The BT_HCI_TX_STACK_SIZE is carefully calculated from other
options. When those options change, e.g. from a menuconfig update, it
is important that the stack size is re-calculated, but it is not when
there is a prompt.
Therefore, make the prompt conditional such that the previously set
value is only used when it has been explicitly configured to be so.
Now users can still change the value through menuconfig and prj.conf,
by also enabling <option>_WITH_PROMPT, but when the value is
calculated by the defaults, it will continue to be calculated by
defaults instead of inheriting the intial value.
This is AFAIK a novel approach, but testing has shown that it gives
the users the behaviour they want, at the cost of some boilerplate of
course. This pattern can be applied to other options if it proves to
work as intended.
Alternatively one could remove the prompt, but then it would no longer
be possible to override the value through menuconfig.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
When a MetaIRQ preempts a cooperative thread, that thread would be
added back to the generic run queue. When the MetaIRQ is done, the
highest priority thread will be selected to run, which may obviously
be a cooperative thread of a higher priority than the one that was
preempted.
But that's wrong, because the original thread was promised that it
would NOT be preempted until it reached a scheduling point on its own
(that's the whole point of a cooperative thread, of course).
We need to track the thread that got preempted (one per CPU) and
return to it instead of whatever else the scheduler might have found.
Fixes#20255
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Add documentation for the possible HCI errors codes received for a
connected callback. The HCI error code received when the initiator is
canceled through the HCI create conn cancel operation is non-intuitive.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Fix for not served error code in fcb_offset_last_n()
implementation.
Issue was reported by static code analyze scan.
Looks like the function might have failed silently
if aggressive access to the fcb instance
from another thread have been occurring while the function
have been processing the fcb storage.
fixes#20512
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
Add a simple block diagram detailing the SMP initialization flow. Not
pretty, but hopefully reasonable.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
This commit modifies the struct array initialisation expression using
the "universal initialiser" to a struct single element initialisation
expression in order to prevent old GCC versions from erroneously
reporting the "missing braces around initializer" warning.
This is an infamous bug in old GCC versions; since there is no side
effect of using this alternate initialisation expression and it does
not violate the C99 standard, this problem should be addressed in the
code for the sake of portability.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
Add support for boards that utilize the TI XDS110 like the CC3220SF,
CC3235SF, CC1352R1 and CC26x2R1 LaunchXL boards. The XDS110 can expose
multiple serial endpoints one for the uart device, but another for a
trace buffer. We assume that endpoint 0 will be the UART device.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
Add setting runner type to jlink for 'J-Link' and openocd for
'STM32 STLink' when generating the hardware map.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
Documentation about scheduling options was burried in the Kconfig help.
It has better visibility as part of the scheduling section of the main
kernel reference pages.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Add a flag for identifing whether this object is on the trace
list. Ensure that link any object to the trace list only one time.
It will avoid the issue about lost object caused by adding a
object to trace list twice.
Fixes#19537
Signed-off-by: Shih-Wei Teng <swteng@andestech.com>
Get this commit in:
check_compliance.py: Run the dtlib/edtlib test suites on Python 3.5
Will work once
https://github.com/zephyrproject-rtos/zephyr/pull/20597 is in. CI
uses Python 3.5, so they have been skipped until now.
That PR is in now.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
Commit 555936f771 ("dts/Kconfig: Remove unused HAS_DTS_SPI")
accidentally removed the wrong #endif, meaning the second part of
dts_fixup.h depended on DT_ADXL372_DEV_NAME not being defined.
Discovered through the reference to the undefined HAS_DTS_SPI symbol in
the comment after the #endif at the end of the file. Adding detection of
unused symbols in samples and tests to CI.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
A reference to CONFIG_CAN_AUTO_BOFF_RECOVERY was added in commit
1b88658f9f ("samples: driver: Extend CAN sample"), but it's never been
defined as a Kconfig symbol.
Should have been CONFIG_CAN_AUTO_BUS_OFF_RECOVERY according to
alexanderwachter, so change it to that.
Adding detection of unused symbols in samples and tests to CI.
Signed-off-by: Ulf Magnusson <Ulf.Magnusson@nordicsemi.no>
When Config client model send publish message, app_idx will
be need, however, currently code clear this value use `&`,
this will be generate error when app_idx not zero.
Signed-off-by: Lingao Meng <mengabc1086@gmail.com>
Corrected spellings in printk messages plus removed
unnecessary blank lines.
In case of Gen. Move client's message to Server, reduced
values of delta & transition time to improve demonstration.
Signed-off-by: Vikrant More <vikrant8051@gmail.com>
We need to fix the formato of the macros for nRF CACHE and NVMC
peripherals in soc.c, so the _S and _NS suffixes are skipped.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
Since the nRF5340 DK contains the required DC/DC circuitry,
enable it by default while keeping it configurable. This
reduces power consumption in application core , network
core and high voltage use.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
This makes the misc-flasher runner usable by passing the build
directory to the underlying tool.
Fixes: #20658
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
One shows how globals are routed for automatic memory
domains. The other illustrates control flow when making
system calls.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Signed-off-by: David B. Kinder <david.b.kinder@intel.com>
The device address can only be 0x18 through 0x1F. C6 is connected to
the FXOS870 and is not exposed on a header: switch to Arduino D0.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
The i2c_msg API details do not work correctly with Nordic TWI. Switch
to the higher-level and simpler API for register read and write. Also
add a tree configuration on a Nordic-based board.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
There is possibility that endrx interrupt will be triggered in the
middle of timer interrupt responsible for UARTE timeout, this
patch handles this case.
Signed-off-by: Mieszko Mierunski <mieszko.mierunski@nordicsemi.no>
Commit 16d8ce519c introduced changes
that caused this sample to no longer behave according to documentation
and for some hardware to no longer work at all.
On nRF51 SoCs for instance, the reported number of cycles per second
is 16M, what makes the calculated max_period and min_period to be 16
and 0 microseconds, respectively, what effectively makes it impossible
for the sample to return to the initial blinking frequency. Moreover,
with such short PWM periods, the blinking is not even noticeable.
This patch partially reverts the changes mentioned above, and instead
of calculating max_period and min_period basing on the reported clock
rate, it tries to only decrease the max_period if needed, accordingly
to what the used hardware can handle.
Documentation is also updated to mention the possible change in
observed behavior of the sample on some hardware.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
LSM303AGR is a special one in terms of raw value scale among all devices
supported by lis2dh.c driver. Apply proper scale factor based on
sensitivity scale provided in LIS2DH and LSM303AGR datasheets.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Add new bindings for already supported (by lis2dh.c driver) lsm303agr
accelerometer. Using st,lsm303agr-accel compatible will allow in
subsequent commit to apply different sensitivity scale.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
All values were scaled by 8 instead of 12 for 16g compile-time selected
range. This resulted with values around 6.5 m/s^2. Scaling for runtime
configurable ranges was broken for all except 2g range option due to
totally broken lis2dh_range_to_reg_val().
Fix wrong scaling for 16g compile-time and all runtime selectable ranges
by reworking code that scales raw value. While doing this, change
lis2dh->scale type from 16 to 32 bits. This allows to slightly increase
final result precision by using the fact that raw values have maximum 12
bits precision, allowing us to multiply lis2dh->scale by (1 << 4)
compared to previous implementation.
Fix bug #19872.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
According to struct sensor_value documentation, val2 should be negative
for negative result. So drop code that tries to make val2 positive.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
In settings write NVS errors were not handled
in a few places.
This patch improve that.
fixes#20515
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
As presented to the TSC, Zephyr's out-of-box experience for new
developers is, well, complicated. A number of suggestions were
presented including simplifying the getting started material to present
a straight-forward path through the setup and installation steps through
to getting a sample application built, flashed, and running.
This PR is a work-in-progress towards addressing this OOB experience
with a minimal-distractions version of the GSG. Alternatives, warnings,
and material that could lead the developer astray were moved to
alternative/advanced instruction documents (based on the previous
separate Linux/macOS/Windows setup guides) and a new "Beyond the GSG"
document.
We do take advantage of a sphinx-tabs extension for synchronized tabs to
present OS-specific instructions: clicking on one tab will display all
same-named tabs throughout the doc.
We hope (and will continue evaluating) that this new GSG gets developers
set up quickly and then we can send them along to other documents to
continue learning about Zephyr and trying other sample apps.
Thanks for all your previous feedback that I've worked
into this new version.
Signed-off-by: David B. Kinder <david.b.kinder@intel.com>
in SMP, arc uses gfrc as wall clock, so it's ok
to enable SYSTEM_CLOCK_SLOPPY to avoid unnecessary
interrupts
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
* fix the smp timer dirver bugs found in debug and test.
for smp case, GFRC is used as clock source, and local
internal timer is used to trigger time event.
* because 64-bits gfrc is used, so idle can be igored as no kernel
tick will be missed
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
It's found that in nsim_hs_smp, sometimes the cpu
doesn't response inter-core interrupt after executing sleep
instruction.
It may be a bug of nsim, but needs more time to
investigate the root of this issue.
This commit is a workround for this, as nsim is just an
instruction simulator, no direct impact.
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
* necessary fixes after commit 11bd67db where ipi interrupt is used
to notify other cores to do a thread switch if necessary
* then for arc, it's needed to ignore swap_ok and check whether thread
switch is needed in the exit of irq handling.
Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
Explicitly configures the rgb led pinmuxes as gpios. Currently the gpio
driver quietly changes the pinmux to gpio mode when configuring a gpio
pin, but this behavior is about to change.
Signed-off-by: Maureen Helm <maureen.helm@nxp.com>
Some board description files failed to note where gpio was supported,
causing tests to be inappropriately filtered. Add the feature where
the gpio_basic_api test would use it.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Add device tree elements for all gpio ports of the efr32mg12p including
the dts fixup entries.
Also remove gpio port e since this is not available in efr32mg12p socs.
Signed-off-by: Christian Taedcke <hacking@taedcke.com>