Review rework of the fix to remove the jitter in aux offset
and sync offset values.
Force select BT_TICKER_REMAINDER_GET and BT_TICKER_LAZY_GET
features when Extended Advertising and Periodic Advertising
is supported.
Rename ticks and microsecond offset value struct members
for primary PDU event offset (to auxiliary PDU event).
Converted HAL_TICKER_REMOVE_JITTER and HAL_TICKER_ADD_JITTER
macro to functions.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Use remainder value in scheduling the periodic auxiliary
PDUs and use the ticker next slot get interface with
remainder value to fill the auxiliary offsets with
microsecond precision in the primary channel PDUs.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Updated ticker implementation to return remainder value for
a ticker when enumerating active tickers with time
reservations.
This is required to find offsets and to use the remainder
value to correctly calculate auxiliary offsets to the
microsecond resolution.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Fix overlapping 1M and Coded PHY scanning that caused idle
radio time when both PHY use same scan interval and sum of
their scan window duration equals the interval.
Implementation now will use continuous scanning and offset
the start of Coded PHY by the window duration of the 1M
scanning.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
The initialisation of DLE parameters for the peripheral
was done before the intialisation of the PHY settings.
Since the DLE parameters depend on PHY settings this
can result in incorrect parameters for tx/rx time and
octets
One scenario is where a previous connection set the PHY to
2M or CODED, then when a new connection is established
it uses the same memory-locations for connection settings as the
previous connection, and the (uninitialised) PHY settings will be
set to 2M or CODED, and thus the DLE parameters will be wrong
This PR moves the initialisation of DLE parameters after
that of PHY settings
EBQ tests effected include LL/CON/PER/BV-77-C.
Signed-off-by: Andries Kruithof <andries.kruithof@nordicsemi.no>
Add explicit selection of the BT_HAS_HCI_VS for
the BT_RPMSG HCI driver. When the HCI over RPMSG
is used this dependency will not be solved automatically.
It is required to handle correctly Bluetooth identity
initialization.
Signed-off-by: Kamil Gawor <Kamil.Gawor@nordicsemi.no>
This platform does not use RISC-V machine timer, however, the interrupt
controller had references to its IRQ. Remove them.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Obtain machine timer addresses and IRQ from Devicetree. Note that driver
supports multiple compatibles because mtime/mtimecmp registers are
implemented in different ways depending on the vendor. That means
Devicetree representations can be slightly different and so code to
collect the information needs to treat each compatible differently.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Use the sifive,clint0 compatible instead of "riscv,clint0" and remove
interrupt controller fields (clint compatible is used for its timer
registers). Refer to the Linux or previous commits for more context.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Use the sifive,clint0 compatible instead of "riscv,clint0" and remove
interrupt controller fields (clint compatible is used for its timer
registers). Refer to the Linux or previous commits for more context.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The CLINT (Core Local Interruptor) description was not aligned with
Linux. For example, there's no "riscv,clint0", but "sifive,clint0". The
peripheral is not described as an interrupt-controller either.
Ref. https://elixir.bootlin.com/linux/v5.18.14/source/arch/riscv/boot/
dts/starfive/jh7100.dtsi#L106
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
After some analysis I found out that there's no machine timer provided
by the "riscv" vendor. There are some specs for the mtime/mtimecmp
registers (this is why we can have a single driver), but the actual
register layout or implementations differ amongst vendors. GD32 uses the
Nuclei implementation, named "system timer" in their documentation. This
patch aligns with vendor specs.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Has a lot of great sensors on it including a very fast 6 axis IMU,
atmospheric, and time of flight sensors.
This is a fantastic board to test sensors and algorithms on.
Signed-off-by: Tom Burdick <thomas.burdick@intel.com>
The unit for SENSOR_CHAN_GAUGE_TIME_TO_EMPTY and
SENSOR_CHAN_GAUGE_TIME_TO_FULL are defined in
zephyr/drivers/sensor.h as being in minutes.
Change the max17055 implementation to match the API spec, fix the
annotation for time register units.
Link: https://datasheets.maximintegrated.com/en/ds/MAX17055.pdf
Signed-off-by: Fabio Baltieri <fabiobaltieri@google.com>
Now that we properly destroy the tx context in `bt_conn_process_tx`, we
need to make sure the context is really valid.
Signed-off-by: Jonathan Rico <jonathan.rico@nordicsemi.no>
Symptoms were that a disconnect happening when a buffer was enqueued but
not sent resulted in a meta-data memory leak.
This is problematic because it seemed (as per the OP) that it resulted in
the whole host getting locked up when a device sending long L2CAP packets
was prone to disconnect-reconnect cycles (e.g. in a bad RF environment).
Fixes#47649
Signed-off-by: Jonathan Rico <jonathan.rico@nordicsemi.no>
Code was limiting observations to resources and
resource instances without any reason.
Also if resource is written, and the whole object is observer
it should trigger. Path does not have to be matching on the
same level, if parent is observed.
Signed-off-by: Seppo Takalo <seppo.takalo@nordicsemi.no>
The expression for the condition for using host-based RPA resolution now
simplifies to "Use host-based when we need to resolve more identities
than the controller can handle.".
Proof:
X=((bt_dev.le.rl_entries > 0) && (!bt_dev.le.rl_size ||
bt_dev.le.rl_entries > bt_dev.le.rl_size))
X=(a > 0) && (!b || a > b)
a>=0, b>=0, because they are cardinal / size_t.
If a=0:
// X evaluates to false. (0 > b)=(a > b) also always evaluates to false.
X=false=(0 > b)=(a > b)
If a>0:
X=(!b || a > b)
If b=0: X=true=(a > 0)=(a > b)
If b>0: X=(false || (a > b))=(a > b)
The expression is equivalent with (a > b) for all values of (a,b).
QED.
Signed-off-by: Aleksander Wasaznik <aleksander.wasaznik@nordicsemi.no>
The unicast audio server sample did not set a location value,
and would report an invalid value of 0 when read.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
This adds sample application that can be used as reference
implementation and for HAP qualification purposes.
Signed-off-by: Mariusz Skamra <mariusz.skamra@codecoup.pl>
Mesh Profile Specification errata 11310
Section 4.4.4.1
"This model shall support model publication, as defined in Section
4.2.2 of the Mesh Profile specification and model subscription,
as defined in section 4.2.3 of the Mesh Profile specification."
Signed-off-by: Michal Narajowski <michal.narajowski@codecoup.pl>
Spell out BT_ATT_ERR_* instead of using less specific "ATT error code".
Co-authored-by: Emil Gydesen <Thalley@users.noreply.github.com>
Signed-off-by: Nikolai Kondrashov <spbnick@gmail.com>
Name some types in bluetooth/gatt.h and use those names to document the
expected values of some of the BT_GATT_* macro arguments, in an attempt
to make understanding the API a little easier.
Signed-off-by: Nikolai Kondrashov <spbnick@gmail.com>
Refer to BT_GATT_WRITE_FLAG_* specifically, instead of the more general
BT_GATT_WRITE_* namespace to avoid future clashes with new symbols
there.
Signed-off-by: Nikolai Kondrashov <spbnick@gmail.com>
Move from using Kconfig GDBSTUB_SERIAL_BACKEND_NAME to a devicetree
chosen property ("zephyr,gdbstub-uart"). This is similar to a number
of other functions like "zephyr,shell-uart" or "zephyr,bt-uart".
Signed-off-by: Benjamin Björnsson <benjamin.bjornsson@gmail.com>
Enable a workaround for boards on which an issue has been reported.
Issue occurs only when serial works with virtual environment.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
Add workaround for development kits on which interface chip (segger chip)
is dropping bytes when they are sent in bursts. Issue has been seen on
DK which have multiple virtual com ports on the USB.
A workaround is adding periodic busy waits to introduce gaps in the
transmission.
Signed-off-by: Krzysztof Chruscinski <krzysztof.chruscinski@nordicsemi.no>
