For the standard multi-theading builds, we will
enforce FP context stacking only when FPU_SHARING
is set. For the single-threading use case we enable
context stacking by default.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
If CONTROL register is done in reset.S we can skip
clearing the FPCA when enabling the floating point
support, to save a few instructions. The CONTROL
register is cleared right after boot, if the symbol
CONFIG_INIT_ARCH_HW_AT_BOOT is enabled.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
There was an edge case in the timeout handling (exposed by, but not
strictly related to, the recent timeslice fix): the next_timeout()
computation would include time slice expiration as a clamp on the
result, but this would be invoked also on the z_set_timeout_expiry()
path which gets hooked on entry to a new thread which is needed to set
the timeout in the first place. So if no other timer interrupt was
scheduled, it was possible to miss the first timeslice interrupt after
thread scheduling.
The explanation is much longer than the fix (use <= as the comparator
instead of <).
In practice this was only being hit in the existing test suite on
riscv miv running under renode using non-default clock rates.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Fix an edge case that snuck in with the recent fix: if timeslicing is
enabled, the CPU's slice_ticks will be zero, and thus match a timeout
object's dticks value of zero, and thus get suppressed (because "we
already have a timeout scheduled for that") incorrectly.
Fixes#31789
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
The devicetree only provides PWM-compatible LEDs. Remove the aliases
that suggest it supports GPIO-compatible LEDs.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
socket_can_generic.h makes some assumptions that are not always valid
for various drivers with regards to numbering. To clean this up we
add defines for SOCKET_CAN_NAME_0 and SOCKET_CAN_NAME_2 in addition to
SOCKET_CAN_NAME_1.
We also move struct socket_can_context into the drivers themselves
since they know best how many CAN interfaces are getting supported and
what naming/number convention they'd have.
Additionally, this also exposes a few other build issues with the
can_mcux_flexcan driver that get fixed.
Finally, we remove the platform_allow from samples/net/sockets/can
since it is no longer needed.
Fixes#31733
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
Fixes: #30591
Now using the `logical_target_for_zephyr_elf` variable for memusage
report.
The `logical_target_for_zephyr_elf` variable will point to the final
elf when two stage linking is done.
On single stage linking the prebuilt target == final elf, in which case
the `logical_target_for_zephyr_elf` points directly to the prebuilt
target.
Thus using `logical_target_for_zephyr_elf` for memusage report ensures
the correct elf file is always used for the memory usage report.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Attempts to disable PM_STATE_SOFT_OFF would index past the end of the
count array. Increase the array length to allow operations on
PM_STATE_SOFT_OFF.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
This patch replaces magic numbers with COAP_TOKEN_MAX_LEN value and
removes unnecessary castings of token buffer type.
Signed-off-by: Hubert Miś <hubert.mis@nordicsemi.no>
RFC 7252 (CoAP) specifies value of the Version (Ver) field in the
protocol header to value 1. This patch defines value of the Version
field to make packet initialization easier. All samples and tests
are updated to use the new COAP_VERSION_1 field when initializing
a CoAP packet.
Signed-off-by: Hubert Miś <hubert.mis@nordicsemi.no>
ESP fetches DNS servers from local network by using DHCP. There is an AT
command to get those DNS addresses. Use that to provide DNS addresses
for Zephyr's DNS resolver.
Signed-off-by: Marcin Niestroj <m.niestroj@grinn-global.com>
Improve the DTS overlay description by adding information that a
`boards/<BOARD>_<REVISION>.overlay` will be merged with
`boards/<BOARD>.overlay` if both files exists.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
This is a follow-up to commit 77ecd6837d.
Add missing check that looks for a revision specific overlay located
in the `<app>/boards/` directory similar to how overlays and Kconfig
fragments are handled.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Adds API reference for sys_mutex and futex to mutex documentation,
adds Doxygen documentation for SYS_MUTEX_DEFINE and fixes typo in
futex documentation.
Fixes#27829
Signed-off-by: Lauren Murphy <lauren.murphy@intel.com>
The RunnerConfig class stores the locations of the Zephyr output files
in various formats (elf, hex, bin). A longstanding issue with the
representation is that these might not exist if the corresponding
Kconfig options are not set. For example, if
CONFIG_BUILD_OUTPUT_BIN=n, there is no .bin file.
Change this so the type system knows these are Optional[str], not str.
Fix the runners that use non-ELF outputs so they check for the
existence of the relevant file before using it, mostly using a new
ZephyrBinaryRunner.ensure_output helper.
I'm not going to bother with checking for the ELF file itself; that's
always there as far as I can tell.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
Fix Tx Buffer Overflow caused by uninitialized node_tx
memory being used by ULL ISR context due to Compiler
Instructions Reordering in the use of MFIFO_ENQUEUE.
The MFIFO last index was committed before the data element
was stored in the MFIFO due to Compiler Instructions
Reordering.
This is fixed now by adding a Data Memory Barrier
instruction alongwith a compiler memory clobber.
Fixes#30378.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Use of Data Memory Barrier instruction with memory clobber
in ARM Cortex M architectures is sufficient in the
controller implementation to keep compiler data access
instructions in order so that an ISR vectoring has memory
accesses in the correct order as intented by design.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
The inheritance test first creates a parent thread, and then
creates a child thread inside this parent thread. At the same
time inside the main thread, a resource pool is assigned to
the parent thread. However, when under SMP, it is possible
that the pool assignment is done between both parent/child
threads are getting pointers to the pool, due to multiple
threads are running. So when doing pool pointer comparison,
there is a mismatch between those pointers (like parent has
a null pool pointer while child is pointing to the actual
pool), and thus failing the test. So fix this by delaying
the parent from running under after pool assignment is done.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
When coverage is enabled on x86_64, GCC uses relative addressing
to increment the gcov counters. The generated code of the test
function assumes execution is in the same location where
the linker places the test function. This does not work with
the execution test as it copies the function into another part
of memory and tries to execute from there. When the copied
function starts to run, the instruction pointer is at the newly
copied function. So any relative addressing with regard to
the instruction pointer now is invalid. Instead of
<generated code RIP + offset> for gcov counter as it should be,
now the copied code is trying to access the counter at
<copied code RIP + offset>, which points to incorrect
memory location (and possibly invalid/non-mapped memory).
To fix this, we need to tell GCC not to use relative addressing.
This can be accomplished by telling GCC to use the large memory
model. This is only used for this test as this option increases
code size quite a bit, and should not be used in general.
Fixes#30434
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This puts the transplanted_function into its own section so that
z_phys_map() can correctly map the whole range of memory used
by the function, in case someone decides to expand the function
to be bigger than a MMU page.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
To give consistency with nrf91. Also, BL2 builds are now faster since
tfm-mcuboot is fetched via west.
This reverts commit 88a865c28d.
Signed-off-by: Øyvind Rønningstad <oyvind.ronningstad@nordicsemi.no>
The scheduler documentation was updated before to define a reschedule
point, but the related term sleep was not clearly described. Add a
definition, and link to it from the API terminology.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Mark the option BT_FIXED_PASSKEY as an insecure option.
During Legacy pairing the passkey is used as encryption key, and
brute-forcing this is easy.
During LE Secure Connections the passkey is checked one bit at a time,
so when it is fixed the passkey can be deduced with series of pairing
attempts.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
The Bluetooth SMP protocol depends on the BT_ECC for the LE Secure
Connections pairing. Since this feature cannot be disabled we must
select this dependency.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Fix fixed passkey for legacy pairing. The pairing procedure fails
because the temporary key (TK) has not been set for this pairing
procedure.
Fix this by always requesting the legacy TK.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
There are more and more tests that fail due to stackoverflow.
Increasing MAIN_STACK_SIZE to fix those issues.
Signed-off-by: Alexandre Bourdiol <alexandre.bourdiol@st.com>
This adds a new GEN_ABSOLUTE_SYM_KCONFIG() specifically for
generating absolute symbols in assembly for kconfig values.
This is needed as the existing GEN_ABSOLUTE_SYM() with
constraints in extended assembly parses the "value" as
signed 32-bit integers. An unsigned 32-bit integer with
MSB set results in a negative number in the final binary.
This also prevents integers larger than 32-bit. So this
new macro simply puts the value inline within the assembly
instrcution instead of having it as parameter.
Fixes#31562
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Fix misleading message in the bluetooth shell when the id-create
command returns en error code. In this case both failure and success
messages are printed.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Add options to control the behavior of the init command.
Option no-settings-load allows the user to create identities
with the id-create command after bt_enable() but before settings_load().
Option no-ready-cb allows the user to test using synchronous enable,
since this sometimes leads to different behavior of the stack.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Prevent the bt_rand function from being called before bt_enable.
Depending on the implementation of bt_rand this function cannot
be called before bluetooth has been initialized. With host supplied
crypto functions the HCI LE rand command is used for example.
The use case for calling bt_id_create before bt_enable is meant for
when the application has storage for the identity instead of the stack.
So we add the requirement that the application has to have storage
for the identity resolving key (IRK) in addition when the local
device is privacy-enabled.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Fix doxygen in addr.h file, missing /** needed for doxygen,
and placing the brief on the first line of the comment consistenly.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Deprecate the bt_set_id_addr API function. This is merely a wrapper
for the bt_id_create function now, except an IRK cannot be given.
When CONFIG_BT_PRIVACY is enabled an IRK has to be given by the
application because the bt_rand function cannot be called before
bt_enable.
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Fix stack overflow in the TX power control sample.
Current stack usage with 0.11.1 zephyr toolchain was at 100% stack
usage of the DYN TX stack.
Fixes: #31433
Signed-off-by: Joakim Andersson <joakim.andersson@nordicsemi.no>
Similarly to what was done for watchdog test,
update watchdog sample to be compatible with stm32 window watchdog
and enable testing by twister.
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Rather than Kconfig vendor symbols, select stm32 watchdog using
compatible.
So user only has to enable the requested node and set
CONFIG_WATCHDOG=y.
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Review tests applicability so that it only relies on watchdog
compatible availability. Additionally, tweak APB prescaler
in case of window watchdog test, so that timeout value can be
achieved.
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Set maximum window value as a test variable.
This allows to set a much shorter value that is compatible with
use in stm32 window watchdog.
Additionally, tweak test configuration so that it is applied on
compatible base (when available).
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Fix the timeout computation to provide more accurate timeouts
versus requested timeout.
Additionally, the error margin is reworked in order to:
- be relative to the application requested timeout (10% tolerance)
- exclude timeouts inferior to application request
Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
Fix the default ULL_HIGH execution priority to be lower than
LLL execution priority to reduce Radio ISR service latency.
Also, default to Zero Latency IRQ for Radio ISR if enabled
in the kernel.
Fixes#29994.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
The commit adds notes on addition on fs_file_t_init function and its
impact on File system API.
Signed-off-by: Dominik Ermel <dominik.ermel@nordicsemi.no>
