fdcavalcanti opened a new pull request, #18698: URL: https://github.com/apache/nuttx/pull/18698
## Summary <!-- This field should contain a summary of the changes. It will be pre-filled with the commit's message and descriptions. Adjust it accordingly --> * cmake: add -D__NuttX__ to preprocessor This change allows for linker script files to be properly processed when `#ifdef __NuttX__` is used. On Make builds, the `.ld` files added to `ARCHSCRIPT` are preprocessed to keep or remove lines that contain `#ifdef __NuttX__`. This same behavior is not present on CMake builds, which causes many linker scripts (on RISC-V Espressif builds at least) to be used improperly. This changes simply adds `-D__NuttX__` to the `nuttx_generate_preprocess_target` function, solving this problem. ## Impact <!-- Please fill the following sections with YES/NO and provide a brief explanation --> Impact on user: No. <!-- Does it impact user's applications? How? --> Impact on build: Yes, CMake only. Preprocess linker script files with the __NuttX__ option. <!-- Does it impact on building NuttX? How? (please describe the required changes on the build system) --> Impact on hardware: No. <!-- Does it impact a specific hardware supported by NuttX? --> Impact on documentation: No. <!-- Does it impact the existing documentation? Please provide additional documentation to reflect that --> Impact on security: No <!-- Does it impact NuttX's security? --> Impact on compatibility: No. <!-- Does it impact compatibility between previous and current versions? Is this a breaking change? --> ## Testing <!-- Please provide all the testing procedure. Consider that upstream reviewers should be able to reproduce the same testing performed internally --> Generate a ESP32-C6 build with and compare linker files used before and after this change. ### Building <!-- Provide how to build the test for each SoC being tested --> - `cmake -B build -DBOARD_CONFIG=esp32c6-devkitc:nsh -GNinja -DNXTMPDIR=on` - `cmake --build build --verbose > out` ### Results <!-- Provide tests' results and runtime logs --> Original content of `esp32c6.rom.newlib.ld`: ``` ... _isatty_r = 0x400004d4; strdup = 0x40000530; strndup = 0x40000554; #ifndef __NuttX__ rand_r = 0x4000058c; rand = 0x40000590; srand = 0x40000594; #endif atoi = 0x400005a0; atol = 0x400005a4; ... ``` Now we look into the generated `esp32c6.rom.newlib.ld.tmp` under `build` directory. Before the changes: ``` [4/1379] cd /home/fdcavalcanti/nuttxspace5/nuttx/build && /home/fdcavalcanti/work/toolchains/riscv-none-elf-gcc/bin/riscv-none-elf-gcc -E -P -x c -I/home/fdcavalcanti/nuttxspace5/nuttx/build/include -I/home/fdcavalcanti/nuttxspace5/nuttx/include -I/home/fdcavalcanti/nuttxspace5/nuttx/arch/risc-v/src/esp32c6 /home/fdcavalcanti/nuttxspace5/nuttx/build/arch/risc-v/src/common/espressif/esp-hal-3rdparty/components/esp_rom/esp32c6/ld/esp32c6.rom.newlib.ld > /home/fdcavalcanti/nuttxspace5/nuttx/build/esp32c6.rom.newlib.ld.tmp ``` ``` ... _isatty_r = 0x400004d4; strdup = 0x40000530; strndup = 0x40000554; rand_r = 0x4000058c; rand = 0x40000590; srand = 0x40000594; atoi = 0x400005a0; atol = 0x400005a4; ... ``` After the changes: ``` ... [8/1379] cd /home/fdcavalcanti/nuttxspace5/nuttx/build && /home/fdcavalcanti/work/toolchains/riscv-none-elf-gcc/bin/riscv-none-elf-gcc -E -P -x c -D__NuttX__ -I/home/fdcavalcanti/nuttxspace5/nuttx/build/include -I/home/fdcavalcanti/nuttxspace5/nuttx/include -I/home/fdcavalcanti/nuttxspace5/nuttx/arch/risc-v/src/esp32c6 /home/fdcavalcanti/nuttxspace5/nuttx/build/arch/risc-v/src/common/espressif/esp-hal-3rdparty/components/esp_rom/esp32c6/ld/esp32c6.rom.newlib.ld > /home/fdcavalcanti/nuttxspace5/nuttx/build/esp32c6.rom.newlib.ld.tmp ... ``` ``` ... _isatty_r = 0x400004d4; strdup = 0x40000530; strndup = 0x40000554; atoi = 0x400005a0; atol = 0x400005a4; strtol = 0x400005a8; strtoul = 0x400005ac; ... ``` You can see that `rand_r`, `rand` and `srand` are gone, as expected. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
