As already done for the regular linker script, dinamically generates the
memory regions with the 'zephyr,memory-region' compatible also when
using the cmake linker generator.
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
Add a new cmake extension function:
dt_comp_path(<var> COMPATIBLE <compatible> [INDEX <idx>])
to get a list of paths for the nodes with the given <compatible>. This
is useful when we have to cycle through several nodes with the same
compatible.
Signed-off-by: Carlo Caione <ccaione@baylibre.com>
Zeroing the BSS and copying data to RAM with regular memset/memcpy may
cause problems when those functions are assuming a fully initialized
system for their optimizations to work e.g. some instructions require
an active MMU, but turning the MMU on needs the .bss section to be
cleared first, etc.
Commit c5b898743a ("aarch64: Fix alignment fault on z_bss_zero()")
provides a detailed explanation of such a case.
Replacing z_bss_zero() with an architecture specific one is problematic
as the former may see new sections added to it that would be missed by
the later. The same reasoning goes for z_data_copy().
Let's make maintenance much easier by providing weak versions of
memset/memcpy that can be overridden by architecture-specific safe
versions when needed.
Signed-off-by: Nicolas Pitre <npitre@baylibre.com>
As part of #40167 is was discovered that devicetree headers are always
generated when CMake re-runs.
This causes all source files that directly or indirectly through
included headers to recompile when CMake re-runs, even if there are no
changes to devicetree.
This commits introduces `zephyr_file_copy(...)` similar to
`file(COPY_FILE ...)` from CMake 3.21.
However, as CMake 3.20 is supported by Zephyr we need a zephyr variant
of this function to allow usage with CMake 3.20.
This ensures that only when there are changes to devicetree headers,
then source files will recompile.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Fixes: #39503Fixes: #40167
This commit moves `BUILD_VERSION` CMake variable from a compile
definition to be a define inside version.h.
Besides the benefit of having related settings grouped in a common
header, it also means that an updated `BUILD_VERSION` value does not
need to trigger re-compilation of all source files.
When using compile definitions, CMake cannot tell whether a given source
files uses the definition or not, and hence all sources must be
recompiled to be sure they are up-to-date.
Placing `BUILD_VERSION` in version.h, the source dependencies ensures
that only source files including `version.h` gets recompiled.
As part of this, version.h generation is moved so that it is now done
at build time.
This means that re-generation of version.h is no longer depending on a
CMake re-run but can have it's own dependency in `.git/index` when git
described is used to obtain `BUILD_VERSION` information.
Generation of logging dictionary database has been updated to support
BUILD_VERSION from header file instead of CMake configure time variable.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
GCC-based XCC uses GNU Assembler (xt-as). However, CMake doesn't
recognize it when invoking through xt-xcc. This results in CMake
going through all possible combinations of command line arguments
while invoking xt-xcc to determine assembler vendor. This multiple
invocation of xt-xcc unnecessarily lengthens the CMake phase of
build, especially when XCC needs to obtain license information
from remote licensing servers. So here forces the assembler ID
to be GNU to speed things up a bit.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This reverts commit deeb98da53.
A less invasive change has been implemented which does not
require changes to the GCC compiler flag file. So revert
the commit.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Some older versions of XCC Clang would result in the following
error during compilation:
/tmp/file.s: Assembler messages:
/tmp/file.s:20: Error: file number 1 already allocated
clang-3.9: error: Xtensa-as command failed with exit code 1
due to a bug in LLVM: https://bugs.llvm.org/show_bug.cgi?id=11740.
This is fixed in upstream, https://reviews.llvm.org/D20002, in 2016.
However, it seems that it is only fixed after XCC RI-2018.0.
Instead of blanket disabling usage of '-g', use an environment
variable "XCC_NO_G_FLAG" to disable usage of flag '-g' to workaround
this issue. This needs to be manually set because there is no way to
know which XCC version is being used, and compiler flag checking for
'-g' would not result in error (and thus '-g' is not ignored).
This is only needed for older XCC Clang. For sufficiently new XCC
verisons, there is no need for this.
Note that this is an alternative implementation to commit
deeb98da53. This one does not
alter the GCC flag file, and instead, simply clear the debug
compiler property so "-g" is not used.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
Some older versions of XCC Clang would result in the following
error during compilation:
/tmp/file.s: Assembler messages:
/tmp/file.s:20: Error: file number 1 already allocated
clang-3.9: error: Xtensa-as command failed with exit code 1
due to a bug in LLVM: https://bugs.llvm.org/show_bug.cgi?id=11740.
This is fixed in upstream, https://reviews.llvm.org/D20002, in 2016.
However, it seems that it is only fixed after XCC RI-2018.0.
Instead of blanket disabling usage of '-g', use an environment
variable "XCC_NO_G_FLAG" to disable usage of flag '-g' to workaround
this issue. This needs to be manually set because there is no way to
know which XCC version is being used, and compiler flag checking for
'-g' would not result in error (and thus '-g' is not ignored).
This is only needed for older XCC Clang. For sufficiently new XCC
verisons, there is no need for this.
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
This commit adds support for adjust the addresses of the final image.
This is useful when the image is to be flashed at a location different
from the LMA address encoded in the ELF file by the linker.
An example use-case is multicore systems where core A might load image
from a flash partition into RAM in order for core B to execute and load,
but where the image itself is build with the RAM addresses as LMA.
It updates the zephyr_image_info.h header with information of adjustment
value.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
The MIPS Malta is an ATX form factor evaluation board made by MIPS
Technologies. Malta board is the most popular platform for MIPS
full-system emulation.
See https://www.linux-mips.org/wiki/MIPS_Malta for details.
Signed-off-by: Antony Pavlov <antonynpavlov@gmail.com>
Signed-off-by: Remy Luisant <remy@luisant.ca>
This commit defines the toolchain command line options for the
Zephyr MIPS architectural port.
Signed-off-by: Antony Pavlov <antonynpavlov@gmail.com>
Signed-off-by: Remy Luisant <remy@luisant.ca>
Since CMake 3.20 DEPFILE has been supported by Makefile generators.
Simplify the CMake code by using DEPFILE for both Ninja and Makefile
generators.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
With CMake 3.20 relative path inside DEPFILEs are treated relative to
CMAKE_CURRENT_BINARY_DIR and are transformed by CMake in its internal
dep file.
Therefore Zephyr build system must no longer add `base_name` to the
`-MT` argument for the preprocessor.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
The GDB check routine added in the PR #38749 does not suppress the
console outputs (stdout and stderr) and may print out a misleading
error message during a CMake configuration when the required version
of Python is not available on the system:
arm-zephyr-eabi-gdb: error while loading shared libraries:
libpython3.8.so.1.0: cannot open shared object file: No such file or
directory
This commit adds the `OUTPUT_QUIET` and `ERROR_QUIET` options when
executing the GDB process so that the console outputs during the GDB
executable validation are not displayed to the user.
In addition, this commit removes the unused `GDB_PY_NO_PY` standard
output redirection variable since it is unnecessary.
Signed-off-by: Stephanos Ioannidis <root@stephanos.io>
Fixes: #41435
Remove build host specific escaping of start symbol command argument.
The start symbol for armlink is: Image$$device$$Base
and were escaped as: Image\\$$\\$$device\\$$\\$$Base
However, the $ must only be escaped in Linux and MacOS, not on windows
hosts.
Instead of escaping the start symbol in the CMake code then it is better
to use the VERBATIM flag on `add_custom_command()` which ensures correct
escaping for the build host.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Fixes: #41435
The use of %<symbol>% works well in Linux and MacOS but when passed
to the linker script generator in Windows the %<symbol>% is processed
by the windows shell causing the value to disappear and thus an empty
value inside the generated ld linker script or scatter file.
This is fixed by using the @ character instead of the % character.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
We used to test that the compiler works at all, before testing
compiler flags. But during some toolchain refactoring this got
re-arranged.
This commit corrects the ordering.
Signed-off-by: Sebastian Bøe <sebastian.boe@nordicsemi.no>
The commit dee79d2b66 introduced a new
linker output section ztest_suite_node in linker/common-ram.ld but
without updating the corresponding linker_script/common/common-ram.cmake
linker script generator code.
This causes test cases for arm Compiler 6 to fail.
Add the same section to common-ram.ld.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
The PR #40174 changed `configure_linker_script()` macro from handling a
single string to handle a list of strings for the
linker_pass_define argument.
This changed the armlink/target.cmake from STREQUAL to IN_LIST, however
as `configure_linker_script()` is a macro, then arguments are not
variables in the CMake sense but string replacements, and therefore
IN_LIST doesn't work directly on the argument.
Fix this by creating a true CMake list from the content of the argument
and use this list for IN_LIST checking.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
The armlink/target.cmake should not source ld/target_base.cmake file
as it provides its own (empty) implementation of the
`toolchain_ld_base` macro.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
This commit adds a change to support running west sign command even if
the keyfile is not provided. Default value of the configuration
is set to n in order to maintain backward compatibility.
Signed-off-by: Shubham Kulkarni <shubham.kulkarni@espressif.com>
The CMake policies CMP0002 and CMP0116 can now safely be removed.
Setting CMake minimum version to 3.20 ensures that policies are set to
NEW per default for those two policies:
> The cmake_minimum_required() command does more than report an error
> if a too-old version of CMake is used to build a project. It also sets
> all policies introduced in that CMake version or earlier to NEW
> behavior.
CMP0002 was introduced in CMake 2.6 and was set to NEW.
CMP0116 was introduced in CMake 3.20 and set to OLD because of:
> Use the old CMake behaviour until we are updating the CMake 3.20 as
> minimum required. This ensure that CMake >=3.20 will be consistent
> with older CMakes.
and can thus be removed as we are now having 3.20 as minimum required.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Remove xtensa specific workaround as this code is now present in Zephyr
SDK cmake code.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
With the revert of commit 820d327b4618d587a9cae0d085f00c6ab9c81cf2 then
some additional code can be cleaned up.
This removes the final left-overs from Zephyr SDK 0.11.1 support and
older.
It further aligns message printing when including Zephyr SDK toolchain
to other toolchain message printing.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
This reverts commit 820d327b4618d587a9cae0d085f00c6ab9c81cf2.
Commit b973cdc9e8 updated the minimum
required Zephyr SDK version to 0.13.
Therefore revert commit 820d327b4618d587a9cae0d085f00c6ab9c81cf2 as
backward support for 0.11.1 and 0.11.2 is no longer required.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Fix the zephyr_linker_dts_memory to lookup the correct property name
when trying to figre the region name.
Signed-off-by: Fabio Baltieri <fabio.baltieri@gmail.com>
After having resolved all of the instances of packed member access,
re-enable the warning.
Fixes#16587.
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Remove the `NAME` function argument from `zephyr_linker_dts_memory` as
the name is now automatically derived from the devicetree node.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
Update the linker script generator to automatically derive memory
region names from the devicetree node according to the same logic in
`devicetree_regions.h`.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
This causes any devicetree error messages to refer to the source files
which contain the errors instead of intermediate <BOARD>.dts.pre.tmp
files in the build directory.
Do this using a new python-devicetree feature which can correctly
handle cpp's generated line marker directives.
To use this feature, rework the way that the C preprocessor is run on
devicetree inputs so that line marker directives are preserved by
removing the -P option.
This is a backwards incompatible change to the way the generated files
in the build directory used to work, as not all tools can consume line
markers. In particular, dtc can't handle these lines. We therefore
pass dtc zephyr.dts instead (the final parsed devicetree output from
python-devicetree).
Since <BOARD>.dts.pre.tmp is a publicly documented file, this is
changing existing behavior, so use a new file name for the
intermediate files to make it more obvious that something changed. In
particular, use zephyr.dts.pre instead of <BOARD>.dts.pre.tmp.
(The $BOARD.dts.pre.tmp name is a little cumbersome anyway.)
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
Curently, any and all devicetree syntax errors are being handled by
dtc, because it runs before gen_defines.py. This means that any
improvements we make to devicetree error handling and reporting in
dtlib or edtlib will never be noticed by users that have dtc
installed.
(This also has the unfortunate effect of hiding dtlib error reporting
from me, since my systems all have dtc installed.)
Move dtc to after gen_defines.py, so that initial error handling and
warnings all come from python. If DTC has some additional warnings or
errors to report, let them come last. They will still halt the build
if anything is wrong.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
When multiple shields are defined, only the shield last in the -DSHIELD
list gets defined in `.config`. This is due to too many backslashes
used defining it for an env setting.
Signed-off-by: Ryan McClelland <ryanmcclelland@fb.com>
In the python APIs we use for devicetree, you can generally use an
alias interchangeably with the full path to a node.
For example, dtlib.py says:
[...] both dt.get_node("/foo/bar") and dt.get_node("bar-alias")
will return the 'bar' node below:
/dts-v1/;
/ {
foo {
bar_label: bar {
baz {
};
};
};
aliases {
bar-alias = &bar-label;
};
};
Fetching subnodes via aliases is supported:
dt.get_node("bar-alias/baz") returns the 'baz' node.
Now that we have the necessary helper functions for resolving alias
names, make the CMake APIs to devicetree behave the same way. This
improves consistency between the two sets of devicetree helpers used
by the build system, making them both match the general devicetree
practice that an alias is as good as a path. This is good for
consistency.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
This is a helper function for looking up the path to a devicetree
alias. It is analogous to dt_chosen(). It has to exist as a separate
function from dt_prop() for similar reasons that dt_chosen() does:
the edtlib APIs for interacting with /aliases are special-cased in the
same way they are for /chosen.
Signed-off-by: Martí Bolívar <marti.bolivar@nordicsemi.no>
Fixes an issue on SOF-supported boards where the build system
incorrectly believed that the rimage target was changing when rerunning
CMake, even though the user did not change any configuration. This issue
was introduced in commit 525fa76f4d.
Create initial build dir:
$ west build --board intel_adsp_cavs18 samples/subsys/audio/sof
Re-run CMake with no arguments in that build dir:
$ cd build
$ cmake .
CMake Warning at /home/mhelm/zephyrproject/zephyr/cmake/extensions.cmake:2403 (message):
The build directory must be cleaned pristinely when changing rimage target,
Current value="", Ignored value="cnl"
Call Stack (most recent call first):
/home/mhelm/zephyrproject/zephyr/cmake/extensions.cmake:829 (zephyr_check_cache)
/home/mhelm/zephyrproject/zephyr/boards/xtensa/intel_adsp_cavs18/board.cmake:6 (board_set_rimage_target)
/home/mhelm/zephyrproject/zephyr/cmake/app/boilerplate.cmake:635 (include)
/home/mhelm/zephyrproject/zephyr/share/zephyr-package/cmake/ZephyrConfig.cmake:24 (include)
/home/mhelm/zephyrproject/zephyr/share/zephyr-package/cmake/ZephyrConfig.cmake:40 (include_boilerplate)
CMakeLists.txt:5 (find_package)
Signed-off-by: Maureen Helm <maureen.helm@intel.com>
Create the ZEPHYR_MODULES_NAMES list of zephyr modules present in the
system before calling roots defined in `MODULE_EXT_ROOT`.
This will allow external module roots to process based on Zephyr modules
available in the system.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Fixes: #40590
This commit updates gen_app_partitions.py to include only files present
in the current build by extracting the information from the CMake
generated `compile_commands.json` file.
This ensures that object files in sub-projects, such as `empty_cpu0`,
will not be considered by the script.
Using the compile_commands.json instead of walking the whole build tree
for finding object files also improves performance:
Time of executing `gen_app_partitions.py` (Old):
__________________________
Executed in 480.06 millis
usr time 425.83 millis
sys time 49.55 millis
Time of executing `gen_app_partitions.py` (New):
________________________________________________________
Executed in 76.22 millis
usr time 49.00 millis
sys time 24.59 millis
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
boilerplate.cmake contained a comment referring to CMake 3.13.1 but
minimal required version is 3.20, so update the comment to 3.20.
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Removes hardcoded logic in the west signing script that translates
Zephyr board names to rimage target names. Instead, use a cached CMake
variable set at the board level to define its respective rimage target
name. This eliminates the need to modify the west signing script when
new SOF-supported boards are introduced to Zephyr.
Signed-off-by: Maureen Helm <maureen.helm@intel.com>
DTC can be built with both traditional Makefiles or Meson. When built
with Makefiles the --version output looks like 'Version: DTC
1.6.1-dirty' but when built with Meson the output is 'Version: DTC
v1.6.1+.
This fails to match the version regex and the cmake then fails:
CMake Error at cmake/host-tools.cmake:28 (if):
if given arguments:
"VERSION_GREATER" "1.4.6"
Unknown arguments specified
Expanding the regex with an optional 'v' covers both cases and the build
succeeds.
Signed-off-by: Ross Burton <ross.burton@arm.com>
This commit reworks the linker script generation and linking stages in
order to better support fixed section location as required by #38836.
Today we have the following generated linker scripts and the elf output
depending on the system configuration:
- linker_app_smem_unaligned.cmd --> app_smem_unaligned_prebuilt.elf
- linker_zephyr_prebuilt.cmd --> zephyr_prebuilt.elf
- linker.cmd --> zephyr.elf
as not all linker scripts may be created and as there is a need for the
possibility to move gen handles earlier then those stages has been
renamed into more generic names so that with this commit we have:
- linker_zephyr_pre0.cmd --> zephyr_pre0.elf
- linker_zephyr_pre1.cmd --> zephyr_pre1.elf
- linker.cmd --> zephyr.elf
This also means that is the stage zephyr_pre1 is not needed, then build
can go from `zephyr_pre0.elf` to `zephyr.elf`.
The gen_handles.py has been changed so it now uses `zephyr_pre0.elf` as
input. This ensures that the handles array are final when invoking the
next build and linking stages.
To keep the current behavior of generating the isr table and kobj hash
of what was `zephyr_prebuilt` stage the code blocks contolling isr
generation and kobj hash has been relocated to be located after
app_smem and device handle generation.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
Signed-off-by: Torsten Rasmussen <Torsten.Rasmussen@nordicsemi.no>
Introduce a new global property for source files that are generated from
the initial Zephyr link (app_smem_unaligned_prebuilt). This source list
is used for files which will introduce address shifts into the final
binary, which need to be present in `zephyr_prebuilt.elf` for
`CONFIG_USERSPACE` scripts to correctly generate `zephyr.elf`.
This resolves#38836.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
The previous description of the `GENERATED_KERNEL_*` cmake variables
was incorrect in that these are generated after the second link stage
when `CONFIG_USERSPACE` is enabled.
The sources in these variables are generated from `zephyr_prebuilt.elf`
for inclusion in `zephyr.elf`. However `zephyr_prebuilt.elf` is preceded
by `app_smem_unaligned_prebuilt.elf` when `CONFIG_USERSPACE` is enabled.
Signed-off-by: Jordan Yates <jordan.yates@data61.csiro.au>
