When compiling for native, some glibc versions has a check on the
use of _FORTIFY_SOURCE without optimizations enabled and gives a
preprocessor warning.
Signed-off-by: Thomas Ebert Hansen <thoh@oticon.com>
Add "gap-fill 0xFF" option for the CMAKE_OBJCOPY command,
in order to reduce the flash programming time for some boards,
such as cc2650.
Signed-off-by: Walter Xie <41377148@qq.com>
Previously the syscall list was generated only from the include
folder. This is a limitation when the application tries to create
system calls. This patch create a simple way to include these
new syscalls without the application touching the kernel.
This can be enabled by a Kconfig CONFIG_APPLICATION_DEFINED_SYSCALL.
Once enabled the application source directory will be scanned to
find all application defined syscalls.
Signed-off-by: Adithya Baglody <adithya.nagaraj.baglody@intel.com>
The entry point can and therefore should be set by linker
scripts. Whenever possible one should express things in the source
language, be it .c or .ld, and not in code generators or in the build
system.
This patch removes the flag -eCONFIG_KERNEL_ENTRY from the linker's
command line and replaces it with the linker script command
ENTRY(CONFIG_KERNEL_ENTRY)
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Add an LLVM backend and a clang toolchain variant to support building
with llvm coming with popular Linux distributions.
This has been tested with X86 boards:
- quark_d2000_crb
- quark_se_c1000_devboard/Arduino 101
Use:
export ZEPHYR_TOOLCHAIN_VARIANT=clang
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Python scripts should be executed indirectly by the python
interpreter, not directly in a CMake COMMAND. This is for portability
reasons.
Directly executing a python script depends on non-portable shebang
logic and/or "default application" behaviour.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Fixes: #8380
This commit fixes the side-effect of PR #8211 where a 'ninja clean'
would try to remove dependency folders.
Changes:
- Symlinks are created during build and CMake targets now depends on
the symlinks. Thus, same behavior is achived with regards to
dependency handling, while at the same time, the output can be
cleaned as the dependencies are now attached to the symlinks.
- Dependencies have been changed so that generation of json files
depends on the trigger file and CMake depends upon the subdir txt
file. This prevents additional CMake runs after a clean.
Signed-off-by: Torsten Rasmussen <torsten.rasmussen@nordicsemi.no>
Both variables were used (with the same value) interchangeably
throughout CMake files and per the discussion in GH issue,
ZEPHYR_BASE is preferred.
Also add a comment with explanation of one vs. the other.
Tested by building hello_world for several boards ensuring no errors.
Fixes#7173.
Signed-off-by: Alex Tereschenko <alext.mkrs@gmail.com>
Zephyr currently allows users to choose a compiler optimization
between -O0, -Og and -Os, the default being the latter as it offers a
good compromise between speed and flash usage. In cases the speed
matters and/or the flash usage doesn't, optimizing for speed with -O2
is another alternative. For example in case of a simple application
doing cryptographic signature validation with mbedtls, the flash size
increase by about 15% compared to -Os, while it provides a 15% speed
boost for a RSA signature and 30% speed boost for a ECC signature.
This patches therefore adds a new option CONFIG_SPEED_OPTIMIZATIONS
corresponding to the -O2 flag, but keep the default set to
CONFIG_SIZE_OPTIMIZATIONS.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
It is currently not possible to define multiple compiler options using
the CONFIG_COMPILER_OPT Kconfig option. The string is interpreted as a
single quoted option, for example "-opt1 -opt2".
This patch fixes that by splitting the CONFIG_COMPILER_OPT string into
multiple options using the separate_arguments cmake function.
Signed-off-by: Aurelien Jarno <aurelien@aurel32.net>
Fixes: #8210
Following changes has been made to ensure correct behavior on different
system:
- Python script to detect changes to directories, including empty ones.
When files are modified the list is updated
If sub-directories are added / removed a trigger file is touched to
notify cmake to re-run
- Windows: To detect changes to header files in include for
parse_syscalls.py all files must be individual monitored.
Hence all headers are globbed added to dependencies.
CMake configure depends on the folders so the added /
removed files are picked up.
- Other: Folders are monitored through the python list file so that
added / remove / modified
Added / removed sub-directories are detected through trigger
file in order to re-run cmake.
Signed-off-by: Torsten Rasmussen <torsten.rasmussen@nordicsemi.no>
Fixes: #8210
This commit removes the need for always rebuild when generating the
syscalls.json.
It find the files and subfolders of the include directory and makes
syscalls depends on those.
To find added files / subfolders on all platforms CMake configure
is also depending on the folders found.
This ensures that added files / folders will also be detected at later
stages.
Signed-off-by: Torsten Rasmussen <torsten.rasmussen@nordicsemi.no>
Users are sometimes mistakenly invoking cmake with $ZEPHYR_BASE set as
the source directory.
This user-error can occur if the user believes that the toplevel
CMakeLists.txt file is at the root of the repo, or if the user uses
the wrong path when invoking cmake.
The error given when this user-error occurs is cryptic and undefined,
so we replace it with what should be an easy-to-understand error
message.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
This commit provides a minor fix to a CMake output text which
warns the user that ASSERT() statements are enabled.
Signed-off-by: Ioannis Glaropoulos <Ioannis.Glaropoulos@nordicsemi.no>
We haven't ever really had a Kconfig symbol to enable the support
related to DEBUG_SECTION_MISMATCH. Thus this isn't really used by
anyone and we will remove it.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
This patch fixes the dependency chain for priviliged stack
generation. This fixes a problem when compiling after making
significant changes that would shift the privileged stack area.
Signed-off-by: Andy Gross <andy.gross@linaro.org>
This patchset provides Xtensa's xcc compiler support for Xtensa
projects in Cmake. This requires the below environment variables
to be defined aptly. The appropriate xcc license information also
need to be supplied.
ZEPHYR_GCC_VARIANT=xcc
TOOLCHAIN_VER=RF-2015.3-linux
XTENSA_CORE=cavs21_LX6HiFi3_RF3_WB16
XTENSA_SYSTEM=/opt/xtensa/XtDevTools/install/tools/
RF-2015.3-linux/XtensaTools/config/
XTENSA_BUILD_PATHS=/opt/xtensa/XtDevTools/install/builds/
Change-Id: Ib3c10e8095439b0e32276ff37c00eca8420773ec
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
This reverts commit d9ac1d4b4a.
This option can be compiler specific and is not required anymore.
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
This commit ports nearly all usage of check_c_compiler_flag,
check_cxx_compiler_flag, and check_compiler_flag to use
zephyr_check_compiler_flag instead.
This has a significant CMake configure-time performance impact.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Adding a new kernel object type or driver subsystem requires changes
in various different places. This patch makes it easier to create
those devices by generating as much as possible in compile time.
No behavior change.
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
Driver APIs might not implement all operations, making it possible for
a user thread to get the kernel to execute a function at 0x00000000.
Perform runtime checks in all the driver handlers, checking if they're
capable of performing the requested operation.
Fixes#6907.
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
This is required to have BUILD_ASSERT() from toolchain/common.h to
work without issuing a warning.
Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
Specifying a custom linker script is supposed to be supported through
Kconfig, but evidently this has been unused and 100% broken at least
since the CMake migration.
This commit fixes
https://github.com/zephyrproject-rtos/zephyr/issues/6983
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
This uses the version and hash (git describe) and replaces the timestamp
currently used in the boot banner. This works much better than using
timestamps. It lets us point to the exact commit being used to run a
certain application or test.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Moving the ccache build script earlier in allows us to export ccache
to external projects. Using ccache with external projects, such as
OpenThread, can significantly speed up the build time.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
We need access to SOC_PATH in dts.cmake so we need to move the
definitions of SOC_NAME, SOC_SERIES, SOC_FAMILY, and SOC_PATH out of the
toplevel CMakeLists.txt and into cmake/app/boilerplate.cmake. We place
them before we include cmake/dts.cmake so they will be available to use
in it.
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
This patch adds an additional constraint to make sure that we only
do the application memory sizing if it is really necessary.
Signed-off-by: Andy Gross <andy.gross@linaro.org>
We want to support other toolchain not based on GCC, so the variable is
confusing, use ZEPHYR_TOOLCHAIN_VARIANT instead.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
To enable using application-defined sections in linker scripts, the
application source directory must be included in the search path when
running the preprocessor to generate linker scripts.
This prevents build failures when using any of the configuration options
CONFIG_CUSTOM_RODATA_LD, CONFIG_CUSTOM_RWDATA_LD and
CONFIG_CUSTOM_SECTIONS_LD.
Signed-off-by: Kristian Klomsten Skordal <kristian.skordal@nordicsemi.no>
In a scenario where a platform harbours multiple interrupts to the
extent the core cannot support it, an interrupt controller is added
as an additional level of interrupt. It typically combines several
sources of interrupt into one line that is then routed to the parent
controller.
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
This patch adds the generation and incorporation of privileged stack
regions that are used by ARM user mode threads. This patch adds the
infrastructure for privileged stacks. Later patches will utilize the
generated stacks and helper functions.
Signed-off-by: Chunlin Han <chunlin.han@linaro.org>
Signed-off-by: Andy Gross <andy.gross@linaro.org>
This patch adds application data section alignment constraints
to match the region definition requirements for ARM MPUs. Most MPUs
require a minimum of 32 bytes of alignment for any regions, but some
require power of two alignment to the size of a region.
This requires that the linker align the application data section to
the size of the section. This requires a linker pass to determine the
size. Once this is accomplished the correct value is added to a linker
include file that is utilized in subsequent linker operations.
Signed-off-by: Andy Gross <andy.gross@linaro.org>
This patch changes the way the custom linker function works. It uses
names instead of numbers to denote the name of the file and applies
the correct LINKER_PASS variable for the final linker pass.
Signed-off-by: Andy Gross <andy.gross@linaro.org>
We need to allocate extra memory region for the hash function
that the gperf generates. The linker will need to maintain
the same location counter between multiple linker stages. Reserving
memory will help in maintaing correct location counters between
multiple stages. The required memory varies with optimization level.
In order to keep the size of the compiled object consistent,
the kobject_hash.c is always compiled with -Os.
fixes#5226
Signed-off-by: Adithya Baglody <adithya.nagaraj.baglody@intel.com>
This patch does several things, most notably it changes the semantics
of CONFIG_DEBUG. CONFIG_DEBUG continues to behave as a vaguely defined
"debug mode" that enables printf's, -Og, etc. but now the user may
choose to be in "debug mode" while using a different optimization
level than -Og.
Tp support this a new config is defined to enable -Og;
CONFIG_DEBUG_OPTIMIZATIONS.
Additionally CONFIG_SIZE_OPTIMIZATIONS is introduced to allow the user
to explicitly request optimizing for size instead of relying on
defaulting to it.
The three config's {NO,SIZE,DEBUG}_OPTIMIZATIONS are now organized in
a Kconfig choice to ensure that at most one can be enabled at a time.
Finally, selected users of CONFIG_DEBUG have been ported to use one of
the optimizations configs when it was clear from usage that the
intention was to behave differently when using a different
optimization level and not when in "debug mode".
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Added a new config option which lowers the compiler
optimizations to -O0 independently of other flags.
CONFIG_COVERAGE uses it now instead of having its own
choice in CMakeLists.txt
Fixes#5720
Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
CMake's custom_command support is not great. It is technically
supported, but it is not easy to get it right.
As explained in "5. File-level dependencies of custom targets are not
propagated" from http://bit.ly/2GvwwEy. When a custom_command uses a
custom_target as input, the custom_command must DEPEND on both the
custom_target and the underlying file.
Longterm we need to do something more sofisticated to prevent these
issues from popping up. Like add some static analysis to detect badly
written CMake code or introduce an abstraction for custom commands
that is not affect by this issue.
This fixes#5881
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
This CMakeLists.txt is added via add_subdirectory() to the app's
CMakeLists.txt. If the app-CMakeLists.txt has install() rules and
wants to use them, setting CMAKE_SKIP_INSTALL_RULES to ON breaks
the 'make install' call.
Not setting this variable does not harm zephyr because zephyr does
not use install().
Signed-off-by: Patrick Boettcher <patrick.boettcher@posteo.de>
A fix for this issue is in progress, meanwhile warn the user that
they may be susceptible to this problem if they enable user mode on
an x86-based target that is not known to be immune.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Re-order the execution of the arch/ and subsys/ CMakeLists.txt code to
work around a manifestation of issue #6505. When OpenThread created an
External project in subsys, it did not have access to important
toolchain flags added in arch/.
Intuitively, subsystems might depends on how the ARCH is configured,
but the ARCH shouldn't depend on any subsystem.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Instead of accessing the environment variable ZEPHYR_BASE every time we
require accessing the source code root, use an intermediate variable
that has OS path separators correctly set to '/' to avoid issues on
Windows.
Note: This removes the ZEPHYR_SOURCE_DIR CMake variable. External
applications using that will need to change to use the new ZEPHYR_BASE
variable.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
A Zephyr library not having source files is usually due to a
misconfiguration and will lead to an obscure error message. But
imported Zephyr libraries don't have source files so we need to skip
the source file when a library is imported.
An imported Zephyr library is normally used to reference externally
built code, either prebuilt or recursively built through an external
build system.
This fixes#5497
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Depending on a path inside the Zephyr tree to determine if we are a test
does not scale. Also some samples were marked as TEST while they are
not, just to get some options defined for tests.
Idenitfying a test will be addressed in another patch introducing
CONFIG_TEST.
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
This fixes
https://github.com/zephyrproject-rtos/zephyr/issues/5010. CMake has a
simple parser that can parse C-like files to find header
dependencies. But the parser needs to know what the include
directories are to be able to follow #include pragmas.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
Add a comment that explains what the -P flag does. The explanatory
comment is useful when you want to find the flag so you can comment it
out while debugging the link stage.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
The linker script places kernelspace and userspace archives in
different sections. But the linker script itself does not determine
what archives are in what space, that is done by CMake.
CMake passes the list of kernelspace archives to the linker script
through defines, like this:
-DNUM_KERNEL_OBJECT_FILES=3
-DKERNEL_OBJECT_FILE_0=path/to/archive_a.a
-DKERNEL_OBJECT_FILE_1=path/to/archive_b.a
-DKERNEL_OBJECT_FILE_2=path/to/archive_c.a
These paths are relative, and since Ninja and Make invoke the linker
with different "working directories"[0], the relative paths need to be
different. This patch rectifies the relative path when using Ninja.
This fixes#5343
[0] https://gitlab.kitware.com/cmake/cmake/issues/17448
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
