---
- branch: netbsd-8
  date: Fri Dec  8 06:05:15 UTC 2017
  files:
  - new: 1.97.8.1
    old: '1.97'
    path: src/sys/arch/sparc/include/cpu.h
    pathrev: src/sys/arch/sparc/include/cpu.h@1.97.8.1
    type: modified
  - new: 1.249.6.1
    old: '1.249'
    path: src/sys/arch/sparc/sparc/cpu.c
    pathrev: src/sys/arch/sparc/sparc/cpu.c@1.249.6.1
    type: modified
  - new: 1.118.22.1
    old: '1.118'
    path: src/sys/arch/sparc/sparc/intr.c
    pathrev: src/sys/arch/sparc/sparc/intr.c@1.118.22.1
    type: modified
  id: 20171208T060515Z.85da10aa98c8c40e7321c0a334dc7d519b1d8ade
  log: "Pull up following revision(s) (requested by macallan in ticket #429):\n\tsys/arch/sparc/sparc/cpu.c:
    revision 1.250\n\tsys/arch/sparc/include/cpu.h: revision 1.99\n\tsys/arch/sparc/sparc/intr.c:
    revision 1.119\n- return early in xcall() if the function is sparc_noop() instead
    of triggering\n  the IPI and then ignoring responses ( or lack thereof )\n- write
    the .tag field last to avoid a race when polling for an incoming\n  IPI\n- add
    event counters for IPIs being caught with the mutex not held, and for\n  messages
    that are already marked as completed\nWith this my SS20 made it through 48 hours
    of pkgsrc with MAKE_JOBS=3 and a\npair of SM81s.\nHypersparcs still crash but
    instead of craziness we get actual error messages,\napparently one CPU will occasionally
    do a watchdog reset, which according to\nthe manual is caused by catching a trap
    with traps disabled. Now to figure\nout how that can even happen...\n"
  module: src
  subject: 'CVS commit: [netbsd-8] src/sys/arch/sparc'
  unixtime: '1512713115'
  user: msaitoh