feat: publish FreeRTOS C FC02 card

2026-07-19 16:36:02 +02:00
commit e3420748ee
208 changed files with 57240 additions and 0 deletions
@@ -0,0 +1,49 @@
+# FC02 card plan
+
+## Position
+
+```text
+C pointers and structures
+→ heap_4 mechanics in C
+→ FC02 raw FreeRTOS task in C
+→ C++ language/OOP prerequisites
+→ K02 Task<Derived> wrapper in FreeRTOS C++
+```
+
+FC02 and K02 are a controlled pair. They use the same kernel, config,
+priorities, stack depths, workload and `E01–E12` meanings. The variable under
+study is the language-facing representation: raw C callback/context versus a
+C++ CRTP wrapper.
+
+## Task01
+
+Two equal-priority CPU-bound workers independently sum `1..20000`. A lower
+priority supervisor checks both results, task-stack evidence and an observed
+`A,B,A` scheduling subsequence.
+
+Enabled viewpoints:
+
+- A1 Context;
+- A2 Structure;
+- A3 Dispatch;
+- A5 Flow;
+- A6 State;
+- A7 Runtime.
+
+A4 is unavailable because application topology is already complete in A2/A5.
+A8 is unavailable because the relevant C idioms are attached directly to
+A2/A3 rather than promoted into a decorative pattern diagram.
+
+## Acceptance
+
+- freestanding C11 build for RV32I/Hazard3;
+- `make check-pair`, context contract, host workload and C-only ABI pass;
+- simulator returns PASS with two results `200010000`;
+- E03 proves `xTaskCreate` argument registers;
+- E06 proves `a0 == &g_worker_a` at callback entry;
+- E08 proves a real `A,B,A` subsequence;
+- context, `&handle`, handle/TCB and separate task-stack range are not
+  conflated;
+- all active UML anchors resolve and every RUN step has an event, strategy and
+  snapshot;
+- navigation alone never mutates the simulator.