On 02/15/2012 08:04 AM, Michal Privoznik wrote:
---
tools/virsh.c | 46 ++++++++++++++++++++++++++++++++++++++++++++++
tools/virsh.pod | 6 ++++++
2 files changed, 52 insertions(+), 0 deletions(-)
ACK.
I'm wondering, though, if we should make this command smarter (as a
followup patch). The API for dompmsuspend made it into 0.9.10, but
dompmwakeup depends on 0.9.11. Fortunately, out of the box, 0.9.10
always fails on dompmsuspend (we didn't implement anything in qemu).
But for backport situations (such as what I'm guessing RHEL will want to
do), where you can backport implementation but not API, then we will
have the (unfortunate) case that dompmsuspend might work while
dompmwakeup fails for anyone that bases a release on the 0.9.10 API. Is
it worth adding a virsh flag, so that 'dompmwakeup dom --qemu' tries
virDomainPMWakeup, and if it fails, it then falls back to trying
virDomainQemuMonitorCommand to issue a raw "system_wakeup" monitor
command rather than completely giving up?
--
Eric Blake eblake(a)redhat.com +1-919-301-3266
Libvirt virtualization library
http://libvirt.org