ERROR_QUIET should be used only when the command is actually expected to
fail, otherwise discarding stderr can be extremely time-consuming. See
recent example of time lost in commit 40c2e08e82 ("xcc/cmake: don't
discard stderr; don't (ever!) use ERROR_QUIET")
If some Python interpreter cannot even evaluate `sys.version_info[:2]`,
then we definitely don't want to hide what its stderr says about that.
Fixes commit 94de3e9f60 ("cmake: FindPython3: Adjust python3 program
search")
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Fixes a mismatch between the cmake and ld linker scripts whereby
the cmake script was missing using the flash load size Kconfig
option (if it was said), which would result in images far larger
than they should have allowed to be
Signed-off-by: Jamie McCrae <jamie.mccrae@nordicsemi.no>
Removes support for the deprecated features of having a prj file
which has the board name on the end of it. Board Kconfig fragments
should be used instead.
Signed-off-by: Jamie McCrae <jamie.mccrae@nordicsemi.no>
Commit 40c2e08e82 ("xcc/cmake: don't discard stderr; don't (ever!) use
ERROR_QUIET") fixed xcc but the corresponding code is duplicated. Fix
the duplicates too.
See that commit for the issue description, longer commit message and
rationale.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Remove ERROR_QUIET which is a bad idea 99.9% of the time. When any
program makes the effort of using stderr, we _really_ don't want to lose
those error messages.
Using ERROR_QUIET in XCC is even worse because of how high maintenance
XCC is; see another example in 4cba9e6d42 ("cmake: warn the user that
the toolchain cache hides errors")
No one expects error messages to be silently discarded and especially
not people not familiar with CMake (= most people); so hiding the
following error stalled progress for a couple days:
```
Error: there is no Xtensa core registered as the default.
You need to either specify the name of a registered Xtensa core (with
the --xtensa-core option or the XTENSA_CORE environment variable) or
specify a different registry of Xtensa cores (with the --xtensa-system
option or the XTENSA_SYSTEM environment variable).
Executing the below command failed. Are permissions set correctly?
```
Also capture stdout and print it on failure because you never know.
Indent the ` ${CMAKE_C_COMPILER} --version` line in the error message so
CMake does not split that line.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
Add support for mutable devices. Mutable devices are those which
can be modified after declaration, in-place, in kernel mode.
In order for a device to be mutable, the following must be true
* `CONFIG_DEVICE_MUTABLE` must be y-selected
* the Devicetree bindings for the device must include
`mutable.yaml`
* the Devicetree node must include the `zephyr,mutable` property
Signed-off-by: Christopher Friedt <cfriedt@meta.com>
Add MCUBOOT_IMGTOOL_OVERWRITE_ONLY Kconfig option which
passes the --overwrite-only option to imgtool to avoid
adding the swap status area size when calculating overflow.
It is used by non-swap update modes.
Signed-off-by: Andrej Butok <andrey.butok@nxp.com>
Fixes an issue whereby a board revision is 0 and the overlay file
exists but would not be included
Signed-off-by: Jamie McCrae <jamie.mccrae@nordicsemi.no>
Change CROSS_COMPILE_TARGET to be defined based on CONFIG_SOC_SERIES.
This fix adjusts CROSS_COMPILE_TARGET to ensure compatibility
with the 'espressif' toolchain variant, following the changes in
commit 6b57b3b786.
Signed-off-by: Pisit Sawangvonganan <pisit@ndrsolution.com>
Add a CMake zephyr library extension for add_dependencies, similar
to the target_<func> functions.
This avoids using ${ZEPHYR_CURRENT_LIBRARY} directly.
Signed-off-by: Pieter De Gendt <pieter.degendt@basalte.be>
Replace the TLS base address pointer from TPIDRURO to TPIDRURW.
The difference between them is that TPIDRURO is read-only in user mode
but TPIDRURW isn't. So TPIDRURO is much more suitable for store
the address of _kernel.CPU[n]. For this reason, this commit replaces
the base pointer of the TLS area.
Signed-off-by: Huifeng Zhang <Huifeng.Zhang@arm.com>
Issue reported on Discord.
Safeguard cmake_path() for SOC_LINKER_SCRIPT so that the path is only
processed when SOC_LINKER_SCRIPT is defined.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
This commit updates the handling of board and SoC linker scripts.
Several SoCs creates a linker.ld file which sole purpose is to include
another arch common linker script, often with content like this:
#include <arch>/linker.ld
instead of 100+ SoC specific linker.ld files containing just a single
include line of above structure, then this commit introduces two now
CMake variables, BOARD_LINKER_SCRIPT and SOC_LINKER_SCRIPT.
This allows the board and SoC CMake code to point directly to a common
linker script instead of creating a dummy linker.ld file doing this.
This removes the need for several dummy linker.ld file.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
This commit adds the public header for the GNSS API, along
with the initial GNSS Kconfig file and an entry in the
common linker file for registered GNSS data callbacks.
A very naive implementation of the GNSS data callback is
provided as well in drivers/gnss/gnss_publish.c
Signed-off-by: Bjarki Arge Andreasen <bjarkix123@gmail.com>
Generate ZEPHYR_{MODULE_NAME}_MODULE for existent modules
for unittests as well since they may be using Zephyr modules.
Fixes: 64348
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
This allows users to set `EXTRA_*FLAGS` variables when using sysbuild.
If these flags are provided in both local and global sysbuild scopes,
then only the local ones will be used for a given image. This is to
circumvent issues with mixing space-separated and `;`-separated lists.
Signed-off-by: Grzegorz Swiderski <grzegorz.swiderski@nordicsemi.no>
File gcc-m-fpu.cmake is responsible for determining what should be
passed to -mfpu option. Fortunately GCC and Clang options are compatible
so we can use the file for clang.
The list of supported -mfpu options can be found at
https://github.com/llvm/llvm-project in
llvm/include/llvm/TargetParser/ARMTargetParser.def file.
Signed-off-by: Patryk Duda <pdk@semihalf.com>
HAS_DTS has become a redundant option. All Zephyr architectures now
select this option, meaning devicetree has become a de-facto
requirement. In fact, if any board does not provide a devicetree
source, the build system uses an empty stub, meaning the devicetree
machinery always runs.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
By providing a devicetree stub file, we make sure some internal macros
required by devicetree.h are generated in devicetree_generated.h. This
makes sure that systems without devicetree can continue working without
extra ifdeffery.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
Add QEMU_EXTRA_FLAGS as QEMU board config option.
This allows Twister tests to provide additional device setup
commands to QEMU in prj.conf or testcase.yaml configuration files.
Example use case: to setup TCP or UDP network interfaces
with non-conflicting port numbers in different test suites
to avoid conflicts when Twister run tests in parallel on the
same host.
Signed-off-by: Dmitrii Golovanov <dmitrii.golovanov@intel.com>
LLVM toolchain provides its own C++ standard library called libc++.
This patch adds new LLVM_LIBCXX config which should be used to indicate
that libc++ is used.
Information about library can be found at https://libcxx.llvm.org
Signed-off-by: Patryk Duda <pdk@semihalf.com>
Functions `zephyr_linker_dts_memory()` and `zephyr_linker_dts_section()`
are described as defining a memory region or section based on a DT node,
as long as it "exists and has status okay". However, only the existence
of the node is actually checked, using `dt_node_exists()`. To fix that,
employ `dt_node_has_status()` instead, which can check both conditions.
The status check is important, because both functions require the given
DT node to contain a `zephyr,memory-region` property (not to be confused
with the compatible string). This property is required by the associated
binding as well, but required properties can be omitted from nodes which
don't have status "okay". In those cases, edtlib won't raise an error,
and neither should CMake, because those nodes should be ignored.
Speaking of that property, add a missing error check for it as a bonus
(tucked behind the status check, of course).
Signed-off-by: Grzegorz Swiderski <grzegorz.swiderski@nordicsemi.no>
Calling this function with an output variable named `var` or `okay`
could produce an incorrect result, due to the variable being mishandled.
Add simple changes to avoid this.
Signed-off-by: Grzegorz Swiderski <grzegorz.swiderski@nordicsemi.no>
Fixes: #63771
The 'compiler/*/generic.cmake' is intended to set a C preprocessor which
can be used for devicetree preprocessing.
No C++ compiler is needed as host tool and should therefore not be set
in this file.
Remove the code to avoid setting a not required C++ compiler.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Add the exe fle to the runnerconfiguration class,
so we can use it from runners which will need it.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Picolibc dependencies limit ability to use third party minimal
implementations of CPP when enablng PICOLIBC_USE_MODULE.
Signed-off-by: Al Semjonovs <asemjonovs@google.com>
Fixes: #63011
Custom targets are not available in script mode, which cause snippet
loading to fail when using package helper.
Lookup of snippets is supported by package helper, thus allowing twister
to filter test cases based on snippets.
However, loading of snippets itself is not supported as snippets make
use of Zephyr scoping, which uses custom targets, something that is not
available in script mode.
Therefore overload the `zephyr_set()` function, so that CMake package
helper can be used together with snippets.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
The run_native (and therefore run) targets did not depend
on the proper thing for native_simulator based targets
Fix it.
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
Fixes: #35671
Add minimal version required for LLVM LLD linker.
Linking fails with older LLVM LLD, such as v10.0.0.
LLVM v14.0.0 was released in 2022, and latest LLVM is v17.0.1.
Zephyr currently doesn't have a strict minimum version of LLVM
specified, but based on LLVM development and known issues on older
releases, then a minimum version of v14.0.0 has been chosen in this
commit.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
When `toolchain_is_ok` fails, the error message points the user at the
CMake logs. But those logs will be empty if the user tried to compile
more than once (typically: cleans everything and tries again). So tell
the user that cleaning the toolchain cache is required to see the error.
Tell the user to "move" the cache instead of removing it in case
technical support needs the cache for forensics.
Some finicky toolchains can be "non-deterministic" and fail
_sometimes_. For instance a license server can be flaky, or the
toolchain can require an "elaborate" set of environment variables
triggering some configuration "trial-and-error". In such a
non-deterministic case deleting the cache is enough to get rid of the
issue and move on! Looking at logs is not even required; even
better. Once the toolchain cache believes that the toolchain works, any
future toolchain glitch will be obvious at actual compilation time.
To test all this:
```
# Verify that the toolchain can compile a dummy file, if it is not we
# won't be able to test for compatibility with certain C flags.
-zephyr_check_compiler_flag(C "" toolchain_is_ok)
+zephyr_check_compiler_flag(C "-fubar" toolchain_is_ok)
assert(toolchain_is_ok "The toolchain is unable to build a dummy C file.\
```
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
This mirrors TOOLCHAIN_HAS_NEWLIB to picolibc so the cmake
variable TOOLCHAIN_HAS_PICOLIBC can be used in kconfig.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
The armclang version detection introduced in #55133 does not correctly
detect a valid ARM Compiler (armclang) installation in all situations.
When ARM Compiler for Embedded is installed as part of ARM-DS or Keil,
then it may report itself in following output:
> Product: Arm Development Studio ... <year>.<no>
> Component: ARM Compiler x.y(.z)
> Tool: armclang [...]
>
> Target: ...
Correct the version extraction by turning each line into a list of
strings which can then be looped to find the ARM compiler component to
ensure the correct line is used for retrieving the version information.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Fixes: #62589
Follow-up: #60031
The PR #60031 moved CMake code to new folder location causing generated
library names to change.
This change impacted the use of those libraries in the Zephyr armlink
CMake code, causing CMake failures at configure time.
This PR fixes this failure by updating the armlink CMake code to use
the new library names.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Some tests require the zephyr toolchain version 0.16 or newer to
be able to use picolibc functionality.
Signed-off-by: Jamie McCrae <jamie.mccrae@nordicsemi.no>
This function performs topological sorting of CMake targets. Its initial
use case in Zephyr will be for implementing sysbuild image dependencies,
i.e., specifying an image order for CMake configuration and flashing.
Sourced from a comment on PR #57884 (anchor: #discussion_r1206807021)
Authored-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Signed-off-by: Grzegorz Swiderski <grzegorz.swiderski@nordicsemi.no>
Forces setting the cached APPLICATION_CONFIG_DIR variable so that
it gets updated with the absolute path instead of reverting back
to the relative path which then causes problems when sourcing
other files, e.g. app.overlay
Signed-off-by: Jamie McCrae <jamie.mccrae@nordicsemi.no>
This changes to cache variable GNULD_VERSION_STRING across cmake
runs. This variable is used to determine whether -Wl,-no-pie is
being passed to linker. However, if cmake is run multiple times
without clearing the build directory, GNULD_VERSION_STRING was
lost and the script falsely assumed the linker could not take
this argument, and thus omitting it during linking. Depending on
the host, it would warn on something like this:
/usr/bin/ld.bfd: app/libapp.a(main.c.obj):
warning: relocation in read-only section
`.text._posix_zephyr_main'
/usr/bin/ld.bfd: warning: creating DT_TEXTREL in a PIE
To fix this, simply caches GNULD_VERSION_STRING so it can be
used during subsequent cmake runs.
Fixes#61725
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This reverts commit 577d47f3f1.
The --config= syntax with the equal sign does not work with clang 15
which is less than one year old (https://releases.llvm.org/). It does
not work with clang 14 either.
```
clang --verbose
clang version 15.0.7
clang --help | grep config
-cl-std=<value> OpenCL language standard to compile for.
--config <value> Specifies configuration file
--cuda-feature=<value> Manually specify the CUDA feature to use
clang: error: unsupported option '--config=/home/runner/work/...
```
The reverted commit mentioned "issues in some situations" without
providing any example or other information.
This revert fixes the SOF fuzzer build, see more details in #61778.
Signed-off-by: Marc Herbert <marc.herbert@intel.com>
ZBus stores observers in two ways: statically using a list and dynamically
using a memory slab. Both present limitations. Static observers work only
for channel definition. The dynamic observers rely on a memory slab that
forces the user to manage its size to avoid issues with adding
observers. This commit fixes the static allocation problem by using the
iterable sections for allocating observation data and replacing the VDED
execution sequence since now it is possible to prioritize static observer
execution. All the runtime observers are dynamically allocated on the heap
instead of a specific memory pool.
BREAK changes (only internal, not APIs):
* ZBus channel metadata changed. Remove the observers' static array
pointer. Rename the `runtime_observers` pointer to `observers`. Add
`observer_start_idx` and `observer_end_idx`;
* Change the VDED execution sequence. The position (on definition time),
the priority in conjunction with the lexical order, is considered for
static post-definition time observers. At last, the runtime observer
follows the adding sequence;
* Replace the `CONFIG_ZBUS_RUNTIME_OBSERVERS_POOL_SIZE` with
`CONFIG_ZBUS_RUNTIME_OBSERVERS`.
New APIs:
* New iterable section iterators (for channels and observers) can now
receive a user_data pointer to keep context between the function calls;
* New `ZBUS_LISTENER_DEFINE_WITH_ENABLE(_name, _cb, _enable)` and
`ZBUS_SUBSCRIBER_DEFINE_WITH_ENABLE(_name, _queue_size, enable)` that
enable developers define disabled observers. They need to be enabled
during runtime to receive notifications from the bus;
* `ZBUS_CHAN_ADD_OBS` macro for adding post-definition static observers of
a channel.
Important changes:
* Move the ZBus LD file content to the `common-ram.ld` LD file. That was
necessary to make ZBus compatible with some Xtensa and RISCV boards.
Signed-off-by: Rodrigo Peixoto <rodrigopex@gmail.com>
Add a a new source coverage for native builds
and new kconfig choice of COVERAGE mode to select which:
* COVERAGE_NATIVE_GCOV: what we had until now with native builds
* COVERAGE_NATIVE_SOURCE: a new LLVM source coverage mode
* COVERAGE_GCOV: the old COVERAGE_GCOV (embedded gcov data generation).
Signed-off-by: Alberto Escolar Piedras <alberto.escolar.piedras@nordicsemi.no>
