserved for the Nordic SoftDevice.
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
heap.heap_start = (uint32_t *)MICROBIT_HEAP_BASE_BLE_DISABLED;
heap.heap_end = (uint32_t *)MICROBIT_HEAP_SD_LIMIT;
#endif
microbit_initialise_heap(heap);
microbot: Cleanup of #ifdef in MicroBitHeapAllocator Removed compiler warning if MICROBIT_HEAP_REUSE_SD is not defined.
2015-10-09 15:58:10 +00:00
#else
Updates to DFU service to align with BLE standard BLE security - Removed flashcode base authorization - silenced unused parameter in MicroBitHeapAllocator - bugfixed BLUEZONE mode so that sensor data is still available via BLE
2015-11-20 11:32:42 +00:00
heap.heap_start = 0;
heap.heap_end = 0;
microbot: Cleanup of #ifdef in MicroBitHeapAllocator Removed compiler warning if MICROBIT_HEAP_REUSE_SD is not defined.
2015-10-09 15:58:10 +00:00
#endif
Updates to DFU service to align with BLE standard BLE security - Removed flashcode base authorization - silenced unused parameter in MicroBitHeapAllocator - bugfixed BLUEZONE mode so that sensor data is still available via BLE
2015-11-20 11:32:42 +00:00
return MICROBIT_OK;
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
microbit_create_nested_heap(HeapDefinition &heap)
{
uint32_t mb_heap_max;
microbit: Conditional compilation on MicroBitConfig All options defined in MicroBitConfig are now set IFF they already do not have a value. This allows external configuration of these options much more cleanly. Additional macros also defined to provide clean implementation for boolean configuration switches. Also some minor bugfixes: - Correction of typo of "Microbit.h" => "MicroBit.h" - Fixed heap corruption bug on gcc as a result of a debug printf in an unsafe context.
2015-09-02 11:42:24 +00:00
void *p;
// Ensure we're configured to use this heap at all. If not, we can safely return.
if (MICROBIT_HEAP_SIZE <= 0)
microbit: First cut at updates to provide tight validation within the runtime code, with a view to moving NOP/clamping behaviour out of teh micro:bit runtime and into the glue layers of higher level languages. - Updates to many functions to provide explicut return codes. - Updates to many functions to remove heuristic calidation (NOP/clamping/defaults) - Updates to ErrorNo.h to provide clearer return values, and place return values in a better scope - Updates to MicroBitDisplay to use enums where appropriate.
2015-10-25 21:51:33 +00:00
return MICROBIT_INVALID_PARAMETER;
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
// Snapshot something at the top of the main heap.
microbit: Conditional compilation on MicroBitConfig All options defined in MicroBitConfig are now set IFF they already do not have a value. This allows external configuration of these options much more cleanly. Additional macros also defined to provide clean implementation for boolean configuration switches. Also some minor bugfixes: - Correction of typo of "Microbit.h" => "MicroBit.h" - Fixed heap corruption bug on gcc as a result of a debug printf in an unsafe context.
2015-09-02 11:42:24 +00:00
p = native_malloc(sizeof(uint32_t));
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
// Compute the size left in our heap, taking care to ensure it lands on a word boundary.
mb_heap_max = (uint32_t) (((float)(MICROBIT_HEAP_END - (uint32_t)p)) * MICROBIT_HEAP_SIZE);
mb_heap_max &= 0xFFFFFFFC;
// Release our reference pointer.
native_free(p);
// Allocate memory for our heap.
// We do this iteratively, as some build configurations seem to have static limits
// on heap size... This allows us to be keep going anyway!
while (heap.heap_start == NULL)
{
heap.heap_start = (uint32_t *)native_malloc(mb_heap_max);
if (heap.heap_start == NULL)
{
mb_heap_max -= 32;
if (mb_heap_max <= 0)
microbit: First cut at updates to provide tight validation within the runtime code, with a view to moving NOP/clamping behaviour out of teh micro:bit runtime and into the glue layers of higher level languages. - Updates to many functions to provide explicut return codes. - Updates to many functions to remove heuristic calidation (NOP/clamping/defaults) - Updates to ErrorNo.h to provide clearer return values, and place return values in a better scope - Updates to MicroBitDisplay to use enums where appropriate.
2015-10-25 21:51:33 +00:00
return MICROBIT_NO_RESOURCES;
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
}
}
heap.heap_end = heap.heap_start + mb_heap_max / MICROBIT_HEAP_BLOCK_SIZE;
microbit_initialise_heap(heap);
microbit: First cut at updates to provide tight validation within the runtime code, with a view to moving NOP/clamping behaviour out of teh micro:bit runtime and into the glue layers of higher level languages. - Updates to many functions to provide explicut return codes. - Updates to many functions to remove heuristic calidation (NOP/clamping/defaults) - Updates to ErrorNo.h to provide clearer return values, and place return values in a better scope - Updates to MicroBitDisplay to use enums where appropriate.
2015-10-25 21:51:33 +00:00
return MICROBIT_OK;
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
}
/**
* Initialise the microbit heap according to the parameters defined in MicroBitConfig.h
* After this is called, any future calls to malloc, new, free or delete will use the new heap.
* n.b. only code that #includes MicroBitHeapAllocator.h will use this heap. This includes all micro:bit runtime
* code, and user code targetting the runtime. External code can choose to include this file, or
microbit: First cut at updates to provide tight validation within the runtime code, with a view to moving NOP/clamping behaviour out of teh micro:bit runtime and into the glue layers of higher level languages. - Updates to many functions to provide explicut return codes. - Updates to many functions to remove heuristic calidation (NOP/clamping/defaults) - Updates to ErrorNo.h to provide clearer return values, and place return values in a better scope - Updates to MicroBitDisplay to use enums where appropriate.
2015-10-25 21:51:33 +00:00
* simply use the standard mbed heap.
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
microbit_heap_init()
{
microbit: First cut at updates to provide tight validation within the runtime code, with a view to moving NOP/clamping behaviour out of teh micro:bit runtime and into the glue layers of higher level languages. - Updates to many functions to provide explicut return codes. - Updates to many functions to remove heuristic calidation (NOP/clamping/defaults) - Updates to ErrorNo.h to provide clearer return values, and place return values in a better scope - Updates to MicroBitDisplay to use enums where appropriate.
2015-10-25 21:51:33 +00:00
int result;
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
// Disable IRQ temporarily to ensure no race conditions!
__disable_irq();
microbit: First cut at updates to provide tight validation within the runtime code, with a view to moving NOP/clamping behaviour out of teh micro:bit runtime and into the glue layers of higher level languages. - Updates to many functions to provide explicut return codes. - Updates to many functions to remove heuristic calidation (NOP/clamping/defaults) - Updates to ErrorNo.h to provide clearer return values, and place return values in a better scope - Updates to MicroBitDisplay to use enums where appropriate.
2015-10-25 21:51:33 +00:00
result = microbit_create_nested_heap(heap[0]);
if (result != MICROBIT_OK)
{
__enable_irq();
return MICROBIT_NO_RESOURCES;
}
result = microbit_create_sd_heap(heap[1]);
if (result != MICROBIT_OK)
{
__enable_irq();
return MICROBIT_NO_RESOURCES;
}
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
// Enable Interrupts
__enable_irq();
microbit: Conditional compilation on MicroBitConfig All options defined in MicroBitConfig are now set IFF they already do not have a value. This allows external configuration of these options much more cleanly. Additional macros also defined to provide clean implementation for boolean configuration switches. Also some minor bugfixes: - Correction of typo of "Microbit.h" => "MicroBit.h" - Fixed heap corruption bug on gcc as a result of a debug printf in an unsafe context.
2015-09-02 11:42:24 +00:00
#if CONFIG_ENABLED(MICROBIT_DBG) && CONFIG_ENABLED(MICROBIT_HEAP_DBG)
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
microbit_heap_print();
#endif
microbit: First cut at updates to provide tight validation within the runtime code, with a view to moving NOP/clamping behaviour out of teh micro:bit runtime and into the glue layers of higher level languages. - Updates to many functions to provide explicut return codes. - Updates to many functions to remove heuristic calidation (NOP/clamping/defaults) - Updates to ErrorNo.h to provide clearer return values, and place return values in a better scope - Updates to MicroBitDisplay to use enums where appropriate.
2015-10-25 21:51:33 +00:00
return MICROBIT_OK;
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
}
/**
* Attempt to allocate a given amount of memory from the given heap.
* @param size The amount of memory, in bytes, to allocate.
* @param heap The heap the memory is to be allocated from.
* @return A pointer to the allocated memory, or NULL if insufficient memory is available.
*/
void *microbit_malloc(size_t size, HeapDefinition &heap)
{
uint32_t blockSize = 0;
uint32_t blocksNeeded = size % MICROBIT_HEAP_BLOCK_SIZE == 0 ? size / MICROBIT_HEAP_BLOCK_SIZE : size / MICROBIT_HEAP_BLOCK_SIZE + 1;
uint32_t *block;
uint32_t *next;
if (size <= 0)
return NULL;
// Account for the index block;
blocksNeeded++;
// Disable IRQ temporarily to ensure no race conditions!
__disable_irq();
// We implement a first fit algorithm with cache to handle rapid churn...
// We also defragment free blocks as we search, to optimise this and future searches.
block = heap.heap_start;
while (block < heap.heap_end)
{
// If the block is used, then keep looking.
if(!(*block & MICROBIT_HEAP_BLOCK_FREE))
{
block += *block;
continue;
}
blockSize = *block & ~MICROBIT_HEAP_BLOCK_FREE;
// We have a free block. Let's see if the subsequent ones are too. If so, we can merge...
next = block + blockSize;
while (*next & MICROBIT_HEAP_BLOCK_FREE)
{
if (next >= heap.heap_end)
break;
// We can merge!
blockSize += (*next & ~MICROBIT_HEAP_BLOCK_FREE);
*block = blockSize | MICROBIT_HEAP_BLOCK_FREE;
next = block + blockSize;
}
// We have a free block. Let's see if it's big enough.
// If so, we have a winner.
if (blockSize >= blocksNeeded)
break;
// Otherwise, keep looking...
block += blockSize;
}
// We're full!
if (block >= heap.heap_end)
{
__enable_irq();
return NULL;
}
// If we're at the end of memory or have very near match then mark the whole segment as in use.
if (blockSize <= blocksNeeded+1 || block+blocksNeeded+1 >= heap.heap_end)
{
// Just mark the whole block as used.
*block &= ~MICROBIT_HEAP_BLOCK_FREE;
}
else
{
// We need to split the block.
uint32_t *splitBlock = block + blocksNeeded;
*splitBlock = blockSize - blocksNeeded;
*splitBlock |= MICROBIT_HEAP_BLOCK_FREE;
*block = blocksNeeded;
}
// Enable Interrupts
__enable_irq();
return block+1;
}
/**
* Attempt to allocate a given amount of memory from any of our configured heap areas.
* @param size The amount of memory, in bytes, to allocate.
* @return A pointer to the allocated memory, or NULL if insufficient memory is available.
*/
void *microbit_malloc(size_t size)
{
void *p;
// Assign the memory from the first heap created that has space.
for (int i=0; i < MICROBIT_HEAP_COUNT; i++)
{
if(heap[i].heap_start != NULL)
{
p = microbit_malloc(size, heap[i]);
if (p != NULL)
{
microbit: Conditional compilation on MicroBitConfig All options defined in MicroBitConfig are now set IFF they already do not have a value. This allows external configuration of these options much more cleanly. Additional macros also defined to provide clean implementation for boolean configuration switches. Also some minor bugfixes: - Correction of typo of "Microbit.h" => "MicroBit.h" - Fixed heap corruption bug on gcc as a result of a debug printf in an unsafe context.
2015-09-02 11:42:24 +00:00
#if CONFIG_ENABLED(MICROBIT_DBG) && CONFIG_ENABLED(MICROBIT_HEAP_DBG)
minor bugfixes in MICROBIT_HEAP_DEBUG
2015-11-17 16:58:42 +00:00
uBit.serial.printf("microbit_malloc: ALLOCATED: %d [%p]\n", size, p);
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
#endif
return p;
}
}
}
// If we reach here, then either we have no memory available, or our heap spaces
// haven't been initialised. Either way, we try the native allocator.
p = native_malloc(size);
if (p!= NULL)
{
microbit: Conditional compilation on MicroBitConfig All options defined in MicroBitConfig are now set IFF they already do not have a value. This allows external configuration of these options much more cleanly. Additional macros also defined to provide clean implementation for boolean configuration switches. Also some minor bugfixes: - Correction of typo of "Microbit.h" => "MicroBit.h" - Fixed heap corruption bug on gcc as a result of a debug printf in an unsafe context.
2015-09-02 11:42:24 +00:00
#if CONFIG_ENABLED(MICROBIT_DBG) && CONFIG_ENABLED(MICROBIT_HEAP_DBG)
// Keep everything trasparent if we've not been initialised yet
if (microbit_active_heaps())
minor bugfixes in MICROBIT_HEAP_DEBUG
2015-11-17 16:58:42 +00:00
uBit.serial.printf("microbit_malloc: NATIVE ALLOCATED: %d [%p]\n", size, p);
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
#endif
return p;
}
// We're totally out of options (and memory!).
microbit: Conditional compilation on MicroBitConfig All options defined in MicroBitConfig are now set IFF they already do not have a value. This allows external configuration of these options much more cleanly. Additional macros also defined to provide clean implementation for boolean configuration switches. Also some minor bugfixes: - Correction of typo of "Microbit.h" => "MicroBit.h" - Fixed heap corruption bug on gcc as a result of a debug printf in an unsafe context.
2015-09-02 11:42:24 +00:00
#if CONFIG_ENABLED(MICROBIT_DBG) && CONFIG_ENABLED(MICROBIT_HEAP_DBG)
// Keep everything trasparent if we've not been initialised yet
if (microbit_active_heaps())
minor bugfixes in MICROBIT_HEAP_DEBUG
2015-11-17 16:58:42 +00:00
uBit.serial.printf("microbit_malloc: OUT OF MEMORY\n");
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
#endif
microbit: Conditional compilation on MicroBitConfig All options defined in MicroBitConfig are now set IFF they already do not have a value. This allows external configuration of these options much more cleanly. Additional macros also defined to provide clean implementation for boolean configuration switches. Also some minor bugfixes: - Correction of typo of "Microbit.h" => "MicroBit.h" - Fixed heap corruption bug on gcc as a result of a debug printf in an unsafe context.
2015-09-02 11:42:24 +00:00
#if CONFIG_ENABLED(MICROBIT_PANIC_HEAP_FULL)
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
panic(MICROBIT_OOM);
#endif
return NULL;
}
/**
* Release a given area of memory from the heap.
* @param mem The memory area to release.
*/
void microbit_free(void *mem)
{
uint32_t *memory = (uint32_t *)mem;
uint32_t *cb = memory-1;
microbit: Conditional compilation on MicroBitConfig All options defined in MicroBitConfig are now set IFF they already do not have a value. This allows external configuration of these options much more cleanly. Additional macros also defined to provide clean implementation for boolean configuration switches. Also some minor bugfixes: - Correction of typo of "Microbit.h" => "MicroBit.h" - Fixed heap corruption bug on gcc as a result of a debug printf in an unsafe context.
2015-09-02 11:42:24 +00:00
#if CONFIG_ENABLED(MICROBIT_DBG) && CONFIG_ENABLED(MICROBIT_HEAP_DBG)
if (microbit_active_heaps())
minor bugfixes in MICROBIT_HEAP_DEBUG
2015-11-17 16:58:42 +00:00
uBit.serial.printf("microbit_free: %p\n", mem);
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
#endif
// Sanity check.
if (memory == NULL)
return;
// If this memory was created from a heap registered with us, free it.
for (int i=0; i < MICROBIT_HEAP_COUNT; i++)
{
if(memory > heap[i].heap_start && memory < heap[i].heap_end)
{
// The memory block given is part of this heap, so we can simply
// flag that this memory area is now free, and we're done.
*cb |= MICROBIT_HEAP_BLOCK_FREE;
return;
}
}
// If we reach here, then the memory is not part of any registered heap.
// Forward it to the native heap allocator, and let nature take its course...
native_free(mem);
}
Reference in a new issue Copy permalink