On 02/04/13 19:50, Eric Blake wrote:
On 04/01/2013 11:42 PM, Osier Yang wrote:
> This prohibits all clear cpumap eariler in virsh, for both vcpupin
> and emulatorpin.
Is that really necessary?
Honestly, I'm not sure. :-)
Sometimes, letting virsh allow obviously
wrong constructs, in order to prove that libvirt itself will flag things
as sensible errors, makes for nicer testing. Virsh should only filter
out obviously wrong things if it can give a nicer error message than
what the underlying API would produce.
I tended to error out earlier, but I also agreed with you that sometimes
the good way is to let it fall through to driver. So I'm fine to drop this
patch.
Osier