On Wed, Aug 21, 2013 at 06:43:13PM +0200, Paolo Bonzini wrote:
The pvpanic mess is even bigger than anticipated. Let's fix the monitor's behavior (patch 1), get rid of all traces that the broken pvpanic existed (patch 2), and give it a new name so that libvirt can detect a design that works (patch 3).
All downstreams are urged to apply patches 1+2 as soon as they are merged in QEMU.
Still, there are still other problems to solve.
In QEMU, exposing "-device isa-pvpanic" in the ACPI tables. Quite frankly I don't have the time to fix this. We have ~3 months though. Patch 3 should not be applied until it is fixed.
Also, libvirt needs to know under which circumstances to add "-device isa-pvpanic", besides obviously the availability of the device.
IMHO it is just too complicated to retrofit all complications in <on_crash>. In fact, I suspect <on_crash> would match more closely QEMU's "internal error" state, and it would be quite useful to add that to the QEMU driver.
Thus, libvirt could add support for an <on_panic> element with the following values:
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. 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 :|