On Mon, Jan 14, 2013 at 10:09:45AM +0800, Wenchao Xia wrote: > 于 2013-1-11 21:56, Luiz Capitulino 写道: > >On Thu, 10 Jan 2013 14:01:27 +0800 > >Wenchao Xia <xiaw...@linux.vnet.ibm.com> wrote: > > > >>于 2013-1-10 6:34, Eric Blake 写道: > >>>On 01/07/2013 12:27 AM, Wenchao Xia wrote: > >>>> These patch added a seperated layer to take internal or external > >>>> snapshots > >>>>in a unified way, the granularity is block device, so other functions can > >>>>just combine the request and submit, such as group snapshot, savevm. > >>>> > >>>>Total goal are: > >>>> Live back up vm in external or internal image, which need three > >>>> functions: > >>>>1 live snapshot block device internal/external. > >>>>2 live save vmstate internal/external. > >>>>3 combination of the function unit. > >>>> > >>>> This serial provide part one. > >>>> > >>> > >>>The design on this series needs to be coordinated with Pavel's > >>>attempts[1] to convert the existing HMP snapshot commands into QMP. I > >>>don't want to spend too much time reviewing two different divergent > >>>designs that both have the same goal of finally managing snapshots under > >>>QMP. > >>> > >>>[1] https://lists.gnu.org/archive/html/qemu-devel/2013-01/msg01326.html > >>> > >> A bit difference: This serial are taking internal/external block > >>snapshots but Pavel's are focusing are vm whole internal snapshots. It > > > >Pavel's work makes savevm, loadvm and delvm available in QMP. Also note > >that we do have blockdev-snapshot-sync. How does this series relate to it? > > > >I don't know if it's just me but, honestly speaking, I'm getting confused > >about the various proposals and APIs for snapshots. > > > >IMHO, it's time to go back a bit from code and discuss the status of the > >current APIs, what's missing and what the proposals are. > > > >This should also include writing a wiki page, as Anthony has done for > >previous ideas (which worked quite well). > > > I agree, can you help create an account on qemu wiki for me?
I have created an account for you. Please see the private message. Stefan
