fix(components): linux host-build failure fixes+ disable esp_tee in buildv2 + esp_hw_support and xtensa fixes
See merge request espressif/esp-idf!48743
httpd_stop() and httpd_req_async_handler_complete() both pushed
messages onto the control mbox via cs_send_to_ctrl_sock() without
reserving a slot in ctrl_sock_semaphore. Once the silent-drop fix
made the semaphore unconditional, the bypass became a real bug:
when the mbox is saturated by pending httpd_queue_work() items the
unguarded sendto() can return ENOBUFS, and even when it succeeds it
leaves the semaphore overstating free slots until the consumer
drains the message — a window during which a concurrent
httpd_queue_work() can take a slot but still find the mbox full.
Acquire the semaphore (portMAX_DELAY) before both sends and give it
back on send failure so the take/give invariant is preserved. The
httpd task is the consumer in both paths, so blocking is bounded
and deadlock-free. Reword the stale "no-op give on full" comment in
httpd_process_ctrl_msg() to reflect that only the recv-error path
relies on the cap behavior now.
The skip_common_name flag was named for the legacy CN field but actually
suppresses the entire mbedtls_ssl_set_hostname() call -- disabling
hostname matching against CN/SAN AND Server Name Indication. Update the
doxygen to describe the real effect, and emit a per-call WARN inside the
SNI-disable branch so debug-only use does not slip into production
unnoticed.
The session-resumption else-if in set_client_config() short-circuited
the CA verification chain when only client_session was supplied. Remove
the branch so session-only configs fall through to the normal error /
skip-verify path; resumption no longer silently disables CA validation.
esp_http_client_set_redirection() now rejects any redirect target whose
scheme is not https:// when the origin is HTTPS. This catches http, ftp,
ws and any other scheme before client state is mutated. Same-host /
https-to-https redirects are unaffected. Apps that intentionally want
mixed-scheme redirects can set disable_auto_redirect=true and handle
HTTP_EVENT_REDIRECT.
Moved the linux spinlock.h stub into esp_hw_support/include/linux.
esp_hw_support's INCLUDE_DIRS already places include/linux before
include, so the correct include files are picked up for the linux
target.
- Drop the FreeRTOS-Kernel-SMP linux spinlock.h duplicate.
- rv_decode_test: use PRIV_REQUIRES esp_hw_support instead of
hardcoding component include paths.
The baremetal RTOS stub gating tested the invisible-bool CONFIG_FREERTOS_PORT,
which can be set even when freertos is outside the build closure (e.g. the
G0 test app). Query __COMPONENT_REQUIRES_COMMON directly — it is the surface
those apps use to declare their closure.
The linux host_test target has no real flash, so the "Serial flasher
config" menu's options are meaningless there. Gate the menu on
!IDF_TARGET_LINUX so its symbols stay undefined on linux.
- Unset IDF_BUILD_V2 in subproject/CMakeLists.txt before including
project.cmake so the nested esp_tee ExternalProject cmake stays on
cmakev1.
- Read PARTITION_TABLE_BIN_PATH as a generator expression under
IDF_BUILD_V2 to defer the lookup to CMake's generate phase.
wuzhenghui
Wan Lei
Island
Mahavir Jain
Tomas Rohlinek
Armando (Dou Yiwen)
Chen Chen
Rahul Tank
morris
Gao Xu
Erhan Kurubas
Aditya Patwardhan
Sudeep Mohanty
Wang Lei
Ashish Sharma
Laukik Hase
harshal.patil
Renz Christian Bagaporo
Jiang Jiang Jian
Ondrej Kosta
Ivan Grokhotkov