dd: ignore a failed cache drop while copying

[ChihweiLHBird]

Summary

iflag=nocache and oflag=nocache diagnosed every failed cache drop, so copying from or to a pipe printed failed to discard cache for: 'standard input': Illegal seek once per read and exited 1 even though the copy succeeded. GNU ignores those failures while copying and diagnoses them only with count=0.

Source::discard_cache and Dest::discard_cache now return the error, Input and Output pass it through, and the callers decide: dd_copy ignores it, the count=0 path reports it and sets the exit status.

nocache is accepted only on Linux and Android, so other targets are unaffected; the platform stub moved from Input/Output into Source/Dest, which lets the wrappers be plain pass-throughs.

Part of #14705 (tests/dd/nocache.sh).

Example

dd if=/dev/zero bs=1M count=1 status=none | dd iflag=nocache bs=1M of=/dev/null

Before: one failed to discard cache line per read, exit 1. After: silent, exit 0, as GNU 9.12.

count=0 still diagnoses the failure:

$ dd count=0 status=none | dd iflag=nocache count=0 status=none; echo $?
dd: failed to discard cache for: 'standard input': Illegal seek
1

[codspeed]

Merging this PR will not alter performance

⚠️ 9 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

• dd_copy_4k_blocks
• dd_copy_8k_blocks
• dd_copy_with_seek
• dd_copy_with_skip
• dd_copy_partial
• dd_copy_default
• dd_copy_64k_blocks
• dd_copy_1m_blocks
• dd_copy_separate_blocks

✅ 18 untouched benchmarks
⏩ 454 skipped benchmarks¹

---

_{Comparing ChihweiLHBird:zhiwei/fix-dd-nocache-on-a-pipe (cfd839d) with main (774cca5)²}

Footnotes

1. 454 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. ↩

2. No successful run was found on main (fb68e6c) during the generation of this report, so 774cca5 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

[github-actions]

GNU testsuite comparison:

Congrats! The gnu test tests/dd/nocache is no longer failing!

[ChihweiLHBird]

I looked into the CodSpeed report. The regression is real but small, and it comes from compiler register allocation rather than from any added work.

benchmark	executed instructions vs main
dd_copy_default (512 B blocks)	+0.68%
dd_copy_8k_blocks	+0.04%
dd_copy_with_skip	+0.11%
dd_copy_with_seek	+0.11%

Cause. None of these benchmarks pass nocache to dd, so the changed code never runs in them. As a diagnostic I built the PR with the three nocache blocks deleted from the copy loop; that build matches main's instruction count exactly, so all of the extra instructions come from how the compiler lays out the loop around those blocks. The PR changes what the call inside them does (return the error instead of printing it), which changes the compiler's register plan for the whole loop: the same work is done with about ten more register-to-stack moves per block. The default benchmark copies 65,536 blocks, which is where 0.68% comes from.

I couldn't find a variant of this fix reaches zero instructions increase. Another variant, adding a bunch of quietly_discard_cache functions, would lower +0.68% to +0.52%, but the code structure can look, in my opinion, less readable, that way. Let me know if the performance if prioritized though, and I would be happy to adopt that design.

[sylvestre]

almost the same block twice for input and output, could be a small helper taking the name and the result, no?

[ChihweiLHBird]

I don't think building a helper for them is a good idea. i and o are different type, and the error messages in translate! macro are also different. Building a helper might make it looks much messy than the current shape

[sylvestre]

these wrappers are just pass-throughs now, could the non-linux stub live in Source/Dest::discard_cache instead? that would drop both #[allow]

[ChihweiLHBird]

Moved. But it adds an extra stub inside Source/Dest

[Copilot]

🟡 Changes recommended

GNU-source provenance must be clarified and the changed output-side count=0 path needs regression coverage.

2 open findings

• Add regression test for piped output with nocache count=0 · New
• Revalidate patch without relying on GNU source · New

What changed in this PR

Updates dd cache-discard handling so advisory failures during copying are ignored while count=0 failures remain fatal.

Changes:

• Return cache-discard errors to callers.
• Preserve diagnostics for count=0.
• Add pipe-copy regression coverage.

File	Description
src/​uu/​dd/​src/​dd.rs	Moves cache-error policy to callers.
src/​uu/​dd/​src/​bufferedoutput.rs	Propagates cache-discard results.
tests/​by-util/​test_dd.rs	Tests silent pipe copying with nocache.

🧠 Review effort: Balanced

---

💡 Configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

🟡 Changes recommended

The unsupported-target count=0 failure contradicts the stated unchanged behavior and lacks coverage.

3 open findings

• Preserve count=0 no-op on targets without fadvise · New
• Add regression test for piped output with nocache count=0
• Revalidate patch without relying on GNU source

🧠 Review effort: Balanced

[Copilot]

🟡 Changes recommended

Redirected regular-file stdin is incorrectly classified as non-seekable on Unix targets without fadvise.

1 open finding

• Classify redirected stdin by seekability, not StdinFile variant · New

3 resolved since last review

• Preserve count=0 no-op on targets without fadvise
• Add regression test for piped output with nocache count=0
• Revalidate patch without relying on GNU source

🧠 Review effort: Balanced

[Copilot]

🟢 Approval recommended

The implementation matches the stated behavior and includes focused regression coverage.

0 open findings

1 resolved since last review

• Classify redirected stdin by seekability, not StdinFile variant

🧠 Review effort: Balanced

[Copilot]

🟢 Approval recommended

The focused behavior change is consistent across call paths and has suitable regression coverage.

0 open findings

🧠 Review effort: Balanced

[Copilot]

🟢 Approval recommended

The focused behavior change is correctly propagated and covered across copying and count=0 paths.

0 open findings

🧠 Review effort: Balanced