From: Rik van Riel <riel@redhat.com>
To: Wanpeng Li <kernellwp@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>,
Ingo Molnar <mingo@kernel.org>,
LKML <linux-kernel@vger.kernel.org>,
Paolo Bonzini <pbonzini@redhat.com>,
Peter Zijlstra <peterz@infradead.org>,
Wanpeng Li <wanpeng.li@hotmail.com>,
Thomas Gleixner <tglx@linutronix.de>,
Radim Krcmar <rkrcmar@redhat.com>, Mike Galbraith <efault@gmx.de>
Subject: Re: [PATCH] time,virt: resync steal time when guest & host lose sync
Date: Thu, 11 Aug 2016 22:44:52 -0400 [thread overview]
Message-ID: <1470969892.13905.120.camel@redhat.com> (raw)
In-Reply-To: <CANRm+Cye=ipm5XwhBV-AxjCMZRQTb3ahuYEa6KXBMe_HzxPa2Q@mail.gmail.com>
[-- Attachment #1: Type: text/plain, Size: 1880 bytes --]
On Thu, 2016-08-11 at 18:11 +0800, Wanpeng Li wrote:
> 2016-08-11 0:52 GMT+08:00 Rik van Riel <riel@redhat.com>:
> > On Wed, 10 Aug 2016 07:39:08 +0800
> > Wanpeng Li <kernellwp@gmail.com> wrote:
> >
> > > The regression is caused by your commit "sched,time: Count
> > > actually
> > > elapsed irq & softirq time".
> >
> > Wanpeng, does this patch fix your issue?
>
> I test this against kvm guest (nohz_full, four vCPUs running on one
> pCPU, four cpuhog processes running on four vCPUs).
> before this fix patch:
> vCPU0's st is 100%, other vCPUs' st are ~75%.
> after this fix patch:
> all vCPUs' st are ~85%.
> However, w/o commit "sched,time: Count actually elapsed irq & softirq
> time", all vCPUs' st are ~75%.
If you pass ULONG_MAX as the maxtime argument to
steal_account_process_time(), does the steal time
get accounted properly at 75%?
If that is the case, I have a hypothesis:
1) The guest is running so much slower when sharing
a CPU 4 ways, that it is accounting only ~90% of
wall clock time as CPU time, due to missing the
other 10% or so of clock ticks.
2) account_process_tick() only ever processes one tick
at a time - if it gets called only 90x a second for
a 100Hz guest, but all the steal time recorded by
the host is fully accounted (ULONG_MAX limit), then
that could make up for lost/skipped timer ticks.
3) not accounting "extra" steal time (beyond the amount
of time accounted by account_process_tick) would reduce
the total amount of time that gets accounted if there
are missed ticks, taking time away from user/system/etc
Does the above make sense?
Am I overlooking some mechanism through which lost/skipped
ticks are made up for in the kernel? I looked through the
code in kernel/time/ briefly, but did not spot it...
--
All Rights Reversed.
[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 473 bytes --]
next prev parent reply other threads:[~2016-08-12 2:44 UTC|newest]
Thread overview: 42+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-07-13 14:50 [GIT PULL] cputime fixes and cleanups Frederic Weisbecker
2016-07-13 14:50 ` [PATCH 1/5] sched,time: Count actually elapsed irq & softirq time Frederic Weisbecker
2016-07-14 10:37 ` [tip:timers/nohz] sched/cputime: " tip-bot for Rik van Riel
2016-08-09 3:59 ` [PATCH 1/5] sched,time: " Wanpeng Li
2016-08-09 14:06 ` Rik van Riel
2016-08-09 23:07 ` Wanpeng Li
2016-08-10 7:51 ` Wanpeng Li
2016-08-09 23:25 ` Wanpeng Li
2016-08-09 23:31 ` Wanpeng Li
2016-08-09 23:35 ` Wanpeng Li
2016-08-09 23:39 ` Wanpeng Li
2016-08-10 5:07 ` Rik van Riel
2016-08-10 6:33 ` Wanpeng Li
2016-08-10 16:52 ` [PATCH] time,virt: resync steal time when guest & host lose sync Rik van Riel
2016-08-11 10:11 ` Wanpeng Li
2016-08-12 2:44 ` Rik van Riel [this message]
2016-08-12 7:09 ` Wanpeng Li
2016-08-12 15:58 ` Rik van Riel
2016-08-13 15:36 ` Frederic Weisbecker
2016-08-15 8:53 ` Wanpeng Li
2016-08-15 11:38 ` Wanpeng Li
2016-08-15 15:00 ` Rik van Riel
2016-08-15 22:19 ` Wanpeng Li
2016-08-16 1:31 ` Wanpeng Li
2016-08-16 2:11 ` Rik van Riel
2016-08-16 6:54 ` Wanpeng Li
2016-08-16 14:01 ` Rik van Riel
2016-08-16 23:08 ` Wanpeng Li
2016-08-12 16:33 ` Paolo Bonzini
2016-08-12 17:23 ` Rik van Riel
2016-08-13 7:18 ` Paolo Bonzini
2016-08-13 8:42 ` Ingo Molnar
2016-08-14 1:50 ` Rik van Riel
2016-08-18 8:23 ` Wanpeng Li
2016-07-13 14:50 ` [PATCH 2/5] nohz,cputime: Replace VTIME_GEN irq time code with IRQ_TIME_ACCOUNTING code Frederic Weisbecker
2016-07-14 10:37 ` [tip:timers/nohz] sched/cputime: " tip-bot for Rik van Riel
2016-07-13 14:50 ` [PATCH 3/5] sched: Complete cleanup of old vtime gen irqtime accounting Frederic Weisbecker
2016-07-14 10:38 ` [tip:timers/nohz] sched/cputime: Clean up the old vtime gen irqtime accounting completely tip-bot for Frederic Weisbecker
2016-07-13 14:50 ` [PATCH 4/5] sched: Reorganize vtime native irqtime accounting headers Frederic Weisbecker
2016-07-14 10:38 ` [tip:timers/nohz] sched/cputime: " tip-bot for Frederic Weisbecker
2016-07-13 14:50 ` [PATCH 5/5] time: Drop local_irq_save/restore from irqtime_account_irq Frederic Weisbecker
2016-07-14 10:38 ` [tip:timers/nohz] sched/cputime: Drop local_irq_save/restore from irqtime_account_irq() tip-bot for Rik van Riel
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1470969892.13905.120.camel@redhat.com \
--to=riel@redhat.com \
--cc=efault@gmx.de \
--cc=fweisbec@gmail.com \
--cc=kernellwp@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=pbonzini@redhat.com \
--cc=peterz@infradead.org \
--cc=rkrcmar@redhat.com \
--cc=tglx@linutronix.de \
--cc=wanpeng.li@hotmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox