Re: [libvirt] [PATCH 6/6] restore correctly the state of the VM upon migration

2009-10-07 Thread Daniel Veillard
On Wed, Oct 07, 2009 at 03:14:18PM +0200, Paolo Bonzini wrote: > On 10/07/2009 03:03 PM, Daniel Veillard wrote: >> NACK ! an error in migration should not result in a dead daemon >> assert/exit forbidden even in error handling paths. > > I have no problem removing the assertion, however notice th

Re: [libvirt] [PATCH 6/6] restore correctly the state of the VM upon migration

2009-10-07 Thread Paolo Bonzini
On 10/07/2009 03:03 PM, Daniel Veillard wrote: NACK ! an error in migration should not result in a dead daemon assert/exit forbidden even in error handling paths. I have no problem removing the assertion, however notice that I added it simply to match vm->state = migrateFrom ? VIR_DOMA

Re: [libvirt] [PATCH 6/6] restore correctly the state of the VM upon migration

2009-10-07 Thread Daniel Veillard
On Thu, Oct 01, 2009 at 08:18:33PM +0200, Paolo Bonzini wrote: > In order to correctly pass the paused/unpaused state on the remote side, > we use the newly introduced return code of qemudDomainMigratePerform. > > The return code does not need to be public (it is internal to the QEMU > driver). A

[libvirt] [PATCH 6/6] restore correctly the state of the VM upon migration

2009-10-01 Thread Paolo Bonzini
In order to correctly pass the paused/unpaused state on the remote side, we use the newly introduced return code of qemudDomainMigratePerform. The return code does not need to be public (it is internal to the QEMU driver). A return code of 0 specifies the old behavior. * src/qemu/qemu_driver.c (