chown, chgrp: decide --from on the file that is changed

[abendrothj]

--from was checked on a stat of the name and the change was then made by name, so a rename landing in between changed a file that never had the required owner.

On Linux, a non-directory operand and every -R entry are now held open with O_PATH when --from is given, checked to be the file that was stat'd and to still pass --from, and changed through that descriptor with fchownat(AT_EMPTY_PATH). A file found replaced is left alone and counts as a failure, which only -v reports, as GNU 9.12 does for operands and under -L/-H. Under -R -P, GNU 9.12 changes the replacement by name; this leaves it alone too, so here it exits 1 where GNU exits 0. Without --from nothing changes.

Unreadable directory operands, -R -H symlink targets, -R -v failure lines, -h with -H/-L, the descriptor walk on other platforms and holding the file without O_PATH are in follow-ups: #15124, #15125, #15126, #15127, #15128, #15129.

Part of #15027

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[codspeed]

Merging this PR will degrade performance by 22.03%

⚠️ 121 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

• cksum_default
• cksum_multiple_files
• du_wide_tree[(5000, 500)]
• du_all_wide_tree[(5000, 500)]
• factor_multiple_u64s[2]
• cksum_raw_output
• cksum_crc
• cksum_crc32b
• wc_bytes_synthetic[1]
• du_deep_tree[(100, 3)]
• …and 111 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]

❌ 2 regressed benchmarks
✅ 399 untouched benchmarks
⏩ 54 skipped benchmarks¹

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
❌	Simulation	three_39_bit_primes	583.2 ms	846.7 ms	-31.12%
❌	Simulation	five_38_bit_primes	1.6 s	1.8 s	-11.75%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

---

_{Comparing abendrothj:chown-from-by-fd (4b43201) with main (6292c03)}

Footnotes

1. 54 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. ↩

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[sylvestre]

this is too large, can we make it smaller?

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[abendrothj]

Split it up. This PR now only has the core change, three files, +247/-32: on Linux, with --from, the operand and each -R entry are held open with O_PATH, checked to be the file that was stat'd and to still pass --from, and changed through that descriptor with fchownat(AT_EMPTY_PATH). The rest is in follow-ups, each based on the one before: unreadable directory operands (#15124), --from on symlink targets under -R -H (#15125), -v failure lines for -R entries (#15126), -h with -H/-L (#15127), the descriptor walk on the other platforms (#15128), and holding the file without O_PATH (#15129). Until the earlier ones merge, each diff also shows the commits it's based on.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.