From: Chris Wright <chrisw@redhat.com>
To: kvm@vger.kernel.org
Cc: qemu-devel@nongnu.org
Subject: [Qemu-devel] KVM call minutes for June 15
Date: Tue, 15 Jun 2010 08:18:12 -0700 [thread overview]
Message-ID: <20100615151812.GA24131@x200.localdomain> (raw)
Page cache controls
- cache is 60% duplicated between host and guest (when not using cache=none)
- Balbir posted 2 patches to eliminate this
- boot parameter for preferred reclaim
- not always have balloon driver
- need a boot parameter
- perhaps add a balloon cmd to give a hint before a more forceful request
- hard/soft quota
- exceed hard, swap
- exceed soft...no real penalty
- another consideration, block device hint (has large cache in host, don't
agressively cache in guest)
- lru weighted by disk speed might make sense on bare metal
- balloon driver filter
- balloon unmapped page cache pages first
- guest has no incentive to cooperate (give up memory and possibly performance)
Migration Subsections
- looks good...
- doesn't replace need to increment version numbers
- needs some documentation to make sure subtlties are clear
- size for each section would be useful (breaks protocol)
- while size is possibly useful, breaks protocol
- anthony wants a self-descriptive protocol if we break (like ASN.1)
- but all moot until everything is converted to vmstate
- need to consider specifying a new protocol
KVM/qemu patches
- patch rate is high, documentation is low, review is low
- patches need to include better descriptions and documentation
- will slow down patch writers
- will make it easier for patch reviewers
QMP
- spec review
- migration events
- introduce migration connected event
- migration completed/done event (more contentious)
- no data, just indication that it's done
- run query migrate to get status
- 0.14 proper async
- (-rc is coming next Monday)
- important to start simple, fix as go...
- no good indication of completion in general across monitor
- includes feedback from guest (pci remove device, shutdown, balloon, etc...)
- async is the consistent issue
- libvirt should not use QMP in 0.13
- need to declare QMP Unstable 0.13 (still need spec review)
next reply other threads:[~2010-06-15 15:18 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2010-06-15 15:18 Chris Wright [this message]
2010-06-15 15:41 ` [Qemu-devel] Re: KVM call minutes for June 15 Christoph Hellwig
2010-06-15 16:13 ` Anthony Liguori
2010-06-16 6:55 ` Markus Armbruster
2010-06-16 8:58 ` Yoshiaki Tamura
2010-06-16 11:55 ` Arnd Bergmann
2010-06-16 18:07 ` Blue Swirl
2010-06-17 5:15 ` Aurelien Jarno
2010-06-16 8:58 ` Paolo Bonzini
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=20100615151812.GA24131@x200.localdomain \
--to=chrisw@redhat.com \
--cc=kvm@vger.kernel.org \
--cc=qemu-devel@nongnu.org \
/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).