This test uses pytest to generate an EDK from a simple Zephyr
application, and uses this EDK to build a simple extension, to ensure
that EDK generation is sane.
Signed-off-by: Ederson de Souza <ederson.desouza@intel.com>
Instead of hardcoding `autoconf.h` imacro, get the list of imacros from
the llext flags. As those come in the form of absolute paths, they also
need to be massaged to point from the EDK directory without revealing
host complete paths.
Also, the EDK now keeps the imacros on a different flag,
`LLEXT_GENERATED_IMACROS_CFLAGS`, to keep it similar to other generated
includes.
Signed-off-by: Ederson de Souza <ederson.desouza@intel.com>
Some flags are common between in tree extensions and out of tree
supported by the EDK. Instead of duplicating those flags, the EDK reuses
the llext ones.
However, as the EDK has its own needs, two new lists,
`LLEXT_EDK_APPEND_FLAGS` and `LLEXT_EDK_REMOVE_FLAGS` are defined to
allow EDK to append or remove flags as needed.
Signed-off-by: Ederson de Souza <ederson.desouza@intel.com>
Besides the LLEXT_CFLAGS, which have all that is needed to compile,
generate more granular ones, LLEXT_INCLUDE_CFLAGS,
LLEXT_ALL_INCLUDE_CFLAGS, LLEXT_GENERATED_INCLUDE_CFLAGS and
LLEXT_BASE_CFLAGS. These are done for convenience, as they can help on
different setups, such as unit testing.
Signed-off-by: Ederson de Souza <ederson.desouza@intel.com>
A new Kconfig option which generates syscall stubs assuming that
extensions will always run on userspace, thus simplifying linking
them, as there's no need for z_impl_ stubs (used for direct syscalls),
CONFIG_LLEXT_EDK_USERSPACE_ONLY.
While defining __ZEPHYR_USER__ could have the same effect for optmised
builds, people building extensions on debug environments - thus
non-optimised - would suffer, as they'd need to somehow make the stubs
available (by either exporting the symbol or implementing dummy stubs).
Signed-off-by: Ederson de Souza <ederson.desouza@intel.com>
Loadable extensions need access to Zephyr (and Zephyr application)
includes and some CFLAGS to be properly built. This patch adds a new
target, `llext-edk`, which generates a tar file with those includes and
flags that can be loaded from cmake and make files.
A Zephyr application willing to expose some API to extensions it loads
only need to add the include directories describing such APIs to the
Zephyr ones via zephyr_include_directories() CMake call.
A new Kconfig option, CONFIG_LLEXT_EDK_NAME allows one to control some
aspects of the generated file, which enables some customization - think
of an application called ACME, willing to have a ACME_EXTENSION_KIT or
something.
All EDK Kconfig options are behind CONFIG_LLEXT_EDK, which doesn't
depend on LLEXT directly - so that EDK features can be leveraged by
downstream variations of loadable extensions.
Also, each arch may need different compiler flags for extensions: those
are handled by the `LLEXT_CFLAGS` cmake flag. An example is set for GCC
ARM.
Finally, EDK throughout this patch means Extension Development Kit,
which is a bad name, but at least doesn't conflict with SDK.
Signed-off-by: Ederson de Souza <ederson.desouza@intel.com>
TLS for Arm targets seems to be well supported in clang/lld, so mark it as
supported by the toolchain.
Signed-off-by: Jonathon Penix <jpenix@quicinc.com>
This was previously used as a variable internal to the module, but it
can be used in other places in the build system, which may want a list
of DT-specific includes. Therefore, document it as an output variable.
Signed-off-by: Grzegorz Swiderski <grzegorz.swiderski@nordicsemi.no>
Changing from `;` to `|` directly doesn't consider for escaped `;`
inside a list. Fix suggested by @pillo79.
Signed-off-by: Ederson de Souza <ederson.desouza@intel.com>
The `cpu.h` is meant to be a hack around Zephyr's dependencies
for unit-test, relocate it to a folder that's included only by
unit-test.
Signed-off-by: Yong Cong Sin <ycsin@meta.com>
Change the GCC toolchain configuration to make use of the Cortex-R82
target. When Cortex-R82 was added as a GCC toolchain option, the GCC
version of the Zephyr SDK did not support Cortex-R82 tuning. Zephyr was
therefore compiled compiled for the Armv8.4-A architecture. Since Zephyr
SDK 0.15.0 (which updated GCC from 10.3.0 to 12.1.0) coupled with Zephyr
3.2, the Cortex-R82 target is supported.
The Armv8-R AArch64 architecture does not support the EL3 exception level.
EL3 support is therefore made conditional on Armv8-R vs Armv8-A.
Signed-off-by: Debbie Martin <Debbie.Martin@arm.com>
On very early stage build system needs to get access to DTC preprocessor.
MWDT has no it's own preprocessor, so here zephyr-SDK preprocessor is used.
On latest build stages zephyr-SDK objcoby also required as MWDT binutils
don't support all features. This at the same time requires that
ZEPHYR_SDK_INSTALL_DIR must be initialized with valid arch-dependent
prefix.
Zephyr-SDK requires ARCH variable to be initialized before
include "generic.cmake", but in hew HWMv2 model ARCH variable will be
initialized more later. This workaround uses any (first awailable,
independent on ARCH) toolchain from SDK for DTC preprocessing only.
For other build stages ARCMWDT will be used.
Signed-off-by: Nikolay Agishev <agishev@synopsys.com>
Currently, clang produces a warning that the '--specs=nosys.specs' argument
is unused. Remove this flag to fix the warning as clang will generally not
honor '--specs' when targeting arm baremetal.
While this flag has been present for some time, I think this should be safe
to do. For background, clang does not seem to handle '--specs' [1] besides
possibly passing it through to GCC if GCC is used as the driver for
linking. However, whether GCC will be used for linking as a fallback
depends on the "Toolchain" [1] clang uses internally, which in turn depends
on the triple. For arm/thumb baremetal triples, the Toolchain clang uses
will not fall back to GCC to drive linking, so '--specs' will never be used
here.
I believe this behavior in clang is fairly longstanding as well (since
~2017/LLVM 5 [2]). While there isn't a minimum required clang version for
Zephyr, Zephyr currently requires lld >= 14.0.0. So, I don't think removing
this flag should impact current users (besides preventing the warning).
[1] https://clang.llvm.org/docs/DriverInternals.html
[2] https://reviews.llvm.org/D33259
Signed-off-by: Jonathon Penix <jpenix@quicinc.com>
This adds supports for flashing images with sysbuild where there
are multiple images per board to prevent using the same command per
image flash which might cause issues if they are not ran just once
per flash per unique board name. A deferred reset feature is also
introduced that prevents a board (or multiple) from being reset if
multiple images are to be flashed until the final one has been
flashed which prevents issues with e.g. security bits being enabled
that then prevent flashing further images.
These options can be set at a board level (in board.yml) or a SoC
level (in soc.yml), if both are present then the board configuration
will be used instead of the SoC, and regex can be used for matching
of partial names which allows for matching specific SoCs or CPU cores
regardless of the board being used
Signed-off-by: Jamie McCrae <jamie.mccrae@nordicsemi.no>
This commit reverts a breaking change in CMAKE_SYSTEM_NAME introduced by
"Zephyr" back to "Generic") and removes the file
`cmake/modules/Platform/Zephyr`.
Both changes in the aforementioned PR were only introduced to ultimately
modify the value of the global CMake property TARGET_SUPPORTS_SHARED_LIBS
for the special case of building for Xtensa with LLEXT.
The modification of CMAKE_SYSTEM_NAME is considered a breaking change
because it has the potential to alter the build of any non-trivial project
that previously checked for the "Generic" system identifier as
corresponding to Zephyr - for example by doing
`if (CMAKE_SYSTEM_NAME STREQUAL "Generic")`. Such builds may now break in
many ways including silently when there is no `else()` clause with a
`message()` to alert the user that a whole configuration block had been
skipped.
In essence, that CMAKE_SYSTEM_NAME modification was only introduced in
order to have CMake to load `cmake/modules/Platform/Zephyr.cmake` which in
turn adjusted the value of TARGET_SUPPORTS_SHARED_LIBS.
But the use of a CMake platform file like this is ineffective for
non-trivial projects where one or more top level CMake `project()` calls
may happen before the first call to `find_package(Zephyr)` because in such
cases CMAKE_MODULE_PATH will not have been modified yet to contain the
path to <Zephyr_ROOT>/cmake/modules and thus no platform file will be
include by CMake.
This patch moves the conditional override of TARGET_SUPPORTS_SHARED_LIBS
needed by some archs (e.g. Xtensa) into the `kernel.cmake` module which
is known to be the first call to `project()` that enables any language and
thus the one that must come before any artifact target can be defined.
Note commit 64e7d85 added a Kconfig to specify the object type of llext
being built, so it's not tied to the arch anymore but to the
CONFIG_LLEXT_TYPE_ELF_SHAREDLIB option.
Signed-off-by: Nicolas Lebedenco <nicolas@lebedenco.net>
Adds support for SoC overlay files which go in a ``socs`` folder
in application folders and functions similar to the ``boards``
folder, but works for SoCs instead of boards
Signed-off-by: Jamie McCrae <jamie.mccrae@nordicsemi.no>
Adds support for a new ``socs`` folder that can be placed in
application folders and functions similar to the ``boards``
folder, but works for SoCs instead of boards
Signed-off-by: Jamie McCrae <jamie.mccrae@nordicsemi.no>
Some third-party modules uses 'find_package(Threads REQUIRED)' to check
if threads implementation is supported.
The original implementation tries to find threads library using various
methods (e.g. checking if pthread library is present or compiling
example program to check if the implementation is provided by libc), but
it's not able to detect pthread implementation provided by Zephyr.
Signed-off-by: Patryk Duda <patrykd@google.com>
Unlike GNU ld, lld's gp relaxation is disabled by default and must be
explicitly enabled via `--relax-gp`. Pass this flag to enable gp relaxation
for lld when both linker relaxations and gp usage for RISC-V are enabled.
Signed-off-by: Jonathon Penix <jpenix@quicinc.com>
Add a symbol to enable GNU C Extensions. And a hidden option for
toolchains to signal GNU Extensions support.
Signed-off-by: Pieter De Gendt <pieter.degendt@basalte.be>
USB High-Speed devices must be able to operate at both High-Speed and
Full-Speed. The USB specification allows the device to have different
configurations depending on connection speed. Modify the API to reflect
USB Specification requirements on what can (e.g. configurations) and
what cannot (e.g. VID, PID) be speed dependent.
While the class configurations for different speeds are completely
independent, the actual class instances are shared between operating
speeds (because only one speed can be active at a time). Classes are
free to provide different number of interfaces and/or endpoints for
different speeds. The endpoints are assigned for all operating speeds
during initialization.
Signed-off-by: Tomasz Moń <tomasz.mon@nordicsemi.no>
Introduce usbd_class_iter for keeping endpoint assignment variables
and the single-linked list node. No functional changes right now, but
this paves the way for independent speed specific configurations.
Signed-off-by: Tomasz Moń <tomasz.mon@nordicsemi.no>
set_compiler_property does not accept a TARGET argument. Only set_property
does but they are easy to confuse. This patch fixes the wrong instances of
set_compiler_property that should have been set_property.
Signed-off-by: Nicolas Lebedenco <nicolas@lebedenco.net>
That call to enable_languages() is passed C CXX ASM which misleads the
reader into thinking that the three languages are enabled in that call but
at that point only ASM is actually because C and CXX were already
implicitly enabled by the call to project() above. The suggested change
removes the misleading C CXX arguments but we could as well make it clear
in the project() call by passing LANGUAGES C CXX.
Signed-off-by: Nicolas Lebedenco <nicolas@lebedenco.net>
When LLEXT_REMOVE_FLAGS is empty, the regular expression that is
constructed to remove flags from the Zephyr flags would match any
string, resulting in all flags being removed. This is not the intended
behavior, so we need to handle this case by setting the regular
expression to a pattern that does not match anything.
Signed-off-by: Luca Burelli <l.burelli@arduino.cc>
The partial linking option depends on the linker used in the build
process, and this flag is already defined by the Zephyr toolchains.
Use the $<TARGET_PROPERTY:linker,partial_linking> generator expression
to retrieve it instead of hardcoding the value for the GNU linker.
Signed-off-by: Luca Burelli <l.burelli@arduino.cc>
Some toolchains may generate convoluted paths when reporting accessory
tools. This is the case with GCC in the Zephyr SDK toolchain. For
example, for a Zephyr SDK installed under `C:\Portable\Zephyr` a call to
`gcc --print-prog-name=ld.bfd` should normally return something like
`c:/portable/zephyr/zephyr-sdk-0.16.5/arm-zephyr-eabi/arm-zephyr-eabi/bin/ld.bfd.exe`
but because of how the toolchain was created the path reported gets all
messed up with relative fragments. In above case, the actual path reported
was `c:/portable/zephyr/zephyr-sdk-0.16.5/arm-zephyr-eabi/bin/../lib/gcc/arm-zephyr-eabi/12.2.0/../../../../arm-zephyr-eabi/bin/ld.bfd.exe`
One might argue that this should be fixed in the toolchain which could be
possible for the Zephyr SDK but not for other toolchains (definitely not
for proprietary ones).
Signed-off-by: Nicolas Lebedenco <nicolas@lebedenco.net>
Adds a new variable NORMALIZED_BOARD_QUALIFIERS which contains
the board qualifiers in file-name format, this allows for
constructing strings in applications (e.g. for folder names) prior
to Zephyr being found for things like APPLICATION_CONFIG_DIR. Also
adds NORMALIZED_BOARD_TARGET.
Signed-off-by: Jamie McCrae <jamie.mccrae@nordicsemi.no>
Fixes: #71384
A VERSION file placed in `/` or `<drive>:\` was accidentally being
picked up during `find_package(Zephyr)`.
This happened because Zephyr loads the VERSION file to determine if it
is the correct Zephyr to use.
During initial phase of find_package(), then APPLICATION_SOURCE_DIR is
not defined, causing one version file to be picked up from `/VERSION`
instead of `<app>/VERSION`. `/VERSION` is outside any Zephyr repo, west
workspace, or the application itself and therefore should not be picked
up accidentally.
Fix this be checking that APPLICATION_SOURCE_DIR is defined, and only
when defined, look for a VERSION file there.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Fix issue #67244 in which boards from external modules where not
printed in the help message (displayed for `shields` target or
when trying to build with unknown shield). The SHIELD_LIST variable
was reset for each BOARD_ROOT discovered.
Simplify discovery logic. Discover all shields and set relevant
variables (SHIELD_LIST and SHIELD_DIR_${name}) in one step and
process all the expected boards in the next step.
Signed-off-by: Marek Metelski <marek@metelski.dev>
This commit adds support for building relocatable (partially linked)
ELF files as the binary object type for the llext subsystem.
Signed-off-by: Luca Burelli <l.burelli@arduino.cc>
This change allows the `add_llext_target` function to accept multiple
source files when building an ELF shared library. The ELF object
target type is still limited to a single source file, since there is no
linking step in that case.
Also fixes a minor typo in another llext function documentation.
Signed-off-by: Luca Burelli <l.burelli@arduino.cc>
Add a new Kconfig option to select the binary object type for the llext
subsystem. This will allow to fully decouple the architecture type from
the kind of binary object that is expected by the loader.
The defaults have been chosen to match the current behavior of the ARM
and Xtensa architectures, but developers can now more easily experiment
with other object types.
Signed-off-by: Luca Burelli <l.burelli@arduino.cc>
This allow developers to create board files without the SoC name when
the board only defines a single SoC.
This means that a board, such as rpi_pico, which defines only a single
SoC, rp2040, and one variant, now allows the following file names:
Board target: rpi_pico/rp2040
- dts: rpi_pico_rp2040.dts, short: rpi_pico.dts
- defconfig: rpi_pico_rp2040_defconfig, short: rpi_pico_defconfig
- overlay: rpi_pico_rp2040.overlay, short: rpi_pico.overlay
- conf: rpi_pico_rp2040.conf, short: rpi_pico.conf
Board target: rpi_pico/rp2040/w
- dts: rpi_pico_rp2040_w.dts, short: rpi_pico_w.dts
- defconfig: rpi_pico_rp2040_w_defconfig, short: rpi_pico_w_defconfig
- overlay: rpi_pico_rp2040_w.overlay, short: rpi_pico_w.overlay
- conf: rpi_pico_rp2040_w.conf, short: rpi_pico_w.conf
A multi CPU cluster board, nrf5340dk:
Board target: nrf5340dk/nrf5340/cpunet
- dts: nrf5340dk_nrf5340_cpunet.dts, short: nrf5340dk_cpunet.dts
- defconfig: nrf5340dk_nrf5340_cpunet_defconfig,
short: nrf5340dk_cpunet_defconfig
- overlay: nrf5340dk_nrf5340_cpunet.overlay,
short: nrf5340dk_cpunet.overlay
- conf: nrf5340dk_nrf5340_cpunet.conf, short: nrf5340dk_cpunet.conf
A multi SoC board, nrf5340dk (real: nrf52840, emulated: nrf52811):
Board target: nrf52840dk/nrf52840
- dts: nrf52840dk_nrf52840.dts, short: Not possible
- defconfig: nrf52840dk_nrf52840_defconfig, short: Not possible
- overlay: nrf52840dk_nrf52840.overlay, short: Not possible
- conf: nrf52840dk_nrf52840.conf, short: Not possible
If two conflicting files are found, for example both
rpi_pico_rp2040.overlay and rpi_pico.overlay, then an error is raised.
If short form is detected for a board target with multiple SoCs, for
example nrf52840dk_nrf52840.overlay and nrf52840dk.overlay, then an
error is raised.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
This change reverts a change that was introduced with hwmv2 which
allowed for using common board names for overlays, given the board
target ``nrf9160dk@0.7.0/nrf9160/ns``.
In hwmv1 this would have used:
``nrf9160dk_nrf9160_ns.conf``.
``nrf9160dk_nrf9160_ns_0_7_0.conf``.
In hwmv2 this would have used:
``nrf9160dk_nrf9160_ns.conf``
``nrf9160dk_nrf9160_ns_0_7_0.conf``
``nrf9160dk_nrf9160.conf``
``nrf9160dk_nrf9160_0_7_0.conf``
``nrf9160dk.conf``
``nrf9160dk_0_7_0.conf``
With these changes, the following are used (which restores the hwmv1
behaviour):
``nrf9160dk_nrf9160_ns.conf``
``nrf9160dk_nrf9160_ns_0_7_0.conf``
For a board with a default SoC which is not a variant for example
``rpi_pico`` then ``rpi_pico.conf`` will also be used, this file
will not be used for variants e.g. ``rpi_pico/rp2040/w``
This applies to .dts, .conf and .overlay files in the boards
directory, and to .conf and .overlay files in application board
overlay directories.
This revert is needed to avoid issues whereby variants have
incompatible configuration to the parent board target, which has
been affecting samples and tests.
Signed-off-by: Jamie McCrae <jamie.mccrae@nordicsemi.no>
Add a new CMake helper to add new include directories to be parsed by
the syscall machinery. This helper complements the existing
zephyr_syscall_header, which works at a header-level.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>
Instead of INTERFACE_INCLUDE_DIRECTORIES, because header files are used,
not directories. Headers were added to both, target sources and include
directories.
Signed-off-by: Gerard Marull-Paretas <gerard@teslabs.com>