rt-thread.git/src, branch v5.3.0

[utest] Enforce minimum console buffer size

2026-09-10T09:54:43Z

(cherry picked from commit 64f1fa4a0dd610b1dece5755207a54c0b324569f)

[kernel][ipc] Fix timeout wakeup ownership race

2026-08-21T08:52:01Z

Keep thread timer ownership when timer stop races with an in-flight timeout callback, and consume IPC resources only after a waiter is resumed successfully. Complete ownership cleanup in the generic and RT-Smart channel timeout callbacks. Add deterministic scheduler and LWP regressions for Event, Semaphore, Mailbox, Message Queue, sender, and receiver paths.

[kernel/memheap] fix: stop invalid frees from corrupting heap

2026-08-20T01:48:13Z

RT_ASSERT() does not stop execution when RT_DEBUGING_ASSERT is disabled. Return after the existing block validation so a sequential duplicate free cannot update heap accounting or the free list in release builds. Before fix: memheap double-free used: before=64 alloc=224 free=64 after=4294959232 Validation: scons -C bsp/qemu-vexpress-a9 -j$(nproc) Signed-off-by: Hui Su <3164683437@qq.com>

[mutex] 跨线程 rt_thread_delete 被删线程持有的mutex也一并释放

2026-07-22T01:45:59Z

- 增加单元测试 Clarified mutex release conditions and added exception handling for closed owner threads.

mem: guard rt_smem_setname() against a NULL name

2026-07-16T05:31:55Z

rt_smem_setname() dereferenced name[index] without checking for NULL first. Some callers legitimately pass an unnamed allocation through with a NULL name, which crashed instead of just leaving the trace field blank. Treat a NULL name the same as an empty string, so the existing fill-with-spaces loop below handles it without duplicating that logic. Fixes #9793 Signed-off-by: yi chen <94xhn1@gmail.com>

utest: fix hooklist testcase reinit

2026-06-26T09:07:01Z

fix: stabilize smp_bind_affinity_tc on SMP CI

2026-06-17T01:51:13Z

This change fixes intermittent failures of the `core.smp_bind_affinity` utest across SMP CI targets. The original failure was a flaky assertion on unbound threads (`thread_inc[x] != thread_tic[x]` in thread_entry), not incorrect bind_cpu behavior. CI logs already showed T0 binding worked (thread_inc[0] == thread_tic[0] == 100) while the not_equal check failed. The root issue is that unbound-thread CPU placement is platform-dependent and outside the bind_cpu contract. On dual-core RISC-V/ARMv7 CI, an unbound thread may legitimately run all iterations on core 0 while T0 delays on the same core, so any assertion requiring cross-core migration (thread_inc != thread_tic, or core_mask with multiple bits) remains flaky. This patch narrows the test to what bind_affinity actually verifies: T0 bound to core 0 must always execute on core 0. Unbound threads only add scheduling pressure; their core_mask is printed for observation but not asserted. Changes: - Sample `rt_hw_cpu_id()` and update counters under `rt_sched_lock` to avoid migration skew between counting and CPU sampling. - Track `thread_core_mask` for diagnostics; assert only T0 binding via `thread_inc[0] == thread_tic[0]` and `thread_core_mask[0] == 1`. - Remove flaky unbound-thread assertions (`thread_inc != thread_tic` and multi-core core_mask checks). - Move T0 assertions to the main test thread after all workers finish. - Use atomic `finsh_flag`; reset counters and `threads[]` in `utest_tc_init`. - Spin in worker loops after `run_num` until cleanup deletes them (do not return from thread_entry or use `RT_WAITING_FOREVER` with `rt_thread_delay`). - Guard `rt_thread_delete` in cleanup with non-NULL checks. Signed-off-by: GuEe-GUI <2991707448@qq.com>

[bsp][zynqmp-a53-dfzu2eg] enable full dm

2026-06-08T06:57:37Z

This is a builtin fdt example BSP Signed-off-by: GuEe-GUI <2991707448@qq.com>

fix(finsh): validate backtrace thread address

2026-05-11T06:09:25Z

修改vdso时间更新触发位置

2026-04-29T22:19:58Z