Stuart Henderson <s...@spacehopper.org> wrote:
> On 2018/09/26 19:16, Stuart Henderson wrote:
> > N.B. "Ports using KERN_CPTIME2 will need to be updated."
> >
> > This is likely to cause a bunch of breakage in things reporting CPU stats
> > and time is very short to fix them before release.
> >
> > If you look after ports that do this, get onto a kernel with this change
> > (very new commit so it probably won't be in snaps quite just yet) and
> > test ASAP.
>
> Some starting points for investigation,
>
> assorted mozillas
> collectd
> conky
> go
> htop
> libgtop2
> net-snmp
> node
> pgtop
> py-psutil
> libuv (+ embedded copies, at least in cmake, maybe more)
> zabbix
>
> >
> >
> >
> > ----- Forwarded message from Scott Soule Cheloha <chel...@openbsd.org> -----
> >
> > From: Scott Soule Cheloha <chel...@openbsd.org>
> > Date: Wed, 26 Sep 2018 11:23:13 -0600 (MDT)
> > To: source-chan...@openbsd.org
> > Subject: CVS: cvs.openbsd.org: src
> >
> > CVSROOT: /cvs
> > Module name: src
> > Changes by: chel...@cvs.openbsd.org 2018/09/26 11:23:13
> >
> > Modified files:
> > sys/kern : kern_sched.c kern_sysctl.c
> > sys/sys : sched.h
> > usr.bin/systat : cpu.c vmstat.c
> > usr.bin/top : display.c display.h machine.c machine.h top.c
> >
> > Log message:
> > KERN_CPTIME2: set ENODEV if the CPU is offline.
> >
> > This lets userspace distinguish between idle CPUs and those that are
> > not schedulable because hw.smt=0.
> >
> > A subsequent commit probably needs to add documentation for this
> > to sysctl.2 (and perhaps elsewhere) after the dust settles.
> >
> > Also included here are changes to systat(1) and top(1) that account
> > for the ENODEV case and adjust behavior accordingly:
> >
> > - systat(1)'s cpu view prints placeholder marks ('-') instead of
> > percentages for each state if the given CPU is offline.
> >
> > - systat(1)'s vmstat view checks for offline CPUs when computing the
> > machine state total and excludes them, so the CPU usage graph
> > only represents the states for online CPUs.
> >
> > - top(1) does not draw CPU rows for offline CPUs when the view is
> > redrawn. If CPUs "go offline", percentages for each state are
> > replaced by placeholder marks ('-'); the view will need to be
> > redrawn to remove these rows. If CPUs "go online" the view will
> > need to be redrawn to show these new CPUs. In "combined CPU" mode,
> > the count and the state totals only represent online CPUs.
> >
> > Ports using KERN_CPTIME2 will need to be updated. The changes
> > described above to make systat(1) and top(1) aware of the ENODEV
> > case *and* gracefully handle a changing HW_NCPUONLINE while the
> > application is running are not necessarily appropriate for each
> > and every port.
> >
> > The changes described above are so extensive in part to demonstrate
> > one way a program *might* be made robust to changing CPU availability.
> > In particular, changing hw.smt after boot is an extremely rare event,
> > and this needs to be weighed when updating ports.
> >
> > The logic needed to account for the KERN_CPTIME2 ENODEV case is
> > very roughly:
> >
> > if (sysctl(...) == -1) {
> > if (errno != ENODEV) {
> > /* Actual error occurred. */
> > } else {
> > /* CPU is offline. */
> > }
> > } else {
> > /* CPU is online and CPU states were set by sysctl(2). */
> > }
> >
> > Prompted by deraadt@. Basic idea for ENODEV from kettenis@. Discussed at
> > length with kettenis@. Additional testing by tb@.
> >
> > No complaints from hackers@ after a week.
> >
> > ok kettenis@, "I think you should commit [now]" deraadt@
> >
> >
> > ----- End forwarded message -----
> >
conky hw.smt=0 ERROR :
conky
Conky: desktop window (1c00010) is subwindow of root window (e1)
Conky: window type - desktop
Conky: drawing to created window (0x1e00001)
Conky: drawing to single buffer
Conky: sysctl kern.cp_time2 failed
Conky: sysctl kern.cp_time2 failed
Conky: sysctl kern.cp_time2 failed
Conky: sysctl kern.cp_time2 failed
conky hw.smt=1 OK
htop hw.smt=0 OK
htop hw.smt=1 OK
top hw.smt=0 OK
top hw.smt=1 OK
zabbix: hw.smt=0 OK
zabbix: hw.smt=1 OK
tested with zabbix_get -s localhost -k system.cpu.util
xstatbar: hw.smt=0 ERROR
xstatbar: sysinfo update: KERN.CPTIME2.4 failed: Operation not supported by
device
xstatbar: hw.smt=1 OK
collectd: hw.smt=0 ERROR with plugin cpu (using write_log)
Oct 3 12:08:50 solene collectd[37866]: plugin_load: plugin "cpu" successfully
loaded.
Oct 3 12:08:50 solene collectd[89100]: cpu plugin: sysctl failed: Operation
not supported by device.
Oct 3 12:08:50 solene collectd[89100]: read-function of plugin `cpu' failed.
Will suspend it for 20.000 seconds.
collectd: hw.smt=1 OK (using write_log)
Oct 3 12:09:31 solene collectd[50994]: plugin_load: plugin "cpu" successfully
loaded.
Oct 3 12:09:41 solene collectd[82167]: write_log values:
localhost.cpu-0.cpu-user 11768 1538561381\^M
Oct 3 12:09:41 solene collectd[82167]: write_log values:
localhost.cpu-1.cpu-user 25592 1538561381\^M
Oct 3 12:09:41 solene collectd[82167]: write_log values:
localhost.cpu-2.cpu-user 21417 1538561381\^M
Oct 3 12:09:41 solene collectd[82167]: write_log values:
localhost.cpu-5.cpu-user 884 1538561381\^M
Oct 3 12:09:41 solene collectd[82167]: write_log values:
localhost.cpu-4.cpu-user 427 1538561381\^M
Oct 3 12:09:41 solene collectd[82167]: write_log values:
localhost.cpu-3.cpu-user 15805 1538561381\^M
Oct 3 12:09:41 solene collectd[82167]: write_log values:
localhost.cpu-0.cpu-system 11650 1538561381\^M
Oct 3 12:09:41 solene collectd[82167]: write_log values:
localhost.cpu-7.cpu-user 1001 1538561381\^M
thunderbird: hw.smt=0 OK
thunderbird: hw.smt=1 OK
firefox: hw.smt=0 OK
firefox: hw.smt=1 OK
