procctl(PROC_REAP_KILL): align error reporting with pgkill(2)
ClosedPublic
Actions

Authored by kib on Feb 18 2026, 8:26 AM.

Details

Reviewers

markj
olce

Commits

rG5c8af592f69a: procctl(PROC_REAP_KILL): align error reporting with pgkill(2)

Summary

ESRCH from cr_cansignal() means that the target process should be not
visible to the initiator.  So do not report pids with ESRCH result.

Noted by:       markj

Diff Detail

Repository

rG FreeBSD src repository

Lint

Lint Skipped

Unit

Tests Skipped

Event Timeline

kib created this revision.Feb 18 2026, 8:26 AM

Herald added subscribers: olce, imp. · View Herald TranscriptFeb 18 2026, 8:26 AM

kib requested review of this revision.Feb 18 2026, 8:26 AM

kib mentioned this in D55288: procctl(PROC_REAP_KILL): use pgrp pg_killsx sx to sync with fork.

olce added inline comments.Feb 18 2026, 12:34 PM

sys/kern/kern_procctl.c
270–274	Looks wrong. If `ESRCH` is returned, this code will still proceed to `pksignal()`, whereas the purpose of `cr_cansignal()` returning `ESRCH` is exactly to avoid the target being signaled.

Return on any error, not excluding ESRCH.

Thank you. It looks right, I would add a short comment like /* Hide ESRCH errors to ensure that this function cannot be used as an oracle for process visibility. */

This revision is now accepted and ready to land.Feb 18 2026, 2:08 PM

Add comment.

This revision now requires review to proceed.Feb 18 2026, 2:20 PM

markj accepted this revision.Feb 18 2026, 2:21 PM

This revision is now accepted and ready to land.Feb 18 2026, 2:21 PM

olce accepted this revision.Feb 18 2026, 3:54 PM

Closed by commit rG5c8af592f69a: procctl(PROC_REAP_KILL): align error reporting with pgkill(2) (authored by kib). · Explain WhyFeb 21 2026, 10:08 AM

This revision was automatically updated to reflect the committed changes.

kib added a commit: rG5c8af592f69a: procctl(PROC_REAP_KILL): align error reporting with pgkill(2).