This allows HCI drivers to expose vendor-specific functions to set the
public address.
Signed-off-by: Armin Brauns <armin.brauns@embedded-solutions.at>
Increase the SPI RX driver stack size by 128 bytes. Overflows have
previously been observed on real hardware at the default stack size of
512.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
The HCI receive path has a delay between reading the header and payload
from the controller to give the controller time to setup the SPI
peripheral for the next transaction. Add the same delay on the transmit
path for the same reasons.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
Add log output when the HCI interface is forced to retransmit a packet
because the controller is not ready.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
Modify SPI configuration to match the features introduced in PR #63437.
Set the property "controller-data-delay-us" to zero for boards
which are using BlueNRG-MS.
Fix Chip Select configuration for stm32l562e_dk board.
Signed-off-by: Ali Hozhabri <ali.hozhabri@st.com>
The existing driver and sample:
- drivers/bluetooth/hci/rpmsg
- samples/bluetooth/hci_rpmsg
are no longer correctly named, since they now use the IPC subsystem to
send and receive data. The IPC subsystem can use RPMsg as a transport,
but that is one of several selectable backends.
I initially wanted to deprecated both the BT_RPMSG Kconfig option as
well as the zephyr,bt-hci-rpmsg-ipc chosen node in Devicetree. However,
this proved to be undoable in the case of the Kconfig option. This is
because it's a choice option, and those have special behavior. In
particular, the only practical way to deprecate would've been to keep
the old Kconfig option outside the choice (much like it's done in this
commit) but then also add a 'depends on !BT_RPMSG' on each of the
remaining choice symbols *except* on the new BT_HCI_IPC one. This, however,
only works correctly for .conf files. If a board instead sets the
default BT_HCI_BUS_TYPE in the Kconfig.defconfig file then the Kconfig
tree parsing would fail, because it'd try to set it to a value
(BT_RPMSG) that is no longer part of the choice.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Don't re-use the `rxmsg` buffer in the `bt_spi_send` function. This
buffer is still used by the RX thread after releasing the SPI semaphore.
The current re-use can result in buffer corruption if the RX thread is
swapped out as a result of the `k_sem_give`.
Moving the semaphore release later can result in deadlocks due to
buffer allocation being performed while holding the semaphore, so
instead just eliminate the re-use entirely.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
Extract the HCI message handling out to a dedicated function to simplify
the main thread function. This also solves a bug as a side effect.
Previously `discardable` and `timeout` were never being reset after
an advertising report was received, resulting in ALL events after the
first advertising report being treated as discardable.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
Ensure that the Bluetooth rpmsg driver thread is temporarly in
cooperative mode using `k_sched_lock`.
The previous commit added the documentation on `bt_recv` stemming from
the general consensus among maintainers that `bt_recv` may not be called
from preemptible priorites.
Many uses may be affected by this race condition, since the default
configuration of rpmsg driver selects a preemtible priority.
Signed-off-by: Aleksander Wasaznik <aleksander.wasaznik@nordicsemi.no>
The rpmsg_close() call uses the HCI reset command to reset the
controller. But when building as controller-only we do not bring in the
infrastructure to send HCI commands (nor should we) and rpmsg_close()
will not be called anyway.
Fixes#63534.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Add .close() implementation for the HCI RPMsg driver.
When running on the nRF5340 application core, it will power-cycle the
network core.
Signed-off-by: Pawel Osypiuk <pawelosyp@gmail.com>
Signed-off-by: Jonathan Rico <jonathan.rico@nordicsemi.no>
Enable backends by default, instead of requiring some other module to
turn them on. This aligns with the behaviour of sensor drivers and
`BT_RPMSG`.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
Be explicit about the HCI backend that Bluetooth unit tests require.
Some unit tests depend on `BT_HAS_HCI_VS`, so also enable that.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
Configure the Number of Completed Packets event threshold to 1, so the
SiLabs Bluetooth controller will always send the Number of Completed
Packets HCI event to the host, even for small numbers of transmitted
packets.
Fixes#62279
Signed-off-by: Markus Fuchs <markus.fuchs@ch.sauter-bc.com>
Make HCI driver compatible to the updated versions of the
EFR32 Bluetooth library BLOBs consisting of:
* libbluetooth_controller
* libbgcommon
* librail
Signed-off-by: Markus Fuchs <markus.fuchs@ch.sauter-bc.com>
`CONFIG_BT_L2CAP_TX_MTU` drives the maximum supported MTU on a Bluetooth
connection, but the SPI backend also imposes its own hidden limits.
Display a warning if a value is chosen that can result in failures to
send. This is done here instead of as a range on `BT_L2CAP_TX_MTU` as
ranges on that symbol are already defined in terms of software features,
which would conflict with this check.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
Interrupt is enabled before reset is released to make sure that the
first IRQ edge is captured, and rx thread can process it.
Remove delay in spi_open as it was redundant due to the "sem_initialised"
semaphore.
Signed-off-by: Ali Hozhabri <ali.hozhabri@st.com>
After conducting tests with a a virtual Bluetooth controller
over TCP it was noticed that some HCI packets may arrive on the
same buffer if sent over a short period of time.
This update ensures the hci packets are parsed correctly in the case
multiple packets are in the same recieving buffer according to
the Bluetooth Spec v5.4 Part E.
Signed-off-by: Victor Chavez <chavez-bermudez@fh-aachen.de>
Added support to connect to an HCI TCP Server. This
allows to do integration tests with other frameworks
that support a virtual hci interface.
Signed-off-by: Victor Chavez <chavez-bermudez@fh-aachen.de>
Currently on zephyr, RAIL is used only for bluetooth. RAIL library is
needed to use efr32 radio regardless of the protocol used. We add
SOC_GECKO_USE_RAIL kconfig option to indicate if we use radio.
FPU is needed when using RAIL, we configure it if SOC_GECKO_USE_RAIL
is set.
Signed-off-by: Antoine Bout <antoine.bout@silabs.com>
Refactor the ESP32 target SOCs together with
all related boards. Most braking changes includes:
- changing the CONFIG_SOC_ESP32* to refer to
the actual soc line (esp32,esp32s2,esp32s3,esp32c3)
- replacing CONFIG_SOC with the CONFIG_SOC_SERIES
- creating CONFIG_SOC_FAMILY_ESP32 to embrace all
the ESP32 across all used architectures
- introducing CONFIG_SOC_PART_NUMBER_* to
provide a SOC model config
- introducing the 'common' folder to hide all
commonly used configs and files.
- updating west.yml to reflect previous changes in hal
Signed-off-by: Marek Matej <marek.matej@espressif.com>
Currently, HCI packet handling does not consider the BT_RECV_CONTEXT
choice selection. It calls bt_recv() and bt_recv_prio() only depending
on the HCI packet type and event flags.
However, for selections other than BT_RECV_BLOCKING, the "HCI driver
shall not call bt_recv_prio()". Fix that by only calling bt_recv_prio()
when CONFIG_BT_RECV_BLOCKING is enabled.
Signed-off-by: Markus Fuchs <markus.fuchs@ch.sauter-bc.com>
Add a small delay between reading the transport header and reading the
HCI data. Failing to do so on a nRF9160<->nRF52832 link was reliably
resulting in the nRF9160 trying to read data before the nRF52832 had
set up the SPI transaction, resulting in the host reading a buffer full
of 0x00 and having to run the entire read result again.
Transceiving a 10 byte packet takes at least 31uS, while 100 byte
packets are around 150uS (duration of `spi_transceive` call). Waiting
1 tick to eliminate the need for most retransmissions is a valid
tradeoff.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
GPIO callbacks need to be initialized before being registered with a GPIO
driver, otherwise the NULL `handler` hits an assert.
Signed-off-by: Armin Brauns <armin.brauns@embedded-solutions.at>
Don't wait for the IRQ line to be de-asserted in `bt_spi_send`. As this
function can be called by the RX processing thread, the previous
behaviour could cause deadlocks under heavy load:
* SPI RX thread starts blocking on `bt_buf_get_evt` due to load
* BT controller generates another event, asserting the IRQ line
* RX processing thread calls `bt_spi_send` in reponse to event
* RX processing thread blocks forever on the removed condition
There is no need to attempt to rate-limit how often `bt_spi_send` is
called to allow the RX thread to run. If the bus is so congested that
there is no remaining capacity, prioritising RX over TX is not going to
improve the situation.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
This commit enabled the Silicon Labs Bluetooth interface driver on the
EFR32BG27 SoC series in the Kconfig.
Signed-off-by: Filip Kokosinski <fkokosinski@antmicro.com>
This commits adapts the driver to the newer version of the EFR32 BT
blobs:
* libbluetooth_controller
* libbgcommon
* librail
Signed-off-by: Filip Kokosinski <fkokosinski@antmicro.com>
Using EM2 or deeper sleep states (where HF clocks are off) requires
special care if BLE radio is used, since BLE radio relies on that clock,
and its power/clock requirements need to be taken into account
On SiLabs, radio PM is implemented as part of RAIL blob, which relies
on sl_power_manager HAL service. I've implemented SoC PM
state changes using sl_power_manager instead of emlib, and added
call to RAIL PM initialization in Gecko HCI driver.
Signed-off-by: Roman Dobrodii <rdobrodii@antmicro.com>
This commit adds BLE support to the `efr32xg24_dk2601b` board. It also
modifies the SiLabs BT HCI driver to accomodate the EFR32xG24 SoC
series.
Signed-off-by: Filip Kokosinski <fkokosinski@antmicro.com>
Use level triggered interrupts instead of edge triggered interrupts to
make it impossible for the interrupt line to be active without the RX
thread attempting to read pending data.
This fixes a race condition for re-enabling the interrupt in edge
triggered mode, where `gpio_pin_interrupt_configure_dt` is not called
before the controller re-asserts the interrupt line.
As a result the controller can be reset before booting the RX thread.
Fixes#53980.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
As of today it is not possible to use SPI dt-spec macros in C++,
something known and documented. The main reason is because `cs` property
is initialized using a compound literal, something not supported in C++.
This PR takes another approach, that is to not make `cs` a pointer but a
struct member. This way, we can perform a regular initialization, at the
cost of using extra memory for unused delay/pin/flags if `cs` is not
used.
Fixes#56572
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
This commit removes the unused argument form the `slz_bt_init` function,
and makes it consistent with what `SYS_INIT` expects - `int (*)(void)`.
Signed-off-by: Filip Kokosinski <fkokosinski@antmicro.com>
This commit adds the SiLabs Bluetooth HCI driver. It also enables this
BLE HCI driver on the efr32bg_sltb010a board.
Signed-off-by: Filip Kokosinski <fkokosinski@antmicro.com>
The init infrastructure, found in `init.h`, is currently used by:
- `SYS_INIT`: to call functions before `main`
- `DEVICE_*`: to initialize devices
They are all sorted according to an initialization level + a priority.
`SYS_INIT` calls are really orthogonal to devices, however, the required
function signature requires a `const struct device *dev` as a first
argument. The only reason for that is because the same init machinery is
used by devices, so we have something like:
```c
struct init_entry {
int (*init)(const struct device *dev);
/* only set by DEVICE_*, otherwise NULL */
const struct device *dev;
}
```
As a result, we end up with such weird/ugly pattern:
```c
static int my_init(const struct device *dev)
{
/* always NULL! add ARG_UNUSED to avoid compiler warning */
ARG_UNUSED(dev);
...
}
```
This is really a result of poor internals isolation. This patch proposes
a to make init entries more flexible so that they can accept sytem
initialization calls like this:
```c
static int my_init(void)
{
...
}
```
This is achieved using a union:
```c
union init_function {
/* for SYS_INIT, used when init_entry.dev == NULL */
int (*sys)(void);
/* for DEVICE*, used when init_entry.dev != NULL */
int (*dev)(const struct device *dev);
};
struct init_entry {
/* stores init function (either for SYS_INIT or DEVICE*)
union init_function init_fn;
/* stores device pointer for DEVICE*, NULL for SYS_INIT. Allows
* to know which union entry to call.
*/
const struct device *dev;
}
```
This solution **does not increase ROM usage**, and allows to offer clean
public APIs for both SYS_INIT and DEVICE*. Note that however, init
machinery keeps a coupling with devices.
**NOTE**: This is a breaking change! All `SYS_INIT` functions will need
to be converted to the new signature. See the script offered in the
following commit.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
init: convert SYS_INIT functions to the new signature
Conversion scripted using scripts/utils/migrate_sys_init.py.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
manifest: update projects for SYS_INIT changes
Update modules with updated SYS_INIT calls:
- hal_ti
- lvgl
- sof
- TraceRecorderSource
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
tests: devicetree: devices: adjust test
Adjust test according to the recently introduced SYS_INIT
infrastructure.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
tests: kernel: threads: adjust SYS_INIT call
Adjust to the new signature: int (*init_fn)(void);
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Unify the drivers/*/Kconfig menuconfig title strings to the format
"<class> [(acronym)] [bus] drivers".
Including both the full name of the driver class and an acronym makes
menuconfig more user friendly as some of the acronyms are less well-known
than others. It also improves Kconfig search, both via menuconfig and via
the generated Kconfig documentation.
Signed-off-by: Henrik Brix Andersen <hebad@vestas.com>
