Use the common functions instead of
replicating functionality and do a minor cleanup
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Use the common functions instead of
replicating functionality and do a minor cleanup
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Use the common scripts to perform common functions
and do a cleanup.
Do not set variables to their defaults, set things
closer to were they are needed;
Do not perform unnecesary checks;
And in general avoid unnecessary complexity.
Some of the _compile scripts were rotten => Fix them.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Ensure we call the common source script before using
the BOARD variable;
Use the common scripts to perform common functions
and do a cleanup.
Do not set variables to their defaults, set things
closer to were they are needed;
Do not perform unnecesary checks;
And in general avoid unnecessary complexity.
And for the 2 parts of the peripheral rpa sharing
test, use different sim_ids, so the traces for each
part can be checked.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Ensure we call the common source script before using
the BOARD variable;
Use the common scripts to perform common functions
and do a cleanup.
Do not set variables to their defaults, set things
closer to were they are needed;
Do not perform unnecesary checks;
And in general avoid unnecessary complexity.
And return !=0 on error.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Use the common scripts to perform common functions
and do a cleanup.
Do not set variables to their defaults, set things
closer to were they are needed;
Do not perform unnecesary checks;
And in general avoid unnecessary complexity.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Ensure we call the common source script before using
the BOARD variable, and let's not set things
which are already set to the same by default
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Use the common scripts so we have a consitent way of
running the tests, report failures in the same way,
can stop them with ctrl+C, properly wait for
all processes to exit..
Let's also place the test scripts in the test_scripts/
folder as it is convention.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
(f) ests/bsim bt att: Use common s
Ensure we call the common source script before using
the BOARD variable;
Use the common scripts to perform common functions
and do a cleanup.
Do not set variables to their defaults, set things
closer to were they are needed;
Do not perform unnecesary checks;
And in general avoid unnecessary complexity.
Also, remove all timeouts which were unnecessarily too short
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Some developers want to programmatically get the
test name instead of hardcoding it.
So let's have a common function for this.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
As now, with hwmv2, boards can have "/" in their name,
and the board name is not anymore the complete target
name, let's set a new variable with the
whole board target name, but converted into the format
which can be used in paths.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
This test was way too verbose (produced a log over 65000 lines
long).
Let's mute it a bit. If the test fails, developers
can raise the log level as needed.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
arch_interface.h is for architecture and should not be
under sys/. So move it under include/zephyr/arch/.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This functionality is useful on the following scenario:
1) The first buffer in a net_pkt contains user data which is relevant
for the (whole) net_pkt.
2) When inserting a new buffer in front of the net_pkt, the (previously)
first buffer (and its user data) are no longer accessible via
net_pkt->buffer.
3) net_buf_user_data_copy() allows to simply copy the user data from the
old to the new first buffer.
Signed-off-by: Reto Schneider <reto.schneider@husqvarnagroup.com>
debug.coredump.logging_backend is currently failing for most of the targets
I have tried when building with clang. Depending on the target, func_3
invokes undefined behavior by dereferencing addr (which is NULL) which
can lead to the compiler optimizing out significant portions of the code,
resulting in unexpected/incorrect failures.
Here, clang seems to inline func_3 into main then marks the inlined
implementation as unreachable (due to the UB) and removes it and everything
after it in main. So, we fall through to whatever code lies past main,
resulting in a test failure (timeout) from what I've seen. GCC seems to do
similar things, but creates an invalid opcode instruction so the test still
succeeds.
clang is correct in both optimizing this behavior out and leaving
buggy code behind, so disable optimizations for func_3 to keep things under
control and prevent the incorrect failures.
Signed-off-by: Jonathon Penix <jpenix@quicinc.com>
The riscv_virtual, hifive_unleashed, and hifive_unmatched boards all
timeout on this test if they actually perform the null dereference when
they are simulated. Whether the null dereference happens seems to depend
on the compiler used as well as whether optimizations are enabled though.
To make this more consistent, handle these boards in the same way as others
which use Renode simulation and have them directly call k_panic().
Signed-off-by: Jonathon Penix <jpenix@quicinc.com>
Before this commit, cloning a buffer was tested only for a buffer
belonging to a reference counted pool.
By using a buffer from a pool that does not support reference counting,
the added test ensures that the non-ref-counting code path also works.
Signed-off-by: Reto Schneider <reto.schneider@husqvarnagroup.com>
Trying to enable LTO when CODE_DATA_RELOCATION is already enabled
produces a warning.
This commit prevents twister from building such combinations for
arch.shared_interrupt.lto.
Signed-off-by: Reto Schneider <reto.schneider@husqvarnagroup.com>
Trying to enable LTO when CODE_DATA_RELOCATION is already enabled
produces a warning.
This commit prevents twister from building such combinations for
kernel.common.lto.
Signed-off-by: Reto Schneider <reto.schneider@husqvarnagroup.com>
The native_{posix,sim}_64 boards are not anymore twister
targets, native_{posix,sim}/native/64 are, and those
default to just use the overlays for the non variant
version.
These 2 overlays were not used by anything.
Remove them.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
This test is an attempt at formalizing at least part of the behavior
described in commit 6a79c3deae.
Signed-off-by: Aleksander Wasaznik <aleksander.wasaznik@nordicsemi.no>
This function waits for a free connection slot. It should be used to
slow down a busy loop trying to obtains a connection slot.
Signed-off-by: Aleksander Wasaznik <aleksander.wasaznik@nordicsemi.no>
Client ID option is now expected in Request/Ack responses so tests need
to be updated accordingly.
Signed-off-by: Robert Lubos <robert.lubos@nordicsemi.no>
Add a test for confstr(), which is required by the
POSIX_SINGLE_PROCESS Option Group, as per IEEE
1003.1-2017.
Signed-off-by: Christopher Friedt <cfriedt@meta.com>
After https://github.com/zephyrproject-rtos/zephyr/pull/70438
got merged simultaneously with
https://github.com/zephyrproject-rtos/zephyr/pull/70564
all these tests stopped building properly due to the
change of avaliable variables in Kconfig & cmake.
Let's fix them.
As a bonus, as for kconfig the BOARD_TARGET_STRING is the
same for the hwmv1 backwards compatible name, let's
just remove the check for the old names.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
The PBP bsim tests had several issues that was not uncovered
until the PA sync timeout was increased. This commit fixes these,
as well as increase some timeout values due to recently
increased PA sync timeout.
Among the issues were lack of support for the alternating PBA
features, missing reset of values and specific time sleeps.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Increase the timeouts to be 60 instead of 6.
The reason for this is that some controllers (like the Zephyr
controller) will reserve some of these to ensure stability,
and in fact with skip = 5 and retry = 6, it would still
send the PA reports at every interval.
To accomodate a higher timeout value, the functions used to
convert PA intervals to PA timeouts have been updated.
Signed-off-by: Emil Gydesen <emil.gydesen@nordicsemi.no>
Replace all instances of
nrf5340bsim_nrf5340_cpunet & nrf5340bsim_nrf5340_cpuapp
with
nrf5340bsim/nrf5340/cpunet & nrf5340bsim/nrf5340/cpuapp
In these samples/tests twister yamls definitions.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
For all samples which handle the nrf5340bsim in some special way
(for example in sysbuild files, or by checking what samples are
supported) handle also building with the hwmv2 names.
Also, let's fix the cmake info message in these respective
samples which tells the user which board target is used for which
part of the app.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Mesh v1.1, DFU v1.0 and MBT v1.0 are supported by default in the stack.
In this regard, this commit updates the ICS file and Launch Studio
Project file to reflect supported features.
Signed-off-by: Alperen Şener <alperen.sener@nordicsemi.no>
This deduplicates a copy of testlib from before testlib was a globally
available cmake library.
Signed-off-by: Aleksander Wasaznik <aleksander.wasaznik@nordicsemi.no>
Enable CTimer0 and MRT0 channel 0 by default on the board,
and add more of the device instances in the counter api test overlay.
Signed-off-by: Declan Snyder <declan.snyder@nxp.com>
The macro is searching for all instances of specific device
that contain specific property and evaluates to true (1) if
any device does.
The macro used to do that by generating, using
DT_ANY_INST_HAS_PROP_STATUS_OKAY, a logical expression
like (0 || 1 || 0), where each digit represented existence of
property (1) or lack of it (0).
Unfortunately other util macros, like IS_ENABLED, were
not able to evaluate such expression, as they often simply
expect something they can evaluate to 0 or 1.
The commit here changes DT_ANY_INST_HAS_PROP_STATUS_OKAY
to generate a list of tokens (1) where token is added to list
only for instance of a device that has the property;
then such list is processed using IS_EMPTY() macro and
in the end 0 or 1 is generated, depending on whether
any enabled instance of a device has the property or not.
This change allows result of DT_ANY_INST_HAS_PROP_STATUS_OKAY
to be used with macros like IS_ENABLED, IF_ENABLED or
COND_CODE_x.
Signed-off-by: Dominik Ermel <dominik.ermel@nordicsemi.no>
* As we are replacing native_posix with native_sim, let's
refer to native_sim instead of native_posix in the comments
of why we have 1 extra interface.
* scripts/net/run-sample-tests.sh builds for native_sim now,
not native_posix => let's fix it
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
