[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [RFC PATCH v2 0/3] Start fixing the pvpanic mess
From: |
Daniel P. Berrange |
Subject: |
Re: [Qemu-devel] [RFC PATCH v2 0/3] Start fixing the pvpanic mess |
Date: |
Thu, 22 Aug 2013 10:17:59 +0100 |
User-agent: |
Mutt/1.5.21 (2010-09-15) |
On Wed, Aug 21, 2013 at 06:56:52PM +0200, Paolo Bonzini wrote:
> Il 21/08/2013 18:55, Daniel P. Berrange ha scritto:
> > On Wed, Aug 21, 2013 at 06:51:11PM +0200, Paolo Bonzini wrote:
> >> Il 21/08/2013 18:48, Daniel P. Berrange ha scritto:
> >>> No, <on_crash> is the right thing to be using for this from
> >>> libvirt's pov & I don't think we should invent something new.
> >>> The <on_crash> element has always been intended to represent
> >>> handling of guest panics, not qemu internal errors.
> >>
> >> Actually for Xen HVM guests, it mostly traps things such as failed
> >> vmentries. The Xen PV-on-HVM drivers do not register a panic notifier
> >> that moves the guest to the "crashed" state.
> >>
> >> <on_crash> cannot be salvaged, in my opinion, because all domain XMLs in
> >> the wild will have a setting that causes libvirt to add "-device
> >> isa-pvpanic". Thus changing libvirt versions will change guest
> >> hardware, which is _very_ bad.
> >>
> >> In addition, Windows XP and 2003 will show the annoying device wizard
> >> upon a libvirt upgrade, and fixing this is what surfaced all the mess.
> >
> > The existance of a <on_crash> element should not be having any
> > effect on what hardware we create. That is merely a lifecycle
> > policy setting that should be completely independant of the
> > guest device model.
> >
> > eg it is valid to have <on_crash> present in the XML at all
> > times, even if there's no pvpanic device present. That simply
> > means the actions will never be triggered.
>
> So are you suggesting to add a <pvpanic/> element to <devices>? That
> may be fine, but it doesn't seem very user-friendly.
Yes, if we're going to have pvpanic be user controllable, it must be
via an explicit device element.
None of the <on_XXXX> elements should have any impact on guest ABI
model. They're purely lifecycle policy settings.
Daniel
--
|: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org -o- http://virt-manager.org :|
|: http://autobuild.org -o- http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|
- Re: [Qemu-devel] [RFC PATCH v2 0/3] Start fixing the pvpanic mess, (continued)
- Re: [Qemu-devel] [RFC PATCH v2 0/3] Start fixing the pvpanic mess, Daniel P. Berrange, 2013/08/21
- Re: [Qemu-devel] [RFC PATCH v2 0/3] Start fixing the pvpanic mess, Daniel P. Berrange, 2013/08/21
- Re: [Qemu-devel] [RFC PATCH v2 0/3] Start fixing the pvpanic mess, Paolo Bonzini, 2013/08/21
- Re: [Qemu-devel] [RFC PATCH v2 0/3] Start fixing the pvpanic mess, Eric Blake, 2013/08/21
- Re: [Qemu-devel] [RFC PATCH v2 0/3] Start fixing the pvpanic mess, Paolo Bonzini, 2013/08/21
- Re: [Qemu-devel] [RFC PATCH v2 0/3] Start fixing the pvpanic mess,
Daniel P. Berrange <=