linux-kernel - [2.6.28-rc5] TSC sync error and high hrtimer_start (was: Bootup time regression from 2.6.27 to 2.6.28-rc3+)

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <200811162027.44988.elendil@planet.nl>
Date:	Sun, 16 Nov 2008 20:27:43 +0100
From:	Frans Pop <elendil@...net.nl>
To:	Thomas Gleixner <tglx@...utronix.de>,
	Arjan van de Ven <arjan@...radead.org>
Cc:	Lukas Hejtmanek <xhejtman@....muni.cz>,
	"Rafael J. Wysocki" <rjw@...k.pl>,
	Marcin Slusarz <marcin.slusarz@...il.com>,
	linux-kernel@...r.kernel.org, torvalds@...ux-foundation.org,
	corsac@...ian.org
Subject: [2.6.28-rc5] TSC sync error and high hrtimer_start (was: Bootup time regression from 2.6.27 to 2.6.28-rc3+)

On Friday 14 November 2008, Frans Pop wrote:
> Could either of you maybe give a status update on this patch set and
> the remaining issues with it that were reported (especially the high C0
> reported by powertop)?

My previous tests were on my laptop. This is for my desktop running -rc5.
I had not tested my desktop with earlier .28 releases.

For .28 powertop prominently shows hrtimer_start (tick_sched_timer) which
was not present for .27. The list below that is virtually identical to .27.

Total wakeups for .28 is lower though. Something to look into or not?

Possibly this may be related to this diff in the boot messages:
-checking TSC synchronization [CPU#0 -> CPU#1]: passed.
+checking TSC synchronization [CPU#0 -> CPU#1]:
+Measured 696 cycles TSC warp between CPUs, turning off TSC clock.
+Marking TSC unstable due to check_tsc_sync_source failed

With .27.4 TSC _always_ passed. Now, with .28-rc5 I get this warp. I've
done two boots so far, for the other one the warp was 680 cycles.

Cheers,
FJP

Powertop output for idle system:
*** 2.6.27.4 ***
Wakeups-from-idle per second : 68.7     interval: 10.0s
no ACPI power usage estimate available

Top causes for wakeups:
  31.5% ( 17.8)     <kernel core> : __mod_timer (neigh_periodic_timer)
  14.5% (  8.2)   <kernel module> : mod_timer (rh_timer_func)
  11.3% (  6.4)             artsd : __mod_timer (process_timeout)
   9.2% (  5.2)        ksysguardd : __mod_timer (process_timeout)
   4.2% (  2.4)            kicker : __mod_timer (process_timeout)
   3.5% (  2.0)     <kernel core> : add_timer_on (clocksource_watchdog)

*** 2.6.28-rc5 ***
Wakeups-from-idle per second : 55.6     interval: 10.0s
no ACPI power usage estimate available

Top causes for wakeups:
  28.5% ( 23.4)     <kernel core> : hrtimer_start (tick_sched_timer)
  21.7% ( 17.8)     <kernel core> : __mod_timer (neigh_periodic_timer)
   9.7% (  8.0)   <kernel module> : mod_timer (rh_timer_func)
   7.8% (  6.4)             artsd : schedule_hrtimeout_range (hrtimer_wakeup)
   5.4% (  4.4)        ksysguardd : __mod_timer (process_timeout)
   2.8% (  2.3)            kicker : schedule_hrtimeout_range (hrtimer_wakeup)
   2.4% (  2.0)     <kernel core> : add_timer_on (clocksource_watchdog)
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/