Hi, Wenchao In order to determine whether my test method or my system environment have problems, I want to test the patch on the qemu version which you used. You can give me a rough version number.
You told me that patch is based on the version before qemu-1.3. But in your patch, there is a diff: --- a/savevm.c +++ b/savevm.c @@ -698,7 +698,7 @@ int qemu_get_byte(QEMUFile *f) return result; } -static int64_t qemu_ftell(QEMUFile *f) +int64_t qemu_ftell(QEMUFile *f) { return f->buf_offset - f->buf_size + f->buf_index; } @@ -2061,32 +2061,10 @@ out: return ret; } But before qemu-1.3, the "qemu_ftell" function is not a static function. In qemu-1.3 and the following version, the "qemu_ftell" function is change to a static function. So I'm not sure which version you used. Which tool you used to connect the VM, Spice, VNC or others? When I use the Spice, after I "loadvm" the snapshot which is create by the "migrate image:" command, the qemu-kvm will be error or crash. But when I use VNC, it is OK. How do you think this issue? Please give me some suggestion. And can you give me the detailed "qemu-kvm" command, which you used to start the VM. I want compare your "qemu-kvm" command and my. Best Regards Liu Ji ----------------------------------------------------- 2013/2/19 Wenchao Xia <xiaw...@linux.vnet.ibm.com> > Hi, Liu Ji > Thanks for trying my trival patch. That patch is based on the version > before qemu-1.3, and since qemu have important change about migration > from Juan after that, So I think it would not work since 1.3. > > I am trying to fix the size problem now, and then reimplement a formal > patch of "live savevm" after that, as part of the incremental backup > enhancement, you can take a look of the plan at: > http://wiki.qemu.org/Features/VMSnapshotEnchancement > Pls forgive me about the latency of the work, I have just come back > from vacation.:) > > > Hello Wenchao Xia, > > I am very interested in your "live backup vm" patch of qemu, which have > > been sent to qemu-devel mail-list in "8 Dec 2012". > > http://lists.gnu.org/archive/html/qemu-devel/2012-12/msg00800.html > > > > I have test your patch in qemu-1.3.1, but I have two problems: > > > > 1) After I exec the "migrate image:" command, although the CPU of VM is > not > > suspended, the running of VM is very slow. It is very difficult to > operate > > the VM > > > > by Spice. > > 2) When I "loadvm" the snapshot which is create by the "migrate image:" > > command, the Qemu output error. I change the version of qemu to 1.3.0, > the > > "loadvm" > > > > caused a crash. > > Could you tell me which baseline version of qemu you used? And do you > have > > the same problems? > > Do you have optimized the patch, and can you provide the latest patch for > > me? > > > > Thanks. > > > > > -- > Best Regards > > Wenchao Xia > >