Re: [Xen-devel] [PATCH v5 1/2] docs/designs: Add a design document for non-cooperative live migration

["Durrant, Paul" <pdurrant@xxxxxxxxxxxx>]

> -----Original Message-----
> >>> +HVM guests can already be migrated on Xen without guest co-operation but 
> >>> only
> >>> +if they don’t have PV drivers installed[1] or are in power state S3. The
> >>
> >> S3 is very ACPI centric, so I would prefer if we avoid the term. I think
> >> the non-ACPI description is "suspend to RAM". I would be OK is you
> >> mention S3 in parenthesis.
> >
> > I'm actually pulling this from the way the code is currently written, which 
> > is clearly quite x86
> specific:
> >
> > xc_hvm_param_get(CTX->xch, domid, HVM_PARAM_ACPI_S_STATE, &hvm_s_state)
> > .
> > .
> > .
> > if (dsps->type == LIBXL_DOMAIN_TYPE_HVM && (!hvm_pvdrv || hvm_s_state)) {
> >      LOGD(DEBUG, domid, "Calling xc_domain_shutdown on HVM domain");
> >      ret = xc_domain_shutdown(CTX->xch, domid, SHUTDOWN_suspend);
> >      .
> >      .
> > }
> >
> > So actually I should say 'not in power state S0'.
> 
> I understand that the current code is x86 specific. Arm would likely
> have a similar requirement although not based on ACPI.
> 
> However, my point here is nothing in the document says it is focusing on
> x86 only. The concept itself is not arch specific, the document is
> mostly x86 free except in a couple of bits. So I would like them to be
> rewritten in an arch-agnostic way.
> 
> Note that I am ok with arch-specific example.
> 

Sure. I'll try not to be x86 specific where it's not necessary.

  Paul
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel