wall profiler: track live PCPs via an intrusive list

[szegedi]

Summary

Replaces the unordered_set<PersistentContextPtr*> that the wall profiler uses to track live PersistentContextPtr (PCP) instances with an intrusive doubly-linked list threaded through the PCPs themselves. Same shutdown semantics — ~WallProfiler still deletes any PCP that V8 hasn't reclaimed yet — but with materially less per-PCP overhead.

Stacked on top of #322 (thread-local active profiler) and depends on the thread-local already being in place.

Why

liveContextPtrs_ exists because, at process exit, V8's CPED map can still hold PCPs that V8's late-shutdown finalizers haven't reached yet. LSAN's leak check fires before those finalizers, so without an explicit cleanup pass each surviving PCP gets reported as a leak (we verified this on node:22-bookworm with --experimental-async-context-frame: a baseline build with this PR's liveContextPtrs_ removed and no replacement leaks ~288 bytes across ~8 PCP allocations per run). The set is a fine tool for the cleanup job, but it carries hash-table overhead unrelated to that job.

What changed

• New: each PersistentContextPtr carries a WallProfiler* const profiler_, a PersistentContextPtr** pprev_, and a PersistentContextPtr* next_. The two list pointers form a standard intrusive doubly-linked list rooted at WallProfiler::liveContextPtrHead_.
• ~PersistentContextPtr unlinks itself when pprev_ != nullptr. That predicate also serves as proof that profiler_ is still valid, so the destructor dereferences it directly to decrement the live counter — no thread-local lookup, no ID compare.
• ~WallProfiler walks the list, nulls each PCP's pprev_ (so the soon-to-fire destructor short-circuits the unlink instead of poking at the head pointer in dying memory), then deletes each PCP. ~node::ObjectWrap clears the V8 weak callback during that delete, so the dangling internal-field pointer left behind in the wrap object stays inert even if V8 GCs the wrap later.
• GetMetrics reads a std::atomic<size_t> liveContextPtrCount_ updated alongside the list link/unlink, replacing the old liveContextPtrs_.size(). usedAsyncContextCount / totalAsyncContextCount keep their existing semantics (both report the live count, just like before).

Memory savings

Per live PCP, the tracking cost goes from:

	Before (unordered_set)	After (intrusive list)
Inside the PCP object	8 B (back-pointer to set)	8 B (back-pointer) + 16 B (two list pointers) = 24 B
Outside the PCP object	~40 B (hash node + bucket slot in the set)	0 B
Total	~48 B/PCP	24 B/PCP

A net saving of ~24 bytes per live async context. The set's per-bucket fixed overhead also goes away, though that's negligible compared to the per-element savings on real workloads.

The list operations are O(1) (insert-at-head, unlink-by-pointer), matching the previous set's amortized characteristics.

Verification

Ran the project's Linux ASAN+LSAN suite (npm run test:js-asan) in a node:22-bookworm container with --node-option=experimental-async-context-frame so the CPED-mode tests actually execute (without that flag the tests gate themselves out and the code path under test never runs in CI). Compared baseline (this PR reverted) vs patched:

	Tests	LSAN summaries
Baseline	95 passing / 1 pending	40528 B / 75936 B leaked — all V8/Node internal
Patched	95 passing / 1 pending	40528 B / 79648 B leaked — all V8/Node internal

Zero references to PersistentContextPtr, WallProfiler, dd_pprof.node, or wall.cc in the patched LSAN output. The remaining leaks are pure V8/Node-internal cleanup-on-exit noise, present in identical shape on the baseline (small magnitude differences are noise — V8's final GC state at exit is non-deterministic).

Test plan

• npm run build (macOS Release): clean
• npm run test:js (macOS): 96 passing
• npm run test:js-asan (Linux container, Node 22 + --experimental-async-context-frame): 95 passing, 1 pending, no PCP/WallProfiler/dd_pprof leak reports
• CI passes

Jira: PROF-14791

[github-actions]

Overall package size

Self size: 2.08 MB
Deduped: 2.44 MB
No deduping: 2.44 MB

Dependency sizes

| name | version | self size | total size | |------|---------|-----------|------------| | source-map | 0.7.6 | 185.63 kB | 185.63 kB | | pprof-format | 2.2.1 | 163.06 kB | 163.06 kB | | node-gyp-build | 4.8.4 | 13.86 kB | 13.86 kB |

_{🤖 This report was automatically generated by heaviest-objects-in-the-universe}

[datadog-prod-us1-3]

✨ Fix all issues with BitsAI

⚠️ Warnings

🚦 1 Pipeline job failed

Pull Request Labels | label     

🛟 This job is unlikely to succeed on retry. Please review your pipeline configuration.

Label error. Requires exactly 1 of: semver-patch, semver-minor, semver-major.

Useful? React with 👍 / 👎

_{This comment will be updated automatically if new data arrives.
🔗 Commit SHA: 867c9bc | Docs | Datadog PR Page | Give us feedback!}

[nsavoire]

@codex review

[chatgpt-codex-connector]

Codex Review: Didn't find any major issues. Already looking forward to the next diff.

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[nsavoire]

I think that liveContextPtrCount_ is only accessed by isolate thread and not from a signal handler, thus it could be a plain size_t.

[szegedi]

My only worry is GC callbacks, but I'll try to confirm from source code that they always run on the isolate thread (I know that some V8 GC stages can be multithreaded.)

[nsavoire]

nit: the claim that pprev (double pointer) enables O(1) unlink is dubious. Maybe better explain the benefits vs using a single pointer.