Enable ThreadSanitizer across the entire multi-threaded JIT pipeline

[jserv]

ThreadSanitizer (TSAN) can now detect race conditions across the entire multi-threaded JIT pipeline with full 4GB address space emulation. This enables testing of the tier-2 LLVM compilation thread while maintaining production memory layout.

Memory Layout (TSAN-compatible):

• Main memory: MAP_FIXED at 0x7d0000000000 (4GB)
• JIT buffer: MAP_FIXED at 0x7d1000000000
• Both allocations within TSAN app range (0x7cf-0x7ff trillion)
• Prevents conflicts with TSAN shadow memory (0x02a-0x7ce trillion)

ASLR Mitigation:

• Added setarch -R wrapper for TSAN test execution
• Disables ASLR to prevent random allocations in shadow memory
• Only affects test runs, not production builds

SDL Conflict Resolution:

• SDL (uninstrumented system library) creates threads TSAN cannot track
• Disabled SDL when TSAN enabled to focus on built-in race detection
• Production builds still fully support SDL

---

Summary by cubic

Enables ThreadSanitizer across the multi-threaded JIT pipeline with FULL4G memory emulation, including tier-2 compilation thread support. Uses fixed mappings and ASLR-disabled test runs to avoid TSAN shadow conflicts while keeping the production layout.

• New Features

  • TSAN build mode with -fsanitize=thread; tests run via setarch -R; SDL and LTO disabled only when TSAN is enabled.
  • Fixed x86-64 mappings for FULL4G memory (0x7d0000000000) and JIT buffer (0x7d1000000000) using MAP_FIXED within TSAN’s app range.
  • TSAN runtime defaults added to reduce noise and keep tests running after reports.

• Bug Fixes

  • Correct race handling in T2C: atomic load/store for hot2, compiled, n_invoke; acquire/release semantics around func/hot2.
  • Background compiler uses a condition variable and cache key lookup to avoid busy waits and use-after-free; signals on enqueue and shutdown with SATP verification.

[jserv]

Benchmarks

Benchmark suite	Current: 2497867	Previous: 2149996	Ratio
Dhrystone	1328 Average DMIPS over 10 runs	1344 Average DMIPS over 10 runs	1.01
Coremark	961.712 Average iterations/sec over 10 runs	962.342 Average iterations/sec over 10 runs	1.00

This comment was automatically generated by workflow using github-action-benchmark.

[cubic-dev-ai]

1 issue found across 8 files

Prompt for AI agents (all 1 issues)

Understand the root cause of the following 1 issues and fix them.


<file name="src/jit.c">

<violation number="1" location="src/jit.c:2350">
When TSAN is enabled on macOS x86_64, this mmap call loses the MAP_JIT flag that the standard path uses, so hardened macOS failures return MAP_FAILED and the JIT never initializes. Please keep MAP_JIT on macOS even in the TSAN path.</violation>
</file>

_{React with 👍 or 👎 to teach cubic. Mention @cubic-dev-ai to give feedback, ask questions, or re-run the review.}

[cubic-dev-ai]

When TSAN is enabled on macOS x86_64, this mmap call loses the MAP_JIT flag that the standard path uses, so hardened macOS failures return MAP_FAILED and the JIT never initializes. Please keep MAP_JIT on macOS even in the TSAN path.

Prompt for AI agents

Address the following comment on src/jit.c at line 2350:

<comment>When TSAN is enabled on macOS x86_64, this mmap call loses the MAP_JIT flag that the standard path uses, so hardened macOS failures return MAP_FAILED and the JIT never initializes. Please keep MAP_JIT on macOS even in the TSAN path.</comment>

<file context>
@@ -2336,6 +2336,25 @@ struct jit_state *jit_state_init(size_t size)
+     * 0x7ffffffff000) and prevents overlap with main memory or TSAN shadow.
+     */
+    void *jit_addr = (void *) 0x7d1000000000UL;
+    state-&gt;buf = mmap(jit_addr, size, PROT_READ | PROT_WRITE | PROT_EXEC,
+                      MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED, -1, 0);
+    if (state-&gt;buf == MAP_FAILED) {
</file context>

✅ Addressed in f915bc2