Am 03.01.2011 17:37, Glauber Costa wrote: > On Mon, 2011-01-03 at 09:33 +0100, Jan Kiszka wrote: >> From: Jan Kiszka <jan.kiszka@xxxxxxxxxxx> >> >> If kvmclock is used, which implies the kernel supports it, register a >> kvmclock device with the sysbus. Its main purpose is to save and restore >> the kernel state on migration, but this will also allow to visualize it >> one day. >> >> Signed-off-by: Jan Kiszka <jan.kiszka@xxxxxxxxxxx> >> CC: Glauber Costa <glommer@xxxxxxxxxx> > > Hi Jan. > > I've just recently posted a patch (not sure what was made from it), that > fixes a bug that you reintroduce here. Hmm, must have missed it. > > The bug is: if we call KVM_GET_CLOCK ioctl in pre_save, this means that > this value will change every time we issue savevm, even if the machine > is not run in between. > > Ideally, we'd like to have two consecutive savevms returning the exact > same thing in that situation. > > I like the general approach of moving it to sysbus, but please move the > ioctl to change state notifiers. OK, will look into this. Jan
Attachment:
signature.asc
Description: OpenPGP digital signature
