rm: report traversal failed on getdents EIO by Franklin-Qi · Pull Request #15209 · uutils/coreutils

Franklin-Qi · 2026-10-09T10:06:13Z

When directory listing fails mid-walk (e.g. injected getdents EIO), emit GNU's "traversal failed" even under -f instead of later "Directory not empty".

Closes: #15206

codspeed · 2026-10-10T10:47:37Z

Merging this PR will improve performance by 4.45%

⚠️

4 benchmarks spent significant time in system calls

System calls cannot be consistently instrumented, so they are not included in the measure, which understates the real cost. Please switch to the Walltime instrument to accurately measure system calls.

Measurement and system calls

rm_recursive_tree

rm_single_file

rm_force_files

rm_multiple_files

⚡ 1 improved benchmark
✅ 7 untouched benchmarks
⏩ 464 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	`BASE`	`HEAD`	Efficiency
⚡ ⚠️	Simulation	`rm_recursive_tree`	14.4 ms	13.8 ms	+4.45%

Tip

Curious why performance improved? Comment @codspeedbot explain why performance improved on this PR, or directly use the CodSpeed MCP with your agent.

_{Comparing Franklin-Qi:feature-issue#15206-rm-traversal-failed (166a6a6) with main (175f6e1)²}

464 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. ↩
No successful run was found on main (b8657e8) during the generation of this report, so 175f6e1 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

When directory listing fails mid-walk (e.g. injected getdents EIO), emit GNU's "traversal failed" even under -f instead of later "Directory not empty". Closes: uutils#15206

eduardomourar · 2026-10-17T10:53:25Z

+/// even under `-f`. Without that, rm later fails with "Directory not empty".
+#[test]
+#[cfg(target_os = "linux")]
+#[cfg(not(wasi_runner))]


WASMTIME_LOG can't unblock these tests — it's trace-only, with no fault= equivalent to inject the getdents64 EIO — so #[cfg(not(wasi_runner))] is correct. It's still a useful WASI-side strace substitute for debugging; #15453 wires it into wasi.yml.

sylvestre · 2026-10-17T12:13:08Z

+        .current_dir(scene.fixtures.as_string())
+        .output()
+    else {
+        return; // missing strace


this only skips if strace is missing, if ptrace is blocked (containers) strace fails and the test fails too. could you please probe like test_no_preserve_root_literal_root does?

sylvestre · 2026-10-17T12:13:13Z

+#[test]
+#[cfg(target_os = "linux")]
+#[cfg(not(wasi_runner))]
+fn test_recursive_getdents_eio_subdirectory_reports_traversal_failed() {


this is almost the same code as the test above, could you please dedup with a helper taking the when= and the expected path?

I think we can for loop

oech3 reviewed Oct 9, 2026

View reviewed changes