This is technically a breaking change, but:
a) We need the end handle to do descriptor discovery properly.
b) We have no possible use for the existing definition handle in the
characteristic result IRQ. None of the methods can use it, and therefore
no existing code should be using it in a way that changing it to a
different integer value should break.
Unfortunately NimBLE doesn't make it easy to get the end handle, so also
implement a mechanism to use the following characteristic to calculate
the previous characteristic's end handle.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This uses MP_REGISTER_ROOT_POINTER() to register bluetooth_nimble_memory
and bluetooth_nimble_root_pointers and removes the same from all
mpconfigport.h.
Signed-off-by: David Lechner <david@pybricks.com>
We're using the MicroPython fork of NimBLE, which on the
`micropython_1_4_0` branch re-adds support for 64-bit targets and fixes
initialisation of g_msys_pool_list.
Also updates modbluetooth_nimble.c to suit v1.4.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This allows the write to trigger a notification or indication, but only to
subscribed clients. This is different to gatts_notify/gatts_indicate,
which will unconditionally notify/indicate.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
When l2cap_send detects that the sys mempool is running low (used to store
the outgoing HCI payloads), it will report stalled back to the application,
and then only unstall once these HCI payloads have been sent.
This prevents a situation where a remote receiver with very large MTU can
cause NimBLE to queue up more than MYNEWT_VAL_MSYS_1_BLOCK_COUNT (i.e. 12)
payloads, causing further attempts to send to fail with ENOMEM (even though
the channel is not stalled and we have room in the channel mbufs). The
regular credit/stall flow control is not effective here because the
receiver's MTU is large enough that it will not activate (i.e. there are
lots of credits available).
Thresholds of 1/2 (stall) and 1/4 (unstall) chosen to allow headroom for
other payloads (e.g. notifications) and that when a regular stall occurs it
might keep sending (and creating more payloads) in the background.
A correctly-behaved application shouldn't do this, but in the
case where the channel is stalled, there's still enough room
in the mbuf pool, then we'll fail the send (BLE_HS_EBUSY) so
the mbuf needs to be freed.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
The comments in NimBLE for ble_gattc_notify_custom() state that "This
function consumes the supplied mbuf regardless of the outcome.". And
inspection of NimBLE code shows that this is the case. So the comment can
be removed.
Signed-off-by: Damien George <damien@micropython.org>
Previously, the MICROPY_PY_BLUETOOTH_ENABLE_CENTRAL_MODE macro
controlled enabling both the central mode and the GATT client
functionality (because usually the two go together).
This commits adds a new MICROPY_PY_BLUETOOTH_ENABLE_GATT_CLIENT
macro that separately enables the GATT client functionality.
This defaults to MICROPY_PY_BLUETOOTH_ENABLE_CENTRAL_MODE.
This also fixes a bug in the NimBLE bindings where a notification
or indication would not be received by a peripheral (acting as client)
as gap_event_cb wasn't handling it. Now both central_gap_event_cb
and peripheral_gap_event_cb share the same common handler for these
events.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
Zephyr controllers can be queried for a static address (computed from the
device ID). BlueKitchen already supports this, but make them both use the
same macro to enable the feature.
On error, the handle is only available on err->att_handle rather than
in attr->handle used in the non-error case.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
If the _IRQ_L2CAP_RECV handler does the actual consumption of the incoming
data (i.e. via l2cap_recvinto), rather than setting a flag for
non-scheduler-context to handle it later, then two things can happen:
- It can starve the VM (i.e. the scheduled task never terminates). This is
because calling l2cap_recvinto will empty the rx buffer, which will grant
more credits to the channel (an HCI command), meaning more data can
arrive. This means that the loop in hal_uart.c that keeps reading HCI
data from the uart and executing NimBLE events as they are created will
not terminate, preventing other VM code from running.
- There's no flow control (i.e. data will arrive too quickly). The channel
shouldn't be given credits until after we return from scheduler context.
It's preferable that no work is done in scheduler/IRQ context. But to
prevent this being a problem this commit changes l2cap_recvinto so that if
it is called in IRQ context, and the Python handler empties the rx buffer,
then don't grant credits until the Python handler is complete.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This allows sending arbitrary HCI commands and getting the response. The
return value of the function is the status of the command.
This is intended for debugging and not to be a part of the public API, and
must be enabled via mpconfigboard.h. It's currently only implemented for
NimBLE bindings.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This adds `_IRQ_GET_SECRET` and `_IRQ_SET_SECRET` events to allow the BT
stack to request the Python code retrive/store/delete secret key data. The
actual keys and values are opaque to Python and stack-specific.
Only NimBLE is implemented (pending moving btstack to sync events). The
secret store is designed to be compatible with BlueKitchen's TLV store API.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This allows the application to be notified if any of encrypted,
authenticated and bonded state change, as well as the encryption key size.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
Instead of returning None/bool from the IRQ, return None/int (where a zero
value means success). This mirrors how the L2CAP_ACCEPT return value
works.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This widens the characteristic/descriptor flags to 16-bit, to allow setting
encryption/authentication requirements.
Sets the required flags for NimBLE and btstack implementations.
The BLE.FLAG_* constants will eventually be deprecated in favour of copy
and paste Python constants (like the IRQs).
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This allows the application to be notified of changes to the connection
interval, connection latency and supervision timeout.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
Also known as L2CAP "connection oriented channels". This provides a
socket-like data transfer mechanism for BLE.
Currently only implemented for NimBLE on STM32 / Unix.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This changes stm32 from using PENDSV to run NimBLE to use the MicroPython
scheduler instead. This allows Python BLE callbacks to be invoked directly
(and therefore synchronously) rather than via the ringbuffer.
The NimBLE UART HCI and event processing now happens in a scheduled task
every 128ms. When RX IRQ idle events arrive, it will also schedule this
task to improve latency.
There is a similar change for the unix port where the background thread now
queues the scheduled task.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This requires that the event handlers are called from non-interrupt context
(i.e. the MicroPython scheduler).
This will allow the BLE stack (e.g. NimBLE) to run from the scheduler
rather than an IRQ like PENDSV, and therefore be able to invoke Python
callbacks directly/synchronously. This allows writing Python BLE handlers
for events that require immediate response such as _IRQ_READ_REQUEST (which
was previous a hard IRQ) and future events relating to pairing/bonding.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
Using a semaphore (the previous approach) will only run the UART, whereas
for startup we need to also run the event queue.
This change makes it run the full scheduler hook.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
Instead of having the stack indicate a "start", "data"..., "end", pass
through the data in one callback as an array of chunks of data.
This is because the upcoming non-ringbuffer modbluetooth implementation
cannot buffer the data in the ringbuffer and requires instead a single
callback with all the data, to pass to the Python callback.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
This allows `ble.active(1)` to fail correctly if the HCI controller is
unavailable.
It also avoids an infine loop in the NimBLE event handler where NimBLE
doesn't correctly detect that the HCI controller is unavailable and keeps
trying to reset.
Furthermore, it fixes an issue where GATT service registrations were left
allocated, which led to a bad realloc if the stack was activated multiple
times.
Generally a controller should either have its own public address hardcoded,
or loaded by the driver (e.g. cywbt43).
However, for a controller that has no public address where you still want a
long-term stable address, this allows you to use a static address generated
by the port. Typically on STM32 this will be an LAA, but a board might
override this.
Previously the interaction between the different layers of the Bluetooth
stack was different on each port and each stack. This commit defines
common interfaces between them and implements them for cyw43, btstack,
nimble, stm32, unix.
This adds an additional optional parameter to gap_scan() to select active
scanning, where scan responses are returned as well as normal scan results.
This parameter is False by default which retains the existing behaviour.
This commit adds the IRQ_GATTS_INDICATE_DONE BLE event which will be raised
with the status of gatts_indicate (unlike notify, indications require
acknowledgement).
An example of its use is added to ble_temperature.py, and to the multitests
in ble_characteristic.py.
Implemented for btstack and nimble bindings, tested in both directions
between unix/btstack and pybd/nimble.
The goal of this commit is to allow using ble.gatts_notify() at any time,
even if the stack is not ready to send the notification right now. It also
addresses the same issue for ble.gatts_indicate() and ble.gattc_write()
(without response). In addition this commit fixes the case where the
buffer passed to write-with-response wasn't copied, meaning it could be
modified by the caller, affecting the in-progress write.
The changes are:
- gatts_notify/indicate will now run in the background if the ACL buffer is
currently full, meaning that notify/indicate can be called at any time.
- gattc_write(mode=0) (no response) will now allow for one outstanding
write.
- gattc_write(mode=1) (with response) will now copy the buffer so that it
can't be modified by the caller while the write is in progress.
All four paths also now track the buffer while the operation is in
progress, which prevents the GC free'ing the buffer while it's still
needed.
This commit makes sure that all discovery complete and read/write status
events set the status to zero on success.
The status value will be implementation-dependent on non-success cases.
Jim Mussared