---
- branch: netbsd-8
  date: Sun Apr  1 08:45:43 UTC 2018
  files:
  - new: 1.206.6.2
    old: 1.206.6.1
    path: src/sys/kern/kern_proc.c
    pathrev: src/sys/kern/kern_proc.c@1.206.6.2
    type: modified
  id: 20180401T084543Z.f3fa6036e75726ad67dc7830ed024056e164ebba
  log: "Pull up following revision(s) (requested by kamil in ticket #679):\n\n\tsys/kern/kern_proc.c:
    revision 1.211\n\nMake sysctl_doeproc() more predictable\n\nSwap the order of
    looking into zombie and all process lists, start now\nwith the zombie one. This
    prevents a race observed previously that the\nsame process could be detected on
    both lists during a single polling call.\n\nWhile there:\n - Short-circuit break
    for KERN_PROC_PID, once a pid has been detected.\n - Removal of redundant \"if
    (kbuf)\" and \"if (marker)\" checks.\n - Update of comments regarding potential
    optimization, explaining why we\n   don't want to it as of now. Performance gain
    from lookup call vs\n   iteration over a list is neglible on a regular system.\n
    - Return ESRCH when no results have been found. This allows more easily\n   to
    implement a retry or abandon algorithm.\n\nThis corrects races observed in the
    existing ATF ptrace(2) tests, related\nto await_zombie(). This function was expecting
    to check whether a process\nhas been transformed into a zombie, however it was
    causing occasional\ncrashes as it was overflowing the return buffer, returning
    the same pid\ntwice: once from allproc list and the second time from zombieproc
    one.\n\nFix suggested by <christos>\nShort-circuit break suggested by <kre>\n\nDiscussed
    on tech-kern.\n\nSponsored by <The NetBSD Foundation>\n"
  module: src
  subject: 'CVS commit: [netbsd-8] src/sys/kern'
  unixtime: '1522572343'
  user: martin