On Mon, Dec 05, 2016 at 12:44:53AM -0600, Michael Roth wrote: > Quoting Bharata B Rao (2016-12-04 21:51:32) > > On Fri, Dec 2, 2016 at 11:46 PM, Dr. David Alan Gilbert > > <dgilb...@redhat.com> > > wrote: > > > > * Bharata B Rao (bharata....@gmail.com) wrote: > > > Hi, > > > > > > - Add ram object and dimm device at the source > > > > > > (qemu) object_add memory-backend-ram,id=ram0,size=128M > > > (qemu) device_add pc-dimm,id=dimm0,memdev=ram0 > > > > > > - Migrate the VM and remove the dimm device and ram object at the > > target > > > > > > (qemu) device_del dimm0 > > > (qemu) object_del ram0 > > > > > > - Adding the ram object with same id now at the target fails like > > this: > > > > > > (qemu) info memdev > > > > > > (qemu) object_add memory-backend-ram,id=ram0,size=128M > > > Duplicate ID 'ram0' for object > > > > > > Same behaviour is seen on x86 and Power. The problem isn't seen if > > > migration isn't involved (add, remove, add at the source itself). > > > > Is this the same as the one Michael Roth posted a fix for a few days > > back; 'monitor: fix object_del for command-line-created objects' > > > > https://lists.gnu.org/archive/html/qemu-devel/2016-11/msg05479.html > > > > > > Yes, this is same as above, I hadn't noticed Michael's fix. > > > > Will be good to have this fix in 2.8. > > I considered shooting for 2.8, but since it's only triggerable with HMP > (QMP doesn't use the QemuOpts-based parsing so it doesn't hit the > duplicate ID error, so libvirt in turn would be uneffected), and doesn't > seem to be a regression (AFAICT it's been an issue since object_del was > introduced), I figured it could wait till 2.9/2.8.1. > > Definitely a pain for testing though...
How complex is the fix? If it's simple, then AIUI we can still put real bugfixes like this into 2.8. -- David Gibson | I'll have my music baroque, and my code david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_ | _way_ _around_! http://www.ozlabs.org/~dgibson
signature.asc
Description: PGP signature