From: Paolo Bonzini <pbonzini@redhat.com>
To: Marcelo Tosatti <mtosatti@redhat.com>
Cc: "Radim Krčmář" <rkrcmar@redhat.com>,
kvm@vger.kernel.org, qemu-devel <qemu-devel@nongnu.org>,
"Dr. David Alan Gilbert" <dgilbert@redhat.com>,
"Juan Quintela" <quintela@redhat.com>,
"Eduardo Habkost" <ehabkost@redhat.com>,
"Roman Kagan" <rkagan@virtuozzo.com>
Subject: Re: [QEMU PATCH] kvmclock: advance clock by time window between vm_stop and pre_save
Date: Fri, 4 Nov 2016 18:35:23 -0400 (EDT) [thread overview]
Message-ID: <184247986.10964505.1478298923187.JavaMail.zimbra@redhat.com> (raw)
In-Reply-To: <20161104214747.GA17560@amt.cnet>
> > But that's why separating the two cases brings us the best of both worlds.
> > If migrating a paused guest, there's no need for any adjustment, so no
> > advance_clock hack. If pausing at the end of migration, there's no need
> > to pause kvmclock (this patch is effectively working around 00f4d64ee76e)
> > and if we don't do that we can just call KVM_GET_CLOCK at pre_save time.
>
> That was my internal v1. But then, the destination ignores s->clock
> as follows:
>
> if (running) {
> struct kvm_clock_data data = {};
> uint64_t time_at_migration = kvmclock_current_nsec(s);
Right, but this is unnecessary in 4.9-rc, where KVM_GET_CLOCK *already*
returns the same as QEMU's kvmclock_current_nsec. So this is the other
half of my suggestion, where we need to check the behavior of KVM_GET_CLOCK
via KVM_CHECK_EXTENSION.
I think it's okay to only fix the bug with master clock enabled and
for new KVM with sensible KVM_GET_CLOCK semantics.
Paolo
> s->clock_valid = false;
>
> /* We can't rely on the migrated clock value, just discard it */
> if (time_at_migration) {
> s->clock = time_at_migration;
> }
>
> data.clock = s->clock;
> ret = kvm_vm_ioctl(kvm_state, KVM_SET_CLOCK, &data);
>
> So you need to send that "ns" value (difference of two clock reads)
> separately.
>
> >
> > > Oh, and this does introduce a minor bug to this patch -- the time
> > > counted by KVM_GET_CLOCK is has different frequency CLOCK_MONOTONIC.
> > > Not accounting for that is bearable.
> >
> > Not really, I was going to point that out when I got to replying with
> > a review. :)
> >
> > Paolo
>
>
next prev parent reply other threads:[~2016-11-04 22:35 UTC|newest]
Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-11-04 9:43 [QEMU PATCH] kvmclock: advance clock by time window between vm_stop and pre_save Marcelo Tosatti
2016-11-04 12:28 ` Juan Quintela
2016-11-04 12:35 ` Marcelo Tosatti
2016-11-04 14:00 ` Marcelo Tosatti
2016-11-04 15:25 ` Radim Krčmář
2016-11-04 15:33 ` Paolo Bonzini
2016-11-04 15:48 ` Radim Krčmář
2016-11-04 15:57 ` Paolo Bonzini
2016-11-04 17:16 ` Radim Krčmář
2016-11-04 21:29 ` Paolo Bonzini
2016-11-04 21:47 ` Marcelo Tosatti
2016-11-04 22:35 ` Paolo Bonzini [this message]
2016-11-07 14:31 ` Roman Kagan
2016-11-07 19:31 ` Marcelo Tosatti
2016-11-04 16:24 ` Marcelo Tosatti
2016-11-04 17:34 ` Radim Krčmář
2016-11-04 18:29 ` Marcelo Tosatti
2016-11-04 20:07 ` Radim Krčmář
2016-11-04 16:04 ` Marcelo Tosatti
2016-11-04 17:07 ` Marcelo Tosatti
2016-11-04 17:39 ` Radim Krčmář
2016-11-04 18:31 ` Marcelo Tosatti
2016-11-07 13:08 ` Dr. David Alan Gilbert
2016-11-04 16:59 ` [QEMU PATCH v2] " Marcelo Tosatti
2016-11-04 18:57 ` Juan Quintela
2016-11-07 15:46 ` Dr. David Alan Gilbert
2016-11-07 19:41 ` Marcelo Tosatti
2016-11-07 20:03 ` Dr. David Alan Gilbert
2016-11-08 0:06 ` Marcelo Tosatti
2016-11-08 10:22 ` Dr. David Alan Gilbert
2016-11-08 13:32 ` Marcelo Tosatti
2016-11-09 19:32 ` Marcelo Tosatti
2016-11-09 16:23 ` Paolo Bonzini
2016-11-09 16:28 ` Dr. David Alan Gilbert
2016-11-09 16:33 ` Paolo Bonzini
2016-11-10 11:48 ` Marcelo Tosatti
2016-11-10 17:57 ` Paolo Bonzini
2016-11-11 14:23 ` Marcelo Tosatti
2017-02-07 10:02 ` Wanpeng Li
2017-02-07 12:18 ` Marcelo Tosatti
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=184247986.10964505.1478298923187.JavaMail.zimbra@redhat.com \
--to=pbonzini@redhat.com \
--cc=dgilbert@redhat.com \
--cc=ehabkost@redhat.com \
--cc=kvm@vger.kernel.org \
--cc=mtosatti@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=quintela@redhat.com \
--cc=rkagan@virtuozzo.com \
--cc=rkrcmar@redhat.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;
as well as URLs for NNTP newsgroup(s).