Sub-projects like ULP are built as external CMake projects via
externalproject_add() and may use toolchains that are completely
different from the IDF target toolchain. For example, the ULP RISC-V
co-processor on an xtensa target (ESP32-S2/S3) uses
riscv32-esp-elf-gcc while the main project uses xtensa-esp-elf-gcc.
This causes problems in cmakev2:
1. __init_toolchain() in idf.cmake validates that CMAKE_TOOLCHAIN_FILE
matches IDF_TARGET and then overrides it. For sub-projects with
custom toolchains (passed via -DCMAKE_TOOLCHAIN_FILE from the
parent), this validation fails because the toolchain doesn't follow
the IDF naming convention (toolchain-<target>.cmake).
2. Several component project_include.cmake files use functions from
the IDF toolchain response file machinery (idf_toolchain_add_flags,
idf_toolchain_remove_flags, idf_toolchain_rerun_abi_detection) that
are not available in sub-projects with custom toolchains, or perform
validation that assumes the IDF target toolchain is in use.
Introduce IDF_CUSTOM_TOOLCHAIN, a CMake variable passed via -D from
the parent build to indicate that the toolchain was provided externally
and should not be resolved or validated by the IDF build system.
When IDF_CUSTOM_TOOLCHAIN is set:
- __init_toolchain() records the pre-set CMAKE_TOOLCHAIN_FILE and
skips IDF_TARGET-based resolution and validation.
- Component project_include.cmake files that manipulate IDF toolchain
flags (xtensa, esp_libc, esp_psram, soc) return early, as those
operations are not applicable to custom toolchains.
This is a generic mechanism usable by any sub-project that provides
its own toolchain (ULP, or any future sub-project with a non-IDF
toolchain).
Signed-off-by: Frantisek Hrbata <frantisek.hrbata@espressif.com>
Remove ~50 duplicate local definitions of ALIGN_UP/ALIGN_DOWN/ALIGN_UP_BY/
ALIGN_DOWN_BY across the codebase and replace them with canonical
ESP_ALIGN_UP/ESP_ALIGN_DOWN from esp_macros.h.
GCC 16 flags the volatile accumulator in the cache writeback loops even
though it is volatile, as it is only used to update itself. Drop the
accumulator and rely on the volatile-qualified psram pointer to keep
the cache-line reads from being optimized out.
The previous wording "top of PSRAM" was ambiguous: the carve-out is
actually mapped at the highest physical addresses of PSRAM (after the
rodata, text, and main heap mappings). Update the Kconfig help text for
SPIRAM_ENC_EXEMPT and SPIRAM_ENC_EXEMPT_SIZE, the External RAM
documentation, and the internal layout comment to say "upper end of PSRAM
(highest physical addresses)" instead.
Drivers that allocate from the unencrypted PSRAM carve-out via
MALLOC_CAP_SPIRAM_NO_ENC currently have no way to verify after the fact
which pool a buffer came from. This is particularly relevant for callers
using heap_caps_malloc_prefer(MALLOC_CAP_SPIRAM_NO_ENC, MALLOC_CAP_SPIRAM),
where a silent fallback to encrypted PSRAM would still pass the typical
esp_ptr_external_ram() check.
Expose esp_psram_ptr_is_no_enc() in the public esp_psram.h header. It
performs a range check against the carve-out's virtual-address window and
returns false when PSRAM is not initialized or CONFIG_SPIRAM_ENC_EXEMPT is
disabled, so callers do not need to guard the call site with #if.
Also reference the helper from the External RAM documentation alongside
the heap_caps_malloc(MALLOC_CAP_SPIRAM_NO_ENC) usage example.
If the virtual-address pool is exhausted when reserving the unencrypted
PSRAM carve-out (the warning at L388 may already have fired for the main
mapping), esp_mmu_map_reserve_block_with_caps() returns an error rather
than aborting. Convert the previous assert() into a logged fallback that
disables the carve-out for this boot, mirroring the SPIRAM_ENC_EXEMPT_SIZE
>= psram_available_size path.
Also extend esp_psram_extram_test() to run the standard memory test on
the carve-out region when CONFIG_SPIRAM_ENC_EXEMPT is enabled, so the
unencrypted mapping is exercised on startup like the other PSRAM regions.
Extends the External RAM encryption section to describe
CONFIG_SPIRAM_ENC_EXEMPT and the MALLOC_CAP_SPIRAM_NO_ENC heap
capability, including a security warning and a typical DMA-alignment
use case. Mirrors the change in zh_CN.
Adds CONFIG_SPIRAM_ENC_EXEMPT, available on chips that support per-page
PSRAM encryption configuration (esp32c5, esp32c61, esp32p4). When
enabled, esp_psram carves CONFIG_SPIRAM_ENC_EXEMPT_SIZE off the top of
PSRAM and maps it via the new mmu_hal_map_region_no_enc() helper, which
writes MMU entries without the SENSITIVE bit. The region is registered
as a separate heap pool reachable only through the new
MALLOC_CAP_SPIRAM_NO_ENC capability bit, so default SPIRAM allocations
cannot accidentally land there.
PSRAM encryption imposes alignment constraints that some DMA engines
(e.g. 2D-DMA) cannot satisfy. This option lets such workloads place
their buffers in unencrypted PSRAM while keeping the rest of PSRAM
(and flash) encrypted. Default disabled; security implications are
documented in the Kconfig help text.
A typical scenario is: when XIP on PSRAM enabled, compiler optimization level is Os. Under certain binary layout, boot hangs and backtrace points to `esp_sleep_config_gpio_isolate`.
The root cause is that, during PSRAM initialization, it calls esp_gpio_reserve, which happens to place before the reported function. However, after call, there is no barrier before the clock adjustment in `mspi_timing_enter_low_speed_mode`. The clock gets changed when the cache is still fetching data, resulting in the corrupted data in the end of the cache line.
This commits add spi_flash_disable_cache as a barrier to make sure the cache transactions is finished before the clock switch.
Clarify that MALLOC_CAP_DEFAULT describes a memory capability
rather than the malloc() placement policy, so PSRAM-backed
capability allocations are documented correctly when malloc()
remains internal-only by default.
Closes https://github.com/espressif/esp-idf/issues/18391
Made-with: Cursor
- Removed conditional definitions for various RCC_ATOMIC macros across
multiple files, replacing them with a unified PERIPH_RCC_ATOMIC() macro.
- Updated instances where specific RCC_ATOMIC macros were used to ensure
consistent usage of PERIPH_RCC_ATOMIC().
- Deleted unused uart_share_hw_ctrl.h file as its functionality is now
integrated into the new structure.
