feat(freertos): added option for automatically placing task stacks in PSRAM
Closes IDF-15538, IDF-15539, and IDF-15540
See merge request espressif/esp-idf!48953
- Harden prefix_map.cmake for paths with spaces on Windows
- ldgen: fragments-list-file, list normalization, CMake integration (tools/cmake + cmakev2)
- CI exclude list for check_tools; test_spaces bundle tweak
Made-with: Cursor
cmakev2 does not use configdep yet, so the manually specified variable
CONFIGDEP_ENABLE is unused. CMake warns about it, which causes CI to
fail. Until configdep is adopted by cmakev2, suppress the warning.
Co-authored-by: Cursor <cursoragent@cursor.com>
Add build-level tests verifying configdep selective rebuild behavior:
- test_configdep.py: selective rebuild (only affected .obj rebuilt) and
ELF string verification after Kconfig toggle
- test_rebuild_configdep.py: full rebuild graph tests with configdep enabled,
including the two-build "settle" pattern for .cdep stub timestamps
Place the configdep test app in tools/test_build_system/configdep_test_app/
alongside the build system test infrastructure. Split configdep-specific
rebuild coverage from pure CMake/Ninja rebuild tests (test_rebuild.py runs
with configdep disabled via a module-local idf_py fixture).
The esp-idf-configdep utility works as a compiler wrapper. It lets the
compiler compile the app and checks whether -MF option was used to
generate dependency files. If yes, it checks for sdkconfig.h dependency
in those files.
Target file depending on sdkconfig.h is then scanned for all CONFIG_*
occurrences and the dependency file is altered so that the given target
file is marked as "dirty" only if the config options it actually uses
are changed.
This ensures that after a configuration update, only affected files are
rebuilt, reducing incremental build times.
The compiler launcher mechanism is changed from RULE_LAUNCH_COMPILE
(Makefile-only) to CMAKE_*_COMPILER_LAUNCHER (generator-agnostic),
enabling launcher chaining (configdep -> ccache -> compiler).
Made-with: Cursor
When esp-idf-kconfig >= 3.9.0 is installed, we can use the optimized
menuconfig target without the need to preprocess the sdkconfig file by
kconfgen (removing deprecated options - menuconfig - readding deprecated
options).
Also removed unused {menuconfig_depends} from the new target.
Co-authored-by: Cursor <cursoragent@cursor.com>
Refactor the esp_err_to_name() system to decouple esp_common from
higher-level components. Instead of a monolithic generated table,
each component registers its error codes into a dedicated linker
section (.esp_err_msg_table) via idf_define_esp_err_codes() in its
CMakeLists.txt.
New files:
- tools/err_codes_extract.py: extract ESP_ERR_* defines from headers to CSV
- tools/err_codes_to_c.py: generate C source placing entries into linker section
- tools/err_codes_to_rst.py: generate RST documentation from error codes
- tools/cmake/err_codes.cmake: CMake module providing idf_define_esp_err_codes()
- components/esp_common/include/esp_err_codes.h: esp_err_msg_t typedef
- components/esp_common/src/esp_err_to_name_new.c: new lookup using link-time array
- tools/test_apps/build_system/err_codes_check/: CI test app
Changes:
- Remove all optional component dependencies from esp_common/CMakeLists.txt
- Add .esp_err_msg_table section to all 5 linker scripts
- Register error codes in 18 components via idf_define_esp_err_codes()
- Add new scripts to .gitlab/ci/rules.yml build_check patterns
- use new scripts to generate doc and add CI validation
- Update esp_err.rst to add description of composable code registration
Add USB DFU and OTG console SOC capability flags for esp32s2, esp32s3, and esp32p4.
Use these caps in Kconfig, documentation conditionals, and idf.py DFU actions so USB support
is derived from SOC capabilities instead of hardcoded target names.
Calling idf.py size on linux target project would cause an
error. This was a problem for IDEs or build script which
often call this as the last step in the build process,
as it could mark the whole "build" as failed.
In cmakev1, the executable target is named "${project}.elf". In cmakev2,
the executable is named "${project}" with .elf as the output suffix.
Strip the .elf suffix and look up the bare target name when the original
name doesn't exist.
Two compatibility fixes:
- idf_component_register: apply separate_arguments to REQUIRES and
PRIV_REQUIRES after cmake_parse_arguments. Some managed components
publish CMakeLists.txt with PRIV_REQUIRES "log esp_eth" as a single
quoted string (e.g. espressif__rtl8201).
- Add register_component() macro and idf_component_add_link_dependency()
shims for legacy ESP-IDF examples (ULP apps, BLE mesh) and managed
components (esp_flash_nor).
kconfgen runs while the component manager iterates to convergence.
Those passes operate on partial component sets and emit "unknown
kconfig symbol" warnings for symbols defined in not-yet-downloaded
components — idf-build-apps treats those as build failures.
Suppress kconfgen output on the intermediate passes; only the final
pass against the converged set emits warnings.
The cmakev2 compat layer's idf_component_register() was passing
${sources} quoted to add_library(), collapsing the semicolon-delimited
list into a single argument. Drop the quotes so each source becomes a
separate argument, matching cmakev1's idf_component_register().
The CMakeLists injects CONFIG_ESP_REV_MIN_FULL and CONFIG_XTAL_FREQ on
the command line because those symbols live in esp_common's Kconfig,
which is outside the G0 closure and is not loaded during a restricted
COMPONENTS build. When every discovered component's Kconfig is processed
(IDF_BUILD_V2) the symbols are already defined via their Kconfig
defaults, and the additional -D would conflict with the existing
definition.
Gate the injection so it only runs when the Kconfig is not processed.
g1_components forces VFS_SUPPORT_IO=n in sdkconfig.defaults to keep
vfs and its transitive deps out of the closure. When the restricted
COMPONENTS list does not load components/vfs/Kconfig, the symbol is
unknown at config time and kconfgen emits an "unknown kconfig symbol"
warning that idf-build-apps treats as a build failure.
Add an ignore pattern so the warning does not gate the build.
g1_components builds a minimal COMPONENTS list. When every component's
Kconfig is processed, components/vfs/Kconfig sets VFS_SUPPORT_IO=y by
default, and the conditional requires in esp_stdio and esp_libc pull
vfs, esp_driver_uart, and their transitive deps into the closure.
Add an sdkconfig.defaults that forces the symbol off so the dependency
graph stays restricted to the components actually under test.
expected_dep_violations only listed the edges visible when
idf_component_optional_requires was a no-op for the dependency graph.
Under build system v2's DEFERRED mode the call propagates the resolved
dependency back into PRIV_REQUIRES, and the graph also captures
esp_common -> {efuse, bootloader_support, app_update} and esp_system ->
esp_app_format.
Add those entries so the script does not flag them when IDF_BUILD_V2
is set.
main/{esp32s2,esp32c3,esp32s3}/test_panic.c uses WDT registers from
esp_hal_wdt. Declare it explicitly so the build works under cmakev2's
strict component isolation.
JiangGuangMing
Marius Vikhammer
Jakub Kocka
Roland Dobai
Jan BeranandCursor
Marek Fiala
morris
Fu Hanxi
Guillaume Souchere
Igor Udot
Tomas Rezucha
igor.masar
Erhan Kurubas
Sudeep Mohanty
Renz Christian Bagaporo
Ivan Grokhotkov
Renz Bagaporo
Wu Zheng Hui
Chen Ji Chang
Ivan Grokhotkov
Frantisek Hrbata