When something goes wrong in Sphinx or extensions, we'll get a detailed
traceback, so we can easily debug issues in CI.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Add a new CMake option to append options to SPHINXOPTS. This allows us
to easily extend default SPHINXOPTS. This patch also restores the "-j
auto" option in CI (now that we use a custom runner).
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Run Sphinx in warnings as error mode, but keep going so that all issues
are reported to the user.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Supplement the documentation pm_device_is_powered()
adding that a device is assumed powered in case
the device does not support PM or is not on a power domain.
Signed-off-by: Marcel Krüger <marcel.krueger@ithinx.io>
The TLS credentials libraries are purely software constructs with no
external dependencies, run them immediately after the kernel setup to
allow other initialisation functions to add credentials without the
requirement to run in the back half of the `APPLICATION` priority.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
Fixes erroneous packing/unpacking of model app list messages
in the configuration client and server. According to the mesh 1.1
protcol spec (4.3.1.1) two app indexes shall be packed in a 3 octet
interleaved format. The current implementation packs them in 4 octets.
This commit also provide a helper function for unpacking key indexes
as public API to facilitate future config model callback API.
Signed-off-by: Anders Storrø <anders.storro@nordicsemi.no>
Adds a comment stating the state of the peripheral SDU size
and why it is being assigned the PDU size on CIS established.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
This patch adds support for PWM blink which is found in intel's
PCH hardwares.
Signed-off-by: Anisetti Avinash Krishna <anisetti.avinash.krishna@intel.com>
Now that input dump functionalities are at subsystem level, drop the
custom code from the sample, just enable CONFIG_INPUT_EVENT_DUMP.
Signed-off-by: Fabio Baltieri <fabiobaltieri@google.com>
Add two input subsystem options: one for dumping input events, one for
triggering input reports from the shell.
The two are independent from each other, so dumping can be enabled on
any application that has no shell, but if the shell command is present
logging can be toggled on and off with a shell command.
Signed-off-by: Fabio Baltieri <fabiobaltieri@google.com>
Changes the Kconfig option to allow disabling mutex support, this
is to allow other Kconfig options to disable the feature.
Signed-off-by: Jamie McCrae <jamie.mccrae@nordicsemi.no>
Changes the Kconfig option to allow disabling mutex support, this
is to allow other Kconfig options to disable the feature.
Signed-off-by: Jamie McCrae <jamie.mccrae@nordicsemi.no>
The bt_ots_init struct has been renamed to bt_ots_init_param in commit:
cf0ff30b53
That was previously triggering a warning in Sphinx, due to the same
name being used for a struct and function, and this pattern
was designed to silence it. We don't need the pattern any more as the
warning is gone.
Signed-off-by: Michael Jones <m.pricejones@gmail.com>
Change the default mbox initialization priority so that it initializes
before ipc, as some ipc drivers depends on it.
Found with:
$ west build -p -b nrf5340dk_nrf5340_cpuapp \
samples/subsys/ipc/ipc_service/icmsg_me \
-DCONFIG_CHECK_INIT_PRIORITIES=y
...
ERROR: /ipc/ipc1 POST_KERNEL 46 <
/soc/peripheral@50000000/mbox@2a000 POST_KERNEL 50
ERROR: /ipc/ipc0 POST_KERNEL 46 <
/soc/peripheral@50000000/mbox@2a000 POST_KERNEL 50
Signed-off-by: Fabio Baltieri <fabiobaltieri@google.com>
The underlying HAL driver may improperly forward an samplerdy event even
if it's disabled in the configuration. Ignore the event to prevent error
logs until the issue is fixed in HAL.
Signed-off-by: Marek Pieta <Marek.Pieta@nordicsemi.no>
remove the sdk based TOTAL_DMA_CHANNELS macro and instead just use the
zephyr driver's num_of_channels field
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>
Current erroneous usages of dma-channels prop by lpc-dma nodes:
* dma-channels devicetree property should describe the number of
channels supported by the dma controller, not the number of channels
in use.
* LPC55SXX and RTXXX SOCs should be setting dma channels prop at SOC
level, not board level, since it is an SOC property, not a board
property.
* lpc55s28 has 23 channels for dma0, not 20.
* lpc55s28 has 10 channels for dma1, not 0.
* lpc55s69 has 23 channels for dma0, not 20.
* rt5xx has 37 channels for dma1, not 0.
* rt6xx has 33 channels for dma0, not 20.
* rt6xx has 33 channels for dma1, not 0.
Fix all of these issues
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>
Set the DMA number of otrigs DT property at the SOC level instead of the
board DTS because it is an SOC property and does not change on different
boards.
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>
Add a user possibility to set a callback for receiving a new function.
It allows instant performing some actions, that need to be done before
context switch.
Signed-off-by: Dawid Niedzwiecki <dawidn@google.com>
Verify validity of a received command before passing it to the general
handler.
It allows performing some actions, right after receiving the command.
The context switch is not needed. Such feature may be needed for
overloaded system, where instant reboot is required.
Signed-off-by: Dawid Niedzwiecki <dawidn@google.com>
Add a function to signal a new host command by a backend.
Use a function instead of giving semaphore, because it allows more
actions on rx event, common for all backends.
Signed-off-by: Dawid Niedzwiecki <dawidn@google.com>
Add a config to save the final result of a last host command that has
sent EC_HOST_CMD_IN_PROGRESS response. To get the final result use the
ec_host_cmd_send_in_progress_status function.
Signed-off-by: Dawid Niedzwiecki <dawidn@google.com>
Add Arduino compatible LoRaWAN shield based on SX1276 tranceiver from
Semtech. Provide minimal documentation on how to use it.
Tested with `nucleo_l073rz` platform.
Signed-off-by: Marcin Niestroj <m.niestroj@emb.dev>
When reading multiple ADC channel in parallel, an ADC command chain will be
build. This is similar to a linked list, as every command references the
next command.
Before this patch every ADC command after the first, would always reference
this initial command. So that during execution only two commands (the last
and first) would be executed which resulted in readout of only two analog
values. As Zephyr expected more to come in, the `read_adc` function would
block endlessly.
The patch fixes the behaviour and allows a correct chain to build up.
Signed-off-by: Elisabeth Friedrich <auxsys@flowerpot.me>
PCIe devices refer to interrupt nodes, but are initialized with the same
priority, making the sequence depending on the linking order. Add a new
symbol and set it to one unit after intc to ensure that the
initialization sequence is stable.
Found with:
west build -p -b qemu_cortex_a53 \
samples/drivers/virtualization/ivshmem/doorbell \
-DCONFIG_CHECK_INIT_PRIORITIES=y
...
ERROR: /soc/pcie@4010000000 PRE_KERNEL_1 40 <
/soc/interrupt-controller@8000000/its@8080000 POST_KERNEL 40
Signed-off-by: Fabio Baltieri <fabiobaltieri@google.com>
The intc_gicv3_its driver is currently initializing in POST_KERNEL, but
PCIe device nodes depends on it and are initialized earlier. Change the
init level to PRE_KERNEL_1 to sort that out, and also put the driver in
line with other interrupt controller drivers.
Signed-off-by: Fabio Baltieri <fabiobaltieri@google.com>
Multiple reader threads unlocking the read lock simultaneously might
cause the program hang because it's possible that no thread is
identified as the last one to active the writer thread.
To fix the issue, swap the k_sem_give sequence.
Signed-off-by: Jaxson Han <jaxson.han@arm.com>
