build system: explicitly disable LTO plugin to reduce the number of simultaneously open files
Closes IDF-923 and IDFGH-1764
See merge request espressif/esp-idf!6776
... to reduce the number of simultaneously open files at link time.
When plugin support is enabled in the linker, BFD's (and the
corresponding file handles) are cached for the plugin to use. This
results in quite a large number of simultaneously open files, which
hits the default limit on macOS (256 files).
Since we aren't using LTO now, disable it explicitly when invoking the
linker.
Closes IDF-923
Closes IDFGH-1764
Closes https://github.com/espressif/esp-idf/issues/3989
Use rev-parse to get the HEAD directory instead of manually looking for
it. This method works in the main repository, worktrees and submodules.
Closes https://github.com/espressif/esp-idf/issues/4136
Ref. https://github.com/espressif/esp-idf/issues/1684
This change allows RTTI to be enabled in menuconfig. For full RTTI
support, libstdc++.a in the toolchain should be built without
-fno-rtti, as it is done now.
Generally if libstdc++.a is built with RTTI, applications which do not
use RTTI (and build with -fno-rtti) could still include typeinfo
structures referenced from STL classes’ vtables. This change works
around this, by moving all typeinfo structures from libstdc++.a into
a non-loadable section, placed into a non-existent memory region
starting at address 0. This can be done because when the application
is compiled with -fno-rtti, typeinfo structures are not used at run
time. This way, typeinfo structures do not contribute to the
application binary size.
If the application is build with RTTI support, typeinfo structures are
linked into the application .rodata section as usual.
Note that this commit does not actually enable RTTI support.
The respective Kconfig option is hidden, and will be made visible when
the toolchain is updated.
This commit makes it so that BUILD_COMPONENT holds only the component,
and a new property BUILD_COMPONENT_ALIASES hold the full name of the
component.
This also removes erroneous check for duplicate components, as this can
never happen:
(1) if two components have the same name but different prefixes,
the internal names are still unique between them
(2)if two components happen to have the same name and same prefix, the
latter would override the former
Previous implementation only builds list of components included in the
build during component registration.
Since the build components is known as the requirements expansion is
ongoing, update the list here instead.
Use imported library, which does not create additional rules, but still
allows attaching arbitraty properties instead of custom targets. This
allows the targets to not appear in the target list of IDEs such as
CLion.
Fixes an issue where if idf_build_process is called in a CMake
subdirectory, menuconfig looks for the mconf-idf binary in the wrong
place (in the subdirectory build dir instead of root binary dir).
- Makefile didn't quite work with out-of-tree build unless there was already an in-tree build done.
- CMake needs to delete some of the in-tree build artifacts or they're used instead of the correct
files.
- Makefile didn't quite work with out-of-tree build unless there was already an in-tree build done.
- CMake needs to delete some of the in-tree build artifacts or they're used instead of the correct
files.
How idf_build_component and the COMPONENTS argument to idf_build_process
interact is not clear/misleading. Clarify their interaction in the docs.
Closes: https://github.com/espressif/esp-idf/issues/3630
!4452 used setting LINK_LIBRARIES and INTERFACE_LINK_LIBRARIES to link
components built under ESP-IDF build system. However, LINK_LIBRARIES does
not produce behavior same as linking PRIVATE. This MR uses the new
signature for target_link_libraries directly instead. This also moves
setting dependencies during component registration rather than after all
components have been processed.
The consequence is that internally, components have to use the new
signature form as well. This does not affect linking the components to
external targets, such as with idf_as_lib example. This only affects
linking additional libraries to ESP-IDF libraries outside component processing (after
idf_build_process), which is not even possible for CMake<v3.13 as
target_link_libraries is not valid for targets not created in current
directory. See https://cmake.org/cmake/help/v3.13/policy/CMP0079.html#policy:CMP0079
Previously ldgen determines the output file name on its own. This commit
makes it so that user can dictate what the output file name will be
for the processed template, if the user needs it for something else.
!4452 simplified early expansion by using an early expansion script that
only does one thing: get the public and private requirements for each
component, albeit one by one. This was also dependent on parsing
the command output of the expansion script. This commit makes it so that a list of all
components to be processed to passed to the expansion script, generating a cmake
file that sets each component requirements in one go.
This also makes sure that only components that registered themselves get
included in the final build list.
!4452 had config generation first before building the component list
to be used in the build. This proved to be detrimental when a new target
is added as config generation would consider configs from both targets.
Removes the need to know/guess the paths to these libraries. Once we are gcc 8 only, we
can remove -nostdlib and no additional arguments are needed for system libraries.
The catch is: any time IDF overrides a symbol in the toolchain sysroot, we need
an undefined linker marker to make sure this symbol is seen by linker.
Adds --tags and --dirty flags to cmake git_describe() calls, and not
pass the HEAD commit hash (incompatible with --dirty)
Makes IDF_VER output the same as in Make build system
Thanks to @william-ferguson-au for reporting this:
https://github.com/espressif/esp-idf/issues/3378#event-2355460974
This MR removes the common dependency from every IDF components to the SOC component.
Currently, in the ``idf_functions.cmake`` script, we include the header path of SOC component by default for all components.
But for better code organization (or maybe also benifits to the compiling speed), we may remove the dependency to SOC components for most components except the driver and kernel related components.
In CMAKE, we have two kinds of header visibilities (set by include path visibility):
(Assume component A --(depends on)--> B, B is the current component)
1. public (``COMPONENT_ADD_INCLUDEDIRS``): means this path is visible to other depending components (A) (visible to A and B)
2. private (``COMPONENT_PRIV_INCLUDEDIRS``): means this path is only visible to source files inside the component (visible to B only)
and we have two kinds of depending ways:
(Assume component A --(depends on)--> B --(depends on)--> C, B is the current component)
1. public (```COMPONENT_REQUIRES```): means B can access to public include path of C. All other components rely on you (A) will also be available for the public headers. (visible to A, B)
2. private (``COMPONENT_PRIV_REQUIRES``): means B can access to public include path of C, but don't propagate this relation to other components (A). (visible to B)
1. remove the common requirement in ``idf_functions.cmake``, this makes the SOC components invisible to all other components by default.
2. if a component (for example, DRIVER) really needs the dependency to SOC, add a private dependency to SOC for it.
3. some other components that don't really depends on the SOC may still meet some errors saying "can't find header soc/...", this is because it's depended component (DRIVER) incorrectly include the header of SOC in its public headers. Moving all this kind of #include into source files, or private headers
4. Fix the include requirements for some file which miss sufficient #include directives. (Previously they include some headers by the long long long header include link)
This is a breaking change. Previous code may depends on the long include chain.
You may need to include the following headers for some files after this commit:
- soc/soc.h
- soc/soc_memory_layout.h
- driver/gpio.h
- esp_sleep.h
The major broken include chain includes:
1. esp_system.h no longer includes esp_sleep.h. The latter includes driver/gpio.h and driver/touch_pad.h.
2. ets_sys.h no longer includes soc/soc.h
3. freertos/portmacro.h no longer includes soc/soc_memory_layout.h
some peripheral headers no longer includes their hw related headers, e.g. rom/gpio.h no longer includes soc/gpio_pins.h and soc/gpio_reg.h
BREAKING CHANGE
Works around bug where components/soc/CMakeLists.txt was testing "EXISTS
${COMPONENT_PATH}/${SOC_NAME}" and this test could pass during early
expansion if COMPONENT_PATH was empty and a directory /esp32 exists
on Windows.
Closes https://github.com/espressif/esp-idf/issues/3195
A problem if the Python interpreter used for idf.py (or set via PYTHON
variable) didn't match
"/usr/bin/env python" (or the associated executable for .py files, on
Windows).
Closes https://github.com/espressif/esp-idf/issues/3160
Possibly also fix for https://github.com/espressif/esp-idf/issues/2936
Adds build system test to catch any future direct execution of Python in
the standard build process.
Previously, this compiler flag was not being applied
regardless of CONFIG_SPIRAM_CACHE_WORKAROUND setting.
Explanation: add_compile_options() only applies to
source files added after the function is run, or in
subdirectories added after the function is run. In
this case, no new source files were being added after
this function was run.
When IDF_PATH is not set by the user, cmake infers the same through
other means and sets it as an environment variable. However, some
sub-invocation of cmake may not see this variable as it is not set
in the parent. This change adds a custom command for exporting IDF_PATH
just before invoking ldgen so that IDF_PATH can be used in Kconfig
without any errors.
