Adds conditional usage of locking mechanisms to allow building
without multithreading.
Signed-off-by: Mateusz Michalek <mateusz.michalek@nordicsemi.no>
Now RRAMC is set to use up to 512 B burst writes by default as most
time effective.
Requested slot time was changed to 8000 us for that case, as this is
required in order to hold the longest write operation.
Drivers differs slot duration depending on configured RRAMC buffers count
(CONFIG_NRF_RRAM_WRITE_BUFFER_SIZE).
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
When radio-synchronization was on
(CONFIG_SOC_FLASH_NRF_RADIO_SYNC_NONE=n) context of writing was shifted
by 4 instead of write-length (this caused redundant, self-shortening
writes to the same block.)
write-len was unset when requested write len was less than
RRAM_MAX_WRITE_BUFFER.
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
Signed-off-by: Jakub Zymelka <jakub.zymelka@nordicsemi.no>
Added synchronization with BLE radio operation to the RRAM flash
driver. The implementation is using framework which is already
provided for nrf52's flash driver.
Additional added resource locking mechanism while driver does writing
which solves mutual exclusion write access problem.
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
Signed-off-by: Jakub Zymelka <jakub.zymelka@nordicsemi.no>
Added a simple driver for RRAM. It is implemented as a flash driver,
because the "RRAM eFlash" macro obeys flash-like constraints.
Although users are not required to erase before write.
Signed-off-by: Jakub Zymelka <jakub.zymelka@nordicsemi.no>
The test_mqueue_notify_thread testcase had a relatively high
failure rate because sleeping only 10ms was in some cases not
sufficiently long enough for the spawned thread to update
notification_executed.
Sleep 100 ms instead of 10 ms to reduce the failure rate.
Signed-off-by: Christopher Friedt <cfriedt@meta.com>
Instead of rounding up both __tdata_size and __tbss_size at runtime,
perform the calculation when the image is built.
Signed-off-by: Peter Mitsis <peter.mitsis@intel.com>
The print_sirk function that simply logs the SIRK has been removed,
as the SIRK is now directly accessible to the application via the
get_sirk function.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Add driver for ScioSense ENS160 multi-gas sensor. The driver includes
support for I2C and SPI, attributes for setting temperature and
humidity compensation and data ready trigger.
Also add ScioSense to the list of vendor prefixes.
Signed-off-by: Gustavo Silva <gustavograzs@gmail.com>
The presets used BT_AUDIO_CODEC_QOS directly, but since all the
unframed presets use BT_AUDIO_CODEC_QOS_UNFRAMED, it makes sense
that the framed ones used the corresponding macro.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
The LC3 framed macros were unused.
The LC3 unframe macros only added the 7500/10000
parameter, but since the the 7500/10000 requirement is also
defined by BAP, it makes equally much sense to simply use the BAP
presets.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
LC3 does not define any such configurations, and they are in
fact defined by the BAP. The LC3 header file should thus not
pretend that there are these specific configurations for LC3.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Process the GPIO_INT_WAKEUP flag and set appropraite bits
in the SoC to wakeup the system from deep sleep mode.
Signed-off-by: Mahesh Mahadevan <mahesh.mahadevan@nxp.com>
Adds the CAP Commander Change Volume mute procedure.
This procedure changes the volume mute on one or more
CAP Acceptors.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
This moves to more recent versions of Sphinx and Read the Docs theme.
FWIW Sphinx 7.2.0 dropped support for Python 3.8 so we're stuck with 7.1
for now.
Signed-off-by: Benjamin Cabé <benjamin@zephyrproject.org>
CDP1 is mandatory to qualify a mesh device for mesh protocol v1.1.
The dependency on BT_MESH_MODEL_EXTENSIONS is removed because if a node
contains only base models (for example Generic OnOff Server), this API
will consume extra RAM bytes which won't be used.
Signed-off-by: Pavel Vasilyev <pavel.vasilyev@nordicsemi.no>
If the CONFIG_NET_SOCKETS_POLL_MAX is smaller than what is needed
for the socket service API to work properly, then we should not
start the service thread as the service API cannot work and might
cause memory overwrite in ctx.events[] array.
Fixes#69233
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
40k heap size is not enough and results in handshake errors:
<err> net_sock_tls: TLS handshake error: -0x2700
Increasing it to 55k already works (when used with native_sim_64 and
offloaded sockets), but use 60k to get a bit more room for future.
Signed-off-by: Marcin Niestroj <m.niestroj@emb.dev>
The network tests at tests/net use simulated network interfaces
and set CONFIG_NET_TEST to indicate that. If the config option
is set, then we do not want any extra Ethernet driver to
complicate the testing scenario so all external Ethernet
network interfaces should be disabled.
Signed-off-by: Jukka Rissanen <jukka.rissanen@nordicsemi.no>
This commit does two things to k_wakeup():
1. It locks the scheduler before marking the thread as not suspended.
As the the clearing of the _THREAD_SUSPENDED bit is not atomic, this
helps ensure that neither another thread nor ISR interrupts this
action (resulting in a corrupted thread_state).
2. The call to flag_ipi() has been removed as it is already being
made within ready_thread().
Signed-off-by: Peter Mitsis <peter.mitsis@intel.com>
Cast val1 from sensor_value before multiplication in order to avoid
integer overflow, as indicated by Coverity CID 347136.
Signed-off-by: Renato Soma <renatoys08@gmail.com>
nvs-cache-gc test is running into infinite loop because of the wrong
stop condition when filling a sector.
Fix this by keeping an empty ATE in the sector for delete operation as
defined in the NVS filesystem write operations.
Signed-off-by: Riadh Ghaddab <rghaddab@baylibre.com>
This is obviously a copy & paste error. It was introduced by
107cb86bb3 (ztest: Add initial zexpect API
for delayed failing).
Signed-off-by: Reto Schneider <reto.schneider@husqvarnagroup.com>
PR #64399 introduced checks for out-of-bounds filter IDs
in CAN drivers, along with logging of said IDs; however,
the call to LOG_ERR in the native POSIX/Linux driver is
missing the 'filter_id' argument.
This commit adds the missing argument to ensure proper
data is printed when the LOG_ERR call is performed.
Signed-off-by: Mathieu Choplain <mathieu.choplain@st.com>
Generate one single status for all jobs and also report success when
twister run is being skipped to be able to make this a required check.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
In Zephyr, things are often optimized for size first. That's how
we fit into such tight parking spaces.
This change gives more control to the user about whether the
POSIX API does any logging at all, simultaneously shrinking binary
size while improving speed.
No bytes / cycles left behind!
Signed-off-by: Christopher Friedt <cfriedt@meta.com>
Adds v3.6.0 to the list of supported releases. The EOL date corresponds
to the expected release date for v4.0.0.
Signed-off-by: Maureen Helm <maureen.helm@analog.com>
