Re: [Qemu-devel] [RFC][PATCH v4 09/18] virtagent: add va_shutdown RPC

2010-11-18 Thread Jes Sorensen 
On 11/16/10 17:01, Michael Roth wrote:
 RPC to initiate guest shutdown/reboot/powerdown

Do we really need this? After all those events can be passed down via ACPI?

Cheers,
Jes

Re: [Qemu-devel] [RFC][PATCH v4 09/18] virtagent: add va_shutdown RPC

2010-11-18 Thread Anthony Liguori 

On 11/18/2010 08:17 AM, Jes Sorensen wrote:

On 11/16/10 17:01, Michael Roth wrote:
   

RPC to initiate guest shutdown/reboot/powerdown
 

Do we really need this? After all those events can be passed down via ACPI?
   


Reboot can't be passed by ACPI.  This is why virDomainReboot is still 
not functional in KVM.  Every other hypervisor does something identical 
to this series.


And a hypervisor initiated shutdown ought to be treated differently than 
an ACPI shutdown.  Namely, it should correspond to an immediate shut 
down vs. prompting the user.


Not all OSes response to ACPI shutdown either FWIW.

Regards,

Anthony Liguori


Cheers,
Jes

Re: [Qemu-devel] [RFC][PATCH v4 09/18] virtagent: add va_shutdown RPC

2010-11-18 Thread Jes Sorensen 
On 11/18/10 16:35, Anthony Liguori wrote:
 On 11/18/2010 08:17 AM, Jes Sorensen wrote:
 On 11/16/10 17:01, Michael Roth wrote:
   
 RPC to initiate guest shutdown/reboot/powerdown
  
 Do we really need this? After all those events can be passed down via
 ACPI?

 
 Reboot can't be passed by ACPI.  This is why virDomainReboot is still
 not functional in KVM.  Every other hypervisor does something identical
 to this series.
 
 And a hypervisor initiated shutdown ought to be treated differently than
 an ACPI shutdown.  Namely, it should correspond to an immediate shut
 down vs. prompting the user.
 
 Not all OSes response to ACPI shutdown either FWIW.

Ok, fair enough.

I just would like to avoid duplicating functionality if we don't have to.

Cheers,
Jes