Datagram receive code incorrectly calculated the payload length, resulting in
packets being delivered to applications that were two bytes shorter than they
should be.
This commit introduces persistent storage for the compass between
device resets.
setCalibration now also stores the data in FLASH if it differs from
what it contains.
In source/MicroBitSuperMain.cpp, the data is restored if there is
persistent data available that is non-zero.
(i.e. x, y and z are NOT zero).
This ensures that users can code without interference from persistent
storage.
There is an issue (https://github.com/bbc/microbit-extras/issues/2755)
whereby CCCD data is required to be persisted through device resets.
This commit integrates the new MicroBitStorage class, and persistently
stores this information in flash.
P.S. what a long night!
Integrated simple 1K block buffer to host small amount of persistent
data and a 1K FLASH memory block located immediately below the FOTA/DFU
bootloader and APPDATA region that intended to hold core configuration
information, such as:
* BLE bonding information.
* Compass Calibration data.
* Thermometer Calibration data.
The block is resident at 0x0003B400 - 0x0003B800.
The impact of the pixels at lower values was not distinct enough
this commit adds a blocking wait for the first two timings for
greyscale.
Additionally, this commit modifies the timings const array to
take into account the time taken for renderGreyscale to execute.
Previously the DAL relied upon an implementation of a timer that
was very imprecise and lacked resolution, this commit adds a reference
to a frozen repo with us_ticker.c modifications, so that a higher resolution
timer is utilised.
This fixes#13 (finally), and also fixes lower brightness levels when using
DISPLAY_MODE_BLACK_AND_WHITE to not cause a device lockup.
Missed a zero in a previous commit, that caused a system lockup due to
a missed timer event.
This commit also changes the minimum display brightness to 4, to also
minimise another feature that needs fixing in the future, where the
timer does not interrupt in the expected time frame.
Previously the range of MicrobitLightSensor read() was 0 - 100,
this didn't align with other APIs.
The new range is 0-255 which aligns with display brightness to give
a more consistent overall API.
Previous revisions of microbit-dal would handle the case of a bonding request
when the bond table was full by attempting to purge the bonding table on
demand. This suffered two flaws:
1) Nordic's device manager does not support the purging of the bonding table
while a connection is active (including a pairing request).
2) Some devices (e.g. Android 4.4.2 take two entries in the bonding table
when paired), thus making predictions of how 'full' the table is challenging.
This patch employs a simpler replacement policy. The maximum size of the bond
table has been increased to 6 (from 4). However, if the bond table contains 4
or more entries when entering pairing mode it will be purged in advance of any
pairing requetss being initiated.
* This commit fixes an exposed member variable
that was not intended to be public 'tickPeriod'
* normalised the light level to return values between 0 - 100
The BLE stack locks out the CPU for long periods of time when necessary. This
can be alleviated by eans of a flag that release the CPU between radio events.
(BLE_COMMON_OPT_RADIO_CPU_MUTEX).
This can cause issues from applications with tighter timing requirements.
This patch only enables this MUTEX when a BLE connection is active, rather
than automatically at startup. Therefore, application that do not actively use
BLE will not suffer this penalty.
Integration of new radio mode that permits simple local area multicast communication
directly between micro:bits using the device's on board radio hardware.
It should be noted that this implementation build directly on the nrf51822
hardware. It is not a BLE protocol, and must (currently) be operated with BLE
disabled.
A standard on the wire configuration is defined, to promote future
interoperability between projects and languages. A lightweight
packet format is also defined and adhered to, so as to promote future
extensibility without sacrificing compatibility. The notion of groups is also
introduced, that enable multiple groups of children to work in the same area
without compromising each other's projects.
In adition to a low level packet radio driver, two higher level protocols are
defined that allow simple user datagram communication between micro:bits and
an EventBus protocol that allows the micro:bit's MessageBus to be extended to
other micro:bits.
Previously, to light sense a user would have to configure a
MicroBitLightSensor instance themselves, and flip the display mode
manually. This is difficult in languages that target our API.
This commit resolves that by adding a new method readLightLevel which
flips the display mode, and instantiates a light sensor.
When changing the mode, the tickSpeed is also modified to reduce
artefacts on the display.
Added a new class called MicroBitLightSensor which interleaves sensing
with the display.
This class takes advantage of the previous commits which add a dynamic
system tick, and a new display mode for the display.
A future optimisation would be for the display to handle the instantion
of a light sensor when the display mode is changed!
This commit adds a mutator that dynamically reconfigures the
systemTicker to call systemTick at a different period to the default.
The accessor returns the current tick speed in milliseconds.
The accessor is now also used in scheduler_tick to keep timing as
accurate as possible.
MICROBIT_DISPLAY_REFRESH_PERIOD has now been removed, and replaced with
MICROBIT_DEFAULT_TICK_PERIOD.
This commit introduces a new display mode,
DISPLAY_MODE_BLACK_AND_WHITE_LIGHT_SENSE, where the 4th render in
every “frame” is dropped allowing for other components that use the
display to use this dropped frame for processing. This is in
preparation for Light Sensing on the 5x5 LED matrix.
An accessor has also been added to MicroBitDisplay to check the current
display mode!
BLE maintains metadata on its characterisitics, some of which record whether
or not a central device (phone/tablet) wishes to receive notifications when a
given BLE characteristic changes value. One such characterisitic is used as
part of a core BLE service to indicate of the BLE profile of a device has
changed.
IOS is heavily reliant on this feature, and expects the state of this metadata
to be persistent across device reboots. This patch ensures that this metadata
is always set to allow Service Changed indications to be propagated to central
devices.
calibration
MicroBitCompass initiates a calibration procedure on demand if a heading is
requested from an uncalibrated device. The BLE MagnetometerService also
attempts to retrieve a heading periodically, but it is not good behaviour to
initiate calibration as a response to this.
This patch introduces a check such that the BLE MagnetometerService will only
attempt to read heading data from a previously calibrated compass.
BLE central devices (such as mobile phones/tablets) are permitted to cache the
list of services and characteristics provided by a BLE peripheral such as the
micro:bit. This was causing compatibility problems with IOS during FOTA
flashing, as the the list of BLE services may change after flashing.
The Nordic BLE stack contains a fixed size bond table - a list of devices that
have paired with our device (in the case the micro:bit). The default mbed
wrapper around the nordic stack does not account for the case where the bond
table is full.
This patch implements basic bounds checking within the MicroBitBLEManager so
that this table cannot overrun. If a new device attempts to pair when the bond
table is full, the table is emptied prior to completing bonding.
Added configuration option in MicroBitConfig to allow the default power level
to be defined. Based on experiments undertaken by the BBC and subsequent policies defined by
the BBC's child protection panel, to is set by default it's lowest level (0),
in order to protect children's privacy.
It should be noted that this is 30dB below the device's default setting, so will
inevitably leave the device more susceptible to interference, and therefore
significantly reduce the reliability of the device. This will likely present
itself through increased packet loss, reduced thorughout, lost data and
connection drop outs. The normal BLE power level on commercial devices
translates to power level of 6.
The author has witnessed an inability to connect to a micro:bit from a
distance of 40cm under domestic conditions using a Samsung S5 mini
running Android 4.4.2.
Added MicroBitConfig.h option to enable/disable whitelisting
Added MicroBitConfig.h option to enable/disable limited period advertising
Added support for limited period advertising
Minor code cleanups
See bbc/microbit-extras/issues/#1100 or lancaster-university/microbit-dal/issues/5 for full details.
Summary:
A stride of 0 would lock the calling fiber indefinitely, which was not the expected behaviour
for many people.
This update fixes this issue in both animate and scroll, returning immediately when there is
a stride of zero.
issue #73 highlighted an issue whereby the destruction of an instance
registered as an idle or system component, would result in a hardfault.
This was due to not deregistering idle or system callbacks.
This patch has been applied to all components currently in use by the
idle or system callbacks.
Replaced newer struct-style initialisation of CompassSample in MicroBit.cpp
with the more traditional constructor-style initialisation, to enfore
C98 compliance.
Closes#74
The destructor for MicroBitButton didn't take into account the fact that
an instance can be destructed after registering a callback. This would then
result in a HardFault if an instance was destructed.
A clear repro case was registering a Touch input on a pin (abstracted as a button)
and swapping to a DigitalIn.
* Added a number of #defines to MicroBitPin.h which ultimately lead to retabulation of the #defines
* Added two new instance methods:
- setServoValue() which sets the default servo period (20ms) and accepts a value in the range
0 - 180, which maps to a duty cycle between 5% and 10% respectively. This
mapping is also configurable.
- setServoPulseUs() which accepts a pulse width in microseconds, and maps it onto the PWM
channel.
Based on the previous commit, the api for MicroBitPin
has two new accessor methods:
* getAnalogPeriodUs
* getAnalogPeriod
There are also whitespace changes in the MicroBitPin head file
as well as some minor textual changes.
This commit introduces an initial draft fix
for a bug where it would take two "setPeriods" for the
PWM class to update to the expected value.
This commit also adds three new methods to a DynamicPwm instance:
* write() - which is a lightweight wrapper around PwmOut write,
so that the value can be trapped. The trapping of the value
is necessary for the bug fix.
* getPeriod() - which returns the value of a static variable,
containing the current period shared across the pwm bus.
* getValue() - which returns a scaled value in the range 0-1024
representing the current duty cycle.
Dynamic pwm previously didn't round robin transient pwm channels,
and would therefore bias towards the first transient channel it found.
This update ensures that there is now an equal opportunity for all
channels to be considered for redirection.
Previously there would be the possibility of the period being reset
each time a new DynamicPwm was instantiated. This has now been
removed from the constructor, and the comments have been updated
All internal uses of accelerometer sample values now updated to access the
data via accessor methods. This now ensures consistent use irrespective of
the coordinate system used.
native printf() call.
The Matrix4::print() function is useful, but non essential for most
applications, and the underlying printf() call required cost an
additional 1K of RAM on first use, which is currently above the budget for
micro:bit.
An e-compass solution requires knowwlede two pieces of data to provide an
accurate heading:
- Accurate calibration of the magnetometer hardware so that reliable
measurements can be taken.
- Knowledge of the pitch and roll of of device, so that the correct
components of the X/Y and Z axis sensors of the magnetomer can be used
to sense the magnetic field in a horizontal plane regardless of the tilt
of the device.
This commit represent changes to the MicroBitAccelerometer and MicroBitCompass
classes to implemen tthese goals. More specifically, this commit provides:
- The introduciton of an interactive calibration 'game', that can rapidly
gather all the data required to calibrate the compass.
- An improved calibration algorithm based on a Least Mean Squares approach of
compass samples, as documened in Freescale Application Note AN4248.
- The inclusion of a simple Matrix4 class to enable efficient Least Mean
Squares implementation.
- A change from asynchronous to synchronous calibration of the compass when
first used. This is in repsonse to a feature request for this from users
and high level languages using microbit-dal.
- Support for detemrining tilt and roll angle in MicroBitAccelerometer
- Support for multiple co-ordinate spaces in MicroBitAccelerometer and
MicroBitCompass. Data can now be read in either RAW (unaltered) data.
MICORBIT_SIMPLE_CARTESIAN (as used previously) or NORTH_EAST_DOWN
(the industry convention in mobile phones, tablets and aviation)
- Implementation of a tilt compensated algorithm, used when determining
device heading.
Freefall now detected when the net force on the device drops below 0.4g,
rather than 0.2g as previously defined. This improves sensitivity, without
generating false positives under common use.
Also correction of minor typos / layout errors.
- updated acceleration threshold event names to reflect numeric values e.g.
GESTURE_3G, GESTURE_6G, GESTURE_8G.
- Add scoping to elements of SimpleGesture enum.
- Code scrub of code layout caused by editor with incorrect configuration.
- rename of internal state variables and functions to better reflect use.
- added clarification to some comments.
Whilst Android devices permit explicit initiation of the BLE pairing process,
IOS based devices are only capable of implicit initiation of pairing based
on accessing a protected characteristic. Adding read capability to the
ControlPoint characteristic provides a safe and convenient way for IOS devices
to initiate pairing.
Minor amends:
- bleDisconnectionCallback signature change
- bleSetAdvertisingInterval now takes milliseconds as a parameter
- event based invocation of DFU bootloader
More specifically:
- BLUEZONE mode renamed to PairingMode (by popular request).
- Added timeout of 90 seconds before automatic device reboot.
- Move from a scroll to a print based algorithm to improve the amount of time key is displayed fully.
- Added compulsory 'click to bond' button press.
- Improved response time of tick/cross being displayed.
Changes to default BLE name. During normal operation, all micro:bits now have the same name: BBC micro:bit
When pairing mode (AKA BLUEZONE) is entered, device name details re appended as before, to give BBC micro:bit [<name>]
This is to provide a degree of anonymity to users, to make tracking of kids more difficult.
This option is now configured when the BLE stack is initialised. This ensures that SoftDevice is never
starved of CPU during periods of criticality. This does lock out the CPU for application use, but prevents
MIC failures caused by __disable_irq() operations (as found in the mbed Ticker API for example).
Minor amends:
- bleDisconnectionCallback signature change
- bleSetAdvertisingInterval now takes milliseconds as a parameter
- event based invocation of DFU bootloader
The following postures of the device are now detected:
- TILT_UP
- TILT_DOWN
- TILT_LEFT
- TILT_RIGHT
- FACE_UP
- FACE_DOWN
In addition, the following gestures are inferred:
- NONE
- SHAKE
- FREEFALL
- WHEEE (>=3g)
- SICK (>=5g)
- UNCONSCIOUS (>=8g)
Events are now triggered on the MessageBus upon the transition from one posture/gesture to another,
and a synchronous getGesture() method is now also provided to interogate the last gesture recognised.
I should be noted that the default accelerator range of +/-2g will be insufficient to detect some of
the events noted above, and MicroBitAccelerometer::setRange() should be used to increase the range
if required.
- LFSR generated values in a subset of the range requested (rounded to the nearest lower power of 2).
Corrected by increasing the number of bits usedby one, such that it now generated st least the power of 2 greater.
- replaced enable/disable of interrupts with a local snapshot of randomValue. Much of the nordic software is sensitive to interrupt timings,
so best avoided where possible. Chance of race condition is low, and effect is minimal (duplicate number returned).
- Removed flashcode base authorization
- silenced unused parameter in MicroBitHeapAllocator
- bugfixed BLUEZONE mode so that sensor data is still available via BLE
The serial baud rate is set duing construction of the uBit object. We
don't need to do it again, and especially with fixed values that can't
be overridden.
Uisng libc's implementation is likely to be safer than rolling our own. At least
the failure modes are well documented. (glibc's implementation of rand() is
actually not bad).
The LFSR used only provides 1 bit of random data each time it is cycled.
This implementation generates the minimum number of bits needed. Further
it discards numbers that are bigger than required and re-calculates -
this keeps the distribution flat.
If BLE is connected when performing a hard reset, attempt to disconnect
first to allow the remote peer to handle the disconnect gracefully
without haing to wait for a supervision timeout.
- refactored BLE funcitonality into BLEManager class.
- added security requirements standard BLE profile services.
- updated bluezone pairing process to use BLE passkey exchange.
Also recalibrated loop timers in MicroBitDisplay::Error(), as used by panic()...
strange these now seem off by several orders of magnitude.
TODO: Ensure CPU is running at correct internal frequency with an oscilloscope test...
Whilst a little more invasive change than the previous warnings
supression this is more correct and more obvious as to what's being done,
and what warnigns will remain in place for the rest of the comilation
unit.
Unfortunately we can't turn it off just while nrf_soc.h is being included
as the way the defines are used the compiler can't tell the parameters
are unused until the end of the compilation unit. So we can't use the
more normal
as that pops the supression too early.
WARNING: including nrf_soc.h will turn off unused-function warnings for
the rest of the compilation unit - see included nrf_svc.h
- Enures an event listener is not deleted whilst a fiber is activiely processing a queue
- Added support for resurrection of event listeners in cases where identical listeners are removed/added repetitively.
- Add maximum depth for event queues, to prevent buggy scripts causing total memory exhaustion.
- Suppress generation of A/B click events when A+B click is generated
- preservation of event ordering on messagebus for resursive event generation cases.
- bugfix of message bus processing to prevent occasional dual processing of events
- bugfix MicroBitDisplay to behave correctly when delay parameter is zero.
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.
Minor changes to BLE profile services (cleanup prior to merge with master branch):
- Added temperature period characteristic to match design pattern of other sensors.
- Updated endinaness of MicroBitLEDService Matrix to be more intuitive.
- Cleaned up Microbit.cpp by removing unused reference to BLE services.
- Exposed serial number and naming information thr BLE DeviceInformation and API.
- Updated BLE advertisement period to be alightly more repsonsive (200ms vs 1000ms).
- BUGFIX: micro:bit name is now always generated, regardless of BLE services.
- BUGFIX: Scroll Period data length.
Updates to change the behaviour of the scroll/print/animate faily of function away
from being pre-emtive and instead prroviding queing behaviour.
Minor updates to provide complete sets of async equivalent operations
Updates to the scheduler to provide wait/notify/waitone semantics.
All message bus listeners with the micro:bit runtie are non blocking, and safe to be called
in an interrupt context. As such, these have now all been declared MESSAGE_BUS_LISTENER_IMMEDIATE
which should mitigate all unbounded event queueing effects.
Updates to BLE interface to match specification following testing by Bluetooth SIG.
Accelerometer Period: now supports READ and WRITE
Magnetometer Period: now supports READ and WRITE
LED Matrix State: now supports READ and WRITE.
UUIDs of Temperature Service characteristic corrected.
More specifically, the build system now uses the semantic versioning meta-data
held in module.json to define a major.minor.patch version. Additionally, is the branch
being compiled is *not* the master branch, the version is appended with <branchname><githash>.
Specific updates:
- Updates to CMake files to expose this to the micro:bit runtime code.
- Addition of uBit.systemVersion() to expose this to application code.
- Displaying of version string over serial if MICROBIT_DBG is enabled.
- Distribution of version string over BLE via the firmware revision characteristic.
First functionally complete BLE profile, matching BLE speicfication v1.6.
More specifically, the following services are now functional:
- AccelerometerService
- MeganetometerService
- EventService
- TemperatureServide
- IOPinService
- DFUService
- ButtonService
- LEDService
Also, updates to underlying device drivers to enable greater configurability:
- MicroBitCompass now supports variable sample rates and temperature sensing
- MicroBitAccelerometer now supports variable sample rates and ranges
- MicroBitThermometer introduced
- MicroBitMessageBus adapted to permit enumeration and block removal of listeners
Finally, MicroBit DFU Service has been changed to the new UUIDs specificed in v1.6 of BLE spec.
Previous to this patch an explicit constructor was provided for creating
MatrixPoint instances, which meant that such objects could not be put in
ROM (since they needed to be constructed at runtime).
By using the defualt construction method for structs the MatrixPoint map
is now compiled into the rodata section, hence freeing up RAM.
Saves: 4 bytes data, 56 bytes bss, 148 bytes code.
- renamed fiber flag to be more general.
- renamed idle to idleFiber to be consistent with the rest of the code.
- factored content of idle() into a funciton to avoid dupication.
- added test case to MessageBus::process() to avoid forking fibers if not requested.
- removed fiber_allow_run_idle_within mutator in favour of exposing currentFiber, as a more general solution.
If there is only 1 main fiber (plus the idle fiber) then enabling this
new feature for the main fiber (using fiber_allow_run_idle_within())
will mean that no heap is needed to swap the stack out.
Added new URGENT listener type, that ensures immediate, non queuing delivery of events.
Introduced two-pass algorithm into message bus. Only events with an event listener are now
queued at ingress - this reduces unecessary memory usage for 'uninteresting' events.
Hotfix of a minor bug in the MicroBitDFU Service that prevented BLE notifications
from being sent on the FlashCode characterisitic. Needed as the Samsung companion
app is dependent on this functionality to enable BLE programmming of the micro:bit.
Events in the micro:bit runtime are normally launched through making an instance
of the MicroBitEvent class. This update added configuration support to all the
code creating the MicroBitEvent decide if the newly created event should then be
automatically queued on the MessageBus or processed immediately by event handlers.
MessageBus handlers can now have one of four concurrency modes for the eventuality
of an event being raised whilst a previous event is still being processed. An additional
(optional) parameter is provided to the listen() functions to allow this to be selected
on a per event handler basis. The permissable options are:
MESSAGE_BUS_LISTENER_REENTRANT:
The event handler is fired with the new event, regardless of whether or not
a previous event is still be processed by that handler.
MESSAGE_BUS_LISTENER_QUEUE_IF_BUSY:
The new event is queued until such a time as the previous event has completed
execution. The new event is then processed. This option does not preclude the
processing of the new event by other event handlers.
MESSAGE_BUS_LISTENER_DROP_IF_BUSY:
The new event is dropped, and will never be processed the the event handler.
This option does not preclude the processing of the new event by other event handlers.
MESSAGE_BUS_LISTENER_NONBLOCKING:
The event handler is self-declaring that it never blocks. This flag is used purely
for optimisation, as it permits direct execution of the event hadnelr without inducing
any overhead from the scheduler.
In addition, the following minor revisions were made in this release:
* Cleanup of the #include dependencies contained in the microbit-dal .h files
* Bugfix to the scheduler block on event matching code.
* Introduced a MICROBIT_ID_ALERT MessageBus channel, for general purpose eventing using nonces.
MicroBitAccelerometer now contans methods to allow dynamic selection of sample range
(from +/-2g to +/-8g) and sample frequency (up to 800Hz). The default settings and behaviour
remain consistent.
A header file containing custom configuration options for the microbit:runtime can
now be defined through a yotaa configuration option.
For example, including the following inside a top level config.json file
will load configuration options from a file called "MicroBitCustomConfig.h" that is stored
alongside an application's main.cpp. Any options defined there will override the defaults
defined in microbit-dal/inc/MicroBitConfig.h :
{
"microbit" :{
"configfile" : "source/MicroBitCustomConfig.h"
}
}
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.
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.
Updates to MicroBitDisplay::setBrightness() to perform a NOP for out of bound parameters.
This aligns the functionality of the runtime with the TD simulator.
Updates the MicroBitMessageBus implementation to use fork_on_block() calls rather
than create_fiber() calls when servicing events. This provides greater optimisation
for non-blocking event handlers.
More specifically, this update:
- Replaces calls to create_fiber() ith fork_on_block during event send operations.
- Adds a queue of events, to ensure events generated from interrupt context are not serviced in interrupt context.
- Registers a listener with the idle process to empty the event queue.
The microbit fiber scheduler is often used to service event handlers by the microbit message bus.
This provides a very elegant decoupling of user code from system code, interrupt handlers and
also allows users to make blocking calls within event handlers. However, many event handlers are
non-blocking, so launching a dedicated fiber is wasteful in terms of time and SRAM.
This patch adds fork_on_block() to the scheduler. Inspired by the UNIX copy_on_write
technique, this optimisation is semantically equivalent to a create_fiber() call, but will first attempt
to execute the given function in the context of the currently running fiber, and *only* create a fiber
if the given code attempts a blocking operation.
More specifically, this update:
- adds fork_on_block() functions for parameterised and non-parameterised functions.
- adds fields to the fiber context to record the status of parent/child fibers.
- adds optimised ASM functions to store and restore Cortex M0 register context.
- adds a utility function to determine if the processor is executing in interrupt context.
- updates to sleep() and wait_for_event() to handle fork_on_block semantics.
- minor code optimsations within the scheduler.
NO_BLE can now be defined to determine whether ble
will be brought up or not at initialisation of uBit.
This is especially useful for two new Yotta targets:
bbc-microbit-classic-gcc-nosd,
bbc-microbit-classic-armcc-nosd, which do not use
softdevice.
This is the first commit of the microbit-dal on GitHub.
This repository contains the runtime, which is a light
weight operating system developed by Lancaster University.