cat: resize the output pipe once per invocation by limadog9 · Pull Request #15506 · uutils/coreutils

limadog9 · 2026-10-18T19:29:42Z

Prepare stdout's pipe capacity once and reuse the broker copy for every input. On 30 files, resize attempts fall from 31 to 2. A strace regression checks one broker pipe and two resize attempts with cat /dev/null /dev/ null.

Closes #11700.

Sign in to view


 [target.'cfg(unix)'.dependencies]
-rustix = { workspace = true, features = ["fs"] }
+rustix = { workspace = true, features = ["fs", "pipe"] }


Sign in to view

+    // Increasing the output pipe capacity is optional and may be done once by
+    // callers that copy several inputs into the same output.
+    let _ = fcntl_setpipe_size(&mut *dest, MAX_ROOTLESS_PIPE_SIZE);
+    splice_unbounded_broker(source, dest)
+}
+
+/// Copy through a cached broker pipe without changing the output pipe capacity.
+///
+/// Callers copying multiple sources may prepare their output once and reuse
+/// this operation for each source. A failure to resize never prevents copying.
+#[inline]
+pub fn splice_unbounded_broker(source: &impl AsFd, dest: &mut impl AsFd) -> SpliceRes {


Sign in to view

-    // fcntl for input would not improve throughput since
-    // - sender with splice probably increased size already
-    // - sender without splice is bottleneck


oech3 · 2026-10-18T20:38:49Z

+        // All operands share stdout. Resize its pipe once, even if the kernel
+        // refuses the optional increase, rather than retrying for every file.


Suggested change

// All operands share stdout. Resize its pipe once, even if the kernel

// refuses the optional increase, rather than retrying for every file.

// All operands share stdout. Try to resize it once.

sylvestre · 2026-10-18T21:44:41Z

    I: IntoIterator<Item = &'a OsString>,
 {
+    #[cfg(any(target_os = "linux", target_os = "android"))]
+    if options.can_print_fast() {


the description mentions a regression test for mixed file/stdin order but i don't see it in the commit, please add it in tests/by-util/test_cat.rs

I don't think this PR needs a test since it is performance-related PR.

If we test something, it should count pipe2 call by strace.
a regression checks mixed file/stdin output order is unrelated with this PR.

sylvestre · 2026-10-18T21:44:47Z

+/// Copy through a cached broker pipe without changing the output pipe capacity.
+///
+/// Callers copying multiple sources may prepare their output once and reuse
+/// this operation for each source. A failure to resize never prevents copying.


this function never resizes, so "a failure to resize never prevents copying" is confusing here. could you please make this doc one line?

I think we don't need this comment.

codspeed · 2026-10-19T08:31:05Z

Merging this PR will not alter performance

⚠️

119 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

du_all_wide_tree[(5000, 500)]

du_wide_tree[(5000, 500)]

cksum_multiple_files

du_deep_tree[(100, 3)]

cksum_sysv

cksum_raw_output

df_deep_directory

dd_copy_partial

cksum_crc

cksum_crc32b

…and 109 more

ℹ️ 1 benchmark has no flame graph

The flamegraph could not be generated because too much time was spent in runtime internals, making the visualization not useful for performance analysis.

sort_dictionary_order[500000]

✅ 379 untouched benchmarks
⏩ 92 skipped benchmarks¹

_{Comparing limadog9:fix/cat-resize-output-pipe-once-11700 (74bd78d) with main (d2d9694)}

92 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. ↩

limadog9 · 2026-10-19T10:14:41Z

Added the requested strace regression

oech3 · 2026-10-19T10:28:47Z

 }
+
+#[test]
+fn test_fast_copy_mixed_file_and_stdin_order() {


Would you remove unrelated test and update 1st comment to mention strace test instead?

oech3 · 2026-10-19T10:30:07Z

+
+    let scene = TestScenario::new(util_name!());
+    let probe = Command::new("strace")
+        .args(["-qq", "-o", "/dev/null", "-e", "trace=none"])


I think we can check support of syscalls by kernel at here instead of trace=none.

oech3 · 2026-10-19T10:30:58Z

+    let probe = Command::new("strace")
+        .args(["-qq", "-o", "/dev/null", "-e", "trace=none"])
+        .arg(&scene.bin_path)
+        .args(["cat", "--help"])


I think --version is faster, or cat /dev/null is making a sense.

Moreover, we don't need 2 strace at here if we used cat /dev/null /dev/null.

Prepare stdout pipe capacity once and reuse the broker copy for every input. On 30 files, resize attempts fall from 31 to 2. Add a strace regression using cat /dev/null /dev/null to check one broker pipe and two resize attempts. Detect unavailable syscalls from the same trace rather than launching a separate probe. Closes uutils#11700.

oech3 reviewed Oct 18, 2026

View reviewed changes

Comment thread src/uu/cat/Cargo.toml

[target.'cfg(unix)'.dependencies]

rustix = { workspace = true, features = ["fs"] }

rustix = { workspace = true, features = ["fs", "pipe"] }

This comment was marked as low quality.

Sign in to view

This comment was marked as low quality.

Sign in to view

oech3 reviewed Oct 18, 2026

View reviewed changes

This comment was marked as low quality.

Sign in to view

oech3 reviewed Oct 18, 2026

View reviewed changes

Comment thread src/uucore/src/lib/features/pipes.rs

Comment on lines -112 to -114

// fcntl for input would not improve throughput since

// - sender with splice probably increased size already

// - sender without splice is bottleneck

This comment was marked as resolved.

Sign in to view

limadog9 force-pushed the fix/cat-resize-output-pipe-once-11700 branch from c4d75cf to 34525a4 Compare October 18, 2026 19:57

oech3 reviewed Oct 18, 2026

View reviewed changes

limadog9 force-pushed the fix/cat-resize-output-pipe-once-11700 branch from 34525a4 to c7f35d0 Compare October 18, 2026 20:40

sylvestre reviewed Oct 18, 2026

View reviewed changes

limadog9 force-pushed the fix/cat-resize-output-pipe-once-11700 branch from c7f35d0 to 736f4f1 Compare October 18, 2026 22:29

limadog9 force-pushed the fix/cat-resize-output-pipe-once-11700 branch from 736f4f1 to adfcce3 Compare October 19, 2026 10:06

oech3 reviewed Oct 19, 2026

View reviewed changes

limadog9 force-pushed the fix/cat-resize-output-pipe-once-11700 branch from adfcce3 to 74bd78d Compare October 19, 2026 10:43

		// All operands share stdout. Resize its pipe once, even if the kernel
		// refuses the optional increase, rather than retrying for every file.

	// All operands share stdout. Resize its pipe once, even if the kernel
	// refuses the optional increase, rather than retrying for every file.
	// All operands share stdout. Try to resize it once.

Conversation

limadog9 commented Oct 18, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

This comment was marked as low quality.

Uh oh!

This comment was marked as low quality.

Uh oh!

This comment was marked as low quality.

Uh oh!

This comment was marked as low quality.

This comment was marked as resolved.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

oech3 Oct 18, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

codspeed Bot commented Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Merging this PR will not alter performance

Footnotes

Uh oh!

limadog9 commented Oct 19, 2026

Uh oh!

oech3 Oct 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

This branch has not been deployed

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

limadog9 commented Oct 18, 2026 •

edited

Loading

oech3 Oct 18, 2026 •

edited

Loading

codspeed Bot commented Oct 19, 2026 •

edited

Loading

oech3 Oct 19, 2026 •

edited

Loading