mirror of https://github.com/eclipse-threadx/threadx.git synced 2026-10-06 06:59:08 +08:00

Files

T

Frédéric DesbiensandTilen Majerle cf577c7137 Made ThreadX object names const-qualifiable behind an option (#761 )

Fixes #61

Object names are exposed as writable pointers throughout the kernel API, which
rejects string literals in C++ and lets a caller modify a name an object still
holds. Information services return those names through writable double
pointers.

Create services, control blocks, information services, the module manager and
trace registration now preserve const qualification, behind
TX_ENABLE_CONST_NAMES. The option defaults to off, so a build that says nothing
gets exactly the types it got before. It is opt-in rather than opt-out because
it changes the type of a public struct field: application code that copies a
name into a writable CHAR * stops compiling, which is a reasonable thing to ask
of a minor release and not of a patch one. Issue #780 tracks making it the
default in 6.6.

Two things the option reaches that its own call sites do not.
TX_CHAR_TO_UCHAR_POINTER_CONVERT has exactly two users, both of them reading an
object name in _tx_trace_object_register, and every form of that macro but the
MISRA one casts the qualifier away without saying so; the conversion is now
const in and const out, so nothing launders const to make the build pass. The
FreeRTOS adapter holds the name pcTaskGetName retrieves in a TX_NAME_CONST
pointer so that it tracks whichever declaration tx_thread_info_get has, and
keeps its writable return type through an explicit MISRA C:2012 Rule 11.8 cast,
because that signature is part of the FreeRTOS API.

Default build: all seven host configurations and all five SMP configurations
build with zero warnings and pass -- 113/113 on five host configurations,
100/100 on the two MISRA builds, 118/118 on SMP, 3/3 FreeRTOS. With
TX_ENABLE_CONST_NAMES set, the host default and both MISRA configurations, the
SMP trace configuration and the FreeRTOS adapter build with zero warnings and
pass.

Co-authored-by: Tilen Majerle <tilen@majerle.eu>
Assisted-by: Codex (gpt-6-astra) <noreply@openai.com>
Assisted-by: Claude Code (Opus 5) <noreply@anthropic.com>

2026-09-28 14:08:13 -04:00

cmake

Stopped the test runners from reporting an incomplete build as test failures (#709 )

2026-09-09 10:56:28 -04:00

fixtures

Added a regression suite for the FreeRTOS compatibility layer (#583 )

2026-08-09 08:03:16 -04:00

regression

Made ThreadX object names const-qualifiable behind an option (#761 )

2026-09-28 14:08:13 -04:00

readme.md

Added a regression suite for the FreeRTOS compatibility layer (#583 )

2026-08-09 08:03:16 -04:00

readme.md

FreeRTOS compatibility layer regression tests

Regression tests for utility/rtos_compatibility_layers/FreeRTOS/tx_freertos.c, built against the real ThreadX Linux port.

Running them

./scripts/build_freertos.sh
./scripts/test_freertos.sh

Or directly, which is the same thing:

test/freertos/cmake/run.sh build all
test/freertos/cmake/run.sh test all

CI runs both scripts through .github/workflows/regression_test.yml.

What they cover

Every function in the layer that creates an object takes one or two byte pool allocations for its bookkeeping and then creates one or more ThreadX kernel objects. When one of those kernel objects cannot be created, the function returns NULL, or pdFAIL for xTaskCreate(), and the caller is left with no handle and therefore no way to call the matching delete function. Anything the layer fails to release on the way out is lost until the system restarts.

That makes these error paths invisible from the outside: a leaking version and a correct version return exactly the same thing to the caller. The suite therefore counts the ThreadX primitives the layer reaches for, and checks that each error path gives back precisely what it took.

Test	Covers
`txfr_queue_create_test`	`xQueueCreate`, `xQueueCreateStatic`, `vQueueDelete`
`txfr_task_create_test`	`xTaskCreate`, `xTaskCreateStatic`
`txfr_sync_create_test`	semaphores, mutexes, event groups and timers

How the fault injection works

A test asks the harness to fail a chosen kernel creation call, then reads back how many allocations, releases, object creations and object deletions the layer performed. The interception is done with the linker's --wrap option, so tx_freertos.c is compiled exactly as it ships, with no test hooks in it.

Two things are worth knowing before adding tests:

tx_api.h maps the public API onto the error checking entry points, so the symbols that exist at link time are the _txe_ variants, and those are what the wrap list in cmake/regression/CMakeLists.txt names. A --wrap for a name that does not resolve is silently ignored, so a typo there produces a test that quietly never injects anything.
txfr_malloc() and txfr_free() cannot be wrapped, because they are defined in tx_freertos.c and called from within it, so the compiler resolves those calls internally. The byte pool counts stand in for them.

Because --wrap is a GNU ld and lld feature with no MSVC equivalent, this suite is Linux only. The CMake configuration stops with a clear message rather than failing later with confusing link errors.

Fixtures

fixtures/FreeRTOSConfig.h configures the layer for the tests. Two of its settings are not arbitrary:

configASSERT() and TX_FREERTOS_ASSERT_FAIL() are empty, since the tests drive error paths deliberately and neither may halt the run.
portDISABLE_INTERRUPTS() and portENABLE_INTERRUPTS() are defined up front. FreeRTOS.h picks those by compiler rather than by target, so a GNU build otherwise resolves them to the bare metal __disable_interrupts() intrinsic, which does not exist when the layer is hosted on Linux.

fixtures/tx_user.h supplies TX_THREAD_USER_EXTENSION, which the layer requires, as documented in the layer's own readme.md.

Why the build is 32 bit

The Linux port defines ULONG as unsigned int on x86_64, while the layer passes pointers through ULONG arguments, such as the task argument and the timer identifier. A 64 bit build truncates those pointers, which the compiler reports as -Wpointer-to-int-cast and which crashes the timer callback wrapper. The ThreadX and SMP suites build 32 bit for their own reasons; this suite has to.