Collapse ThreadPool::init dead generic and fix cpuid cfg gate

[Jarred-Sumner]

ThreadPool::init dead generic. bundler::ThreadPool::init (and its helper init_with_pool) were generic over V2 because, during the phased Zig→Rust port, bundle_v2.rs carried two BundleV2 definitions — the canonical one plus a bv2_impl draft module — and both needed to call init. The PORT NOTE on the function explicitly said it should collapse to &BundleV2<'_> once bv2_impl was dropped. That module is gone (grep -rn "struct BundleV2" src/bundler/ finds exactly one definition, bundle_v2.rs:75), so this collapses the generic, names the concrete type, and drops the now-stale PORT NOTE. The body was already storing a type-erased raw pointer, so the monomorphised code is identical.

cpuid cfg-gate mismatch. perf/hw_timer.rs defines struct CpuidResult and fn cpuid() under #[cfg(target_arch = "x86_64")], but their only callers live inside a #[cfg(not(any(target_os = "macos", target_os = "freebsd")))] block (macOS/FreeBSD read the boot-time TSC frequency from sysctl instead of probing CPUID). On x86_64-apple-darwin and x86_64-unknown-freebsd the helpers were therefore compiled with no callers, producing never constructed / never used warnings. The cfg gates now mirror the callers' conditions.

Verified with cargo check --workspace --keep-going (clean), bun run rust:check-all (10/10 targets), and cargo fmt -p bun_bundler -p bun_perf --check. Confirmed the dead-code warnings on x86_64-apple-darwin are present without this change and gone with it. File set is disjoint from #30879.

[robobun]

^{Updated 5:18 AM PT - May 16th, 2026}

❌ @Jarred-Sumner, your commit 7c46dda has 1 failures in Build #55204 (All Failures):

• test/js/node/http/node-http-backpressure.test.ts - SIGKILL on 🐧 13 aarch64

---

🧪   To try this PR locally:

bunx bun-pr 30886

That installs a local version of the PR into your bun-30886 executable, so you can run:

bun-30886 --bun

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro

Run ID: 3d378d77-55f2-40e3-800d-ba3c3861a97a

📥 Commits

Reviewing files that changed from the base of the PR and between e750984 and 7c46dda.

📒 Files selected for processing (2)

• src/bundler/ThreadPool.rs
• src/perf/hw_timer.rs

---

Walkthrough

This PR specializes two generic or platform-agnostic APIs: ThreadPool's initialization functions are changed to accept BundleV2 directly instead of a generic type parameter, with explicit pointer casting for the backreference; and CpuidResult's conditional compilation is narrowed to exclude macOS and FreeBSD platforms.

Changes

Type refinements and platform targeting

Layer / File(s)	Summary
ThreadPool specialization to BundleV2 src/bundler/ThreadPool.rs	ThreadPool::init and ThreadPool::init_with_pool are specialized to accept &BundleV2<'_> directly instead of generic V2 parameter. The v2 backreference assignment is updated with explicit casting from &BundleV2<'_> to &'static BundleV2 using std::ptr::from_ref(v2).cast::<BundleV2<'static>>().
CPUID platform restrictions src/perf/hw_timer.rs	CpuidResult conditional compilation guard is narrowed from x86_64 to x86_64 and not(any(target_os = "macos", target_os = "freebsd")), excluding those operating systems from defining the type.

🚥 Pre-merge checks | ✅ 4

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately describes both main changes: collapsing a dead generic in ThreadPool::init and fixing the cpuid cfg gate in hw_timer.rs.
Description check	✅ Passed	The description covers the required template sections with clear explanations of changes and verification methods, though not explicitly labeled per template structure.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[claude]

LGTM — mechanical de-genericization and cfg-gate tightening, both compile-time-only with no runtime behavior change.

Extended reasoning...

Overview

Two unrelated cleanup changes:

• src/bundler/ThreadPool.rs: collapses ThreadPool::init<V2> and init_with_pool<V2> from generic to concrete &BundleV2<'_>, drops the now-stale PORT NOTE that explicitly said to do this once the duplicate BundleV2 definition was gone, and rewrites the pointer cast to name the concrete target type.
• src/perf/hw_timer.rs: narrows the #[cfg] gate on CpuidResult / cpuid() from target_arch = "x86_64" to additionally exclude macOS/FreeBSD, matching the gate on their only call site inside read_frequency().

Verification

• Confirmed only one struct BundleV2 definition exists (bundle_v2.rs:75). The bv2_impl module still exists but only as a method-body container, not a second struct definition — which is what the PORT NOTE actually depended on.
• The sole external call site (bundle_v2.rs:2934) passes &mut *this where this: &mut BundleV2, so de-genericization type-checks with no caller changes.
• The from_ref::<V2>(v2).cast() → from_ref(v2).cast::<BundleV2<'static>>() rewrite is semantically identical: same source pointer, same target type (the field is *const BundleV2<'static>), just the turbofish moved from inference-at-source to explicit-at-cast.
• In hw_timer.rs, the only references to cpuid / CpuidResult are inside the #[cfg(not(any(target_os = "macos", target_os = "freebsd")))] arm of read_frequency(), so the new gate exactly mirrors callers and only removes dead code on x86_64 Darwin/FreeBSD.

Security risks

None. No new logic, no new unsafe, no I/O or auth surface. Both changes are type/cfg-level and produce identical or strictly-subset codegen.

Level of scrutiny

Low. This is the kind of stale-port-scaffolding cleanup the codebase's own PORT NOTE asked for, plus a dead-code-warning fix. PR description reports clean cargo check --workspace and rust:check-all across 10 targets.

Other factors

No bugs from the bug-hunting system, no outstanding reviewer comments, no prior reviews.