We haven't updated the example revisions in the alternate topologies
sections in a while. Update these for the upcoming 2.5 release.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
With the documentation for west 0.9, the 'basics' page is completely
out of control. Reorganize it so it fits in a few pages, moving many
details out into other pages for the truly curious.
As part of this, create new pages for built-in commands (built-in.rst)
and esoteric details about workspaces (workspaces.rst).
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
Rename the 'basics' page to basics.rst, adding a redirect for the old
title.
Rearrange some other index pages to improve the flow:
- Put relatively stale content in moving-to-west.rst to the bottom. I
expect pepole have either moved west or are choosing not to.
- Promote the release notes page higher up, since that seems more
important.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
The commit adds notes on addition on fs_dir_t_init function and its
impact on File system API.
Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
The two calls to unpend a thread from a wait queue were inexplicably*
unsynchronized, as James Harris discovered. Rework them to call the
lowest level primities so we can wrap the process inside the scheduler
lock.
Fixes#32136
* I took a brief look. What seems to have happened here is that these
were originally synchronized via an implicit from an outer caller
(remember the original Uniprocessor irq_lock() API is a recursive
lock), and they were mostly implemented in terms of middle-level
calls that were themselves locked. So those got ported over to the
newer spinlock but the outer wrapper layer got forgotten.
Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
Fix missing advertising PDU length initialization which
prevented advertising PDU to be transmitted by Radio. This
issue was reproduced when connectable undirected advertising
was enabled without setting an advertising data.
Signed-off-by: Vinayak Kariappa Chettimada <vich@nordicsemi.no>
Drivers are using DEVICE_AND_API_INIT which is deprecated so convert
the iproc_pax dma drivers over to using DEVICE_DT_INST_DEFINE and
DEVICE_DT_INST_GET.
Fixes#32153
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
The default frequency for this board in Renode is 66 MHz. This needs to
be even with `SYS_CLOCK_HW_CYCLES_PER_SEC` to avoid such problems as in
the #31726 issue.
Unfortunately, the difference in these values was helpful for some tests
that are failing with 66 MHz set in both places. It created artificial
boost in certain circumstances.
The frequencies in the default Renode platform description (`.repl`)
file for MI-V were overridden with 4 MHz value that seems to be better
tolerated based on testing. The `SYS_CLOCK_HW_CYCLES_PER_SEC` was
adjusted as well.
To solve the rest of the issues, `cpu PerformanceInMips` was set to 4.
It seems tests are completed faster with such a value.
This commit fixes#31726.
Signed-off-by: Adam Jeliński <ajelinski@antmicro.com>
Inline functions reference a __fallthrough macro that is toolchain
specific. Include the header that provides that macro.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Although the master API is stable, the slave API has never been
documented and has only two in-tree implementations with no uses
outside of a single test application. It is marked experimental in
this commit.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
The I2C slave API has been present since 1.12 but lacked documentation
on the behavior of its functions. Provide that, and revise existing
documentation to make clear the type and mode of the device passed to
each API call.
Documented behavior is derived primarily from the LPC11U6X
implementation, as the STM32 implementation generally ignores return
values.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
The callback pointers for uninitiated operations are implicitly null;
making them explicit only confuses maintainers searching for drivers
that implement the API.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Zephyr does not currently allow deviation from standard terminology
for a technology even if it is non-inclusive, until the corresponding
standards body has confirmed intent to change that terminology. The
terms used in a previous attempt to be inclusive do not match the
expected forthcoming standard terms.
Revert to standard terms until the new ones have been announced and
the switch made throughout Zephyr.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Status of addressing inclusive language concerns is embedded into the
coding guidelines. Add a link to the corresponding issue for I2C.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Documentation for subsystems that use offensive terms by standard may
need to reference this section as justification for their continued
and new use.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Nordic watchdog does support a callback, but it has a time limit that
is too tight for the way this sample uses it. Add infrastructure to
bypass using the callback in this sample.
Also add more diagnostics explaining the status of callback use in the
sample.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
Check if a given 'compat' is a key in edt.compat2okay before accessing
its entry in dt_compat_on_bus() and dt_nodelabel_has_compat(), as for
a non-existing key a new entry with an empty list would be created,
and dt_compat_enabled() would then incorrectly return "y" when called
for the same 'compat'.
Signed-off-by: Andrzej Głąbek <andrzej.glabek@nordicsemi.no>
Handle a semantic versioning case used by Kitware that is not
supported by the packaging module we're using for version comparisons.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
Fix twister output. "Test suite" used previously did not correspond
to what was counted (test scenarios). A test scenario is an entry from
testcase/sample.yaml under "tests:", which specify configuration used to
build a test. Changes in docs will follow later on. Also replace the
amount of preselectd platforms with the amount of platforms that
were not filtered out and actually tested.
Signed-off-by: Maciej Perkowski <Maciej.Perkowski@nordicsemi.no>
Power management documentation was still using "deep sleep" and
"sleep" concepts. This commit just remove them, since they are not
longer used, and introduce the new power states defined in pm_state.
As each state has its own meaning and it is documented in the API,
lets just reference them here and avoid getting out of sync.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
This adds a few lines about changes to memory management,
introduction of demand paging, and changes to x86 memory
management code.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This adds a few lines for newly added board and SoC defintions
for Elkhart Lake (ehl_crb) and ACRN configurations (acrn_ehl_crb).
Signed-off-by: Jennifer Williams <jennifer.m.williams@intel.com>
The _update_jump_to_matches() function used when searching from the
jump dialog is using the re module without importing it. Fix that.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
Increase the size of ram/flash given in nrf5340dk_nrf5340_cpuapp.yaml
so they match the actual numbers which are available for default
tests and samples.
Signed-off-by: Maciej Perkowski <Maciej.Perkowski@nordicsemi.no>
Commit 40016a6a92 ("libc/minimal: Use a sys_heap for the malloc
implementation") replaced sys_mem_pool_alloc() with sys_heap_alloc().
The problem is that those aren't equivalent. While the former did
guard against concurrent usage, the later doesn't.
Add the same locking around sys_heap_alloc() that used to be implicit
with sys_mem_pool_alloc().
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
llvm 11 doesn't like having preprocessor directives in the expansion
of a macro used as a preprocessor conditional expression.
Signed-off-by: Peter Bigot <peter.bigot@nordicsemi.no>
After #31176, the network would get stored immediately in
bt_mesh_net_create, causing the address and devicekey to get stored as
their zero-initialized version, as they're only being set in the
bt_mesh_comp_provision call, which fires after.
Fixes#32067.
Signed-off-by: Trond Einar Snekvik <Trond.Einar.Snekvik@nordicsemi.no>
