[adriencaccia]

The v5.7.0 Release workflow failed on two targets that PR CI never built the release way. aarch64-unknown-linux-musl failed to link with undefined reference to __aarch64_ldadd4_sync (job): dist sets RUSTFLAGS for musl targets, which makes cargo ignore the -lgcc in target.aarch64-unknown-linux-musl.rustflags, while musl-build's plain cargo build kept it. aarch64-apple-darwin never started, since GitHub no longer runs the macos-14 runner dist 0.31.0 picks.

PRs now run the release build itself in place of musl-build (pr-run-mode = "upload": artifacts stay on the workflow run, only a tag push publishes). The vendored C code is compiled with -mno-outline-atomics, so the link needs nothing from libgcc; it only changes three attach-time atomics in libbpf, as the ring buffer read path uses plain acquire/release loads. The macOS build runs on macos-latest, as CI does.

	Before	Now
PR check	cargo build of the musl targets (musl-build)	dist build of all 3 release targets
aarch64 musl release	fails: __aarch64_ldadd4_sync undefined	links
macOS release	cancelled: macos-14 not supported	builds on macos-latest

The CI commit alone reproduces the aarch64 failure on this PR (job). With all three commits every target builds (run), and the aarch64 binary is static and runs codspeed --version in Alpine arm64.

Refs COD-4123

[greptile-apps]

Confidence Score: 5/5

[Medium impact] The PR appears safe to merge; no actionable issues were found.

Summary

Replaces explicit libgcc linking with -mno-outline-atomics for AArch64 musl C builds.

• Aarch64 musl C code inlines atomics before linking.

_{Reviews (1) · Last reviewed commit: "build: inline aarch64 atomics in musl C ..." · Reviewed by Greptile}

[codspeed]

Merging this PR will not alter performance

2 warnings

⚠️ Walltime results may be unreliable

Some benchmarks used the Walltime instrument outside CodSpeed Macro Runners, which can make the results inconsistent.

Learn more about Macro Runners

See affected benchmarks

⚠️ 8 benchmarks are dominated by system calls

System calls cannot be simulated consistently and are left out of the measure. The reported value understates the real cost. Switch to the Walltime instrument to measure them.

Learn more about system calls

See affected benchmarks

✅ 53 untouched benchmarks
⏩ 6 skipped benchmarks¹

---

_{Comparing cod-4123-release-valgrind-codspeed-and-runner (c9e6764) with main (34b48f1)}

Footnotes

1. 6 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[GuillaumeLagrange]

lgtm, @not-matthias FYI in case we ever need to troubleshoot this