[adriencaccia]

The v5.7.0 Release workflow failed to link codspeed for aarch64-unknown-linux-musl with undefined reference to __aarch64_ldadd4_sync (job). dist always sets RUSTFLAGS for musl targets, which makes cargo ignore the -lgcc in target.aarch64-unknown-linux-musl.rustflags, and since #531 the runner links libbpf, whose C code calls gcc's outline-atomic helpers. CI's musl-build job runs plain cargo build, kept the flag, and passed.

This compiles the C code with -mno-outline-atomics instead, so the atomics are inlined and the link needs nothing from libgcc.

aarch64 musl build	Before	Now
cargo build (CI musl-build)	links (-lgcc from config)	links
dist build (Release, RUSTFLAGS set)	fails: __aarch64_ldadd4_sync undefined	links

Checked by replaying the dist build in an ubuntu:26.04 arm64 container (gcc 15.2, --profile dist, dist's RUSTFLAGS): main fails with the same error, this branch links a static binary that runs.

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 (df1e3ad) 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