On 12/06/2011 11:54 PM, Lucas Meneghel Rodrigues wrote: > On 12/02/2011 05:10 PM, Chris Evich wrote: >> Signed-off-by: Chris Evich<[email protected]> >> --- >> client/tests/{kvm => libvirt}/tests/boot_savevm.py | 0 >> 1 files changed, 0 insertions(+), 0 deletions(-) >> copy client/tests/{kvm => libvirt}/tests/boot_savevm.py (100%) >> >> diff --git a/client/tests/kvm/tests/boot_savevm.py >> b/client/tests/libvirt/tests/boot_savevm.py >> similarity index 100% >> copy from client/tests/kvm/tests/boot_savevm.py >> copy to client/tests/libvirt/tests/boot_savevm.py > > I think we can get to a single implementation of the test for both > libvirt and kvm autotest. Let's say we verify the vm class has some > certain attributes to tell apart libvirt vm and kvm vm, we can avoid > issuing a system_reset command to the monitor, or simply catch the > exception and logging it.
Agreed. Though really we can boil this all down even simpler by agreeing on common interface behavior. Proposal: vm.save_to_file(): Do everything necessary to store the running state of a VM into a file on the host. VM left in "off" state afterwards. vm.load_from_file(): Do everything necessary to restore a currently "off" VM into running state, based on contents of file on host. If testing requires more fine-grain control over the VM's state, then we should also have a set of functions to provide this, or parameters to save_to/load_from maybe. -- Chris Evich, RHCA, RHCE, RHCDS, RHCSS Quality Assurance Engineer e-mail: cevich + `@' + redhat.com o: 1-888-RED-HAT1 x44214 _______________________________________________ Autotest mailing list [email protected] http://test.kernel.org/cgi-bin/mailman/listinfo/autotest
