Re: [Xen-devel] [PATCH v2 03/17] libxl/arm: Add a configuration option for ARM DomU ACPI

[Julien Grall <julien.grall@xxxxxxx>]

Hi Shannon,

On 12/07/16 04:40, Shannon Zhao wrote:
> On 2016/7/7 23:30, Wei Liu wrote:
> > On Mon, Jun 27, 2016 at 11:40:32AM +0100, Julien Grall wrote:
> > > > On 23/06/16 15:34, Shannon Zhao wrote:
> > > > > > On 2016年06月23日 21:39, Stefano Stabellini wrote:
> > > > > > > > On Thu, 23 Jun 2016, Shannon Zhao wrote:
> > > > > > > > > > > > From: Shannon Zhao <shannon.zhao@xxxxxxxxxx>
> > > > > > > > > > > > +    if (xlu_cfg_get_defbool(config, "acpi", &b_info->arch_arm.acpi, 0)) {
> > > > > > > > > > > > +        libxl_defbool_set(&b_info->arch_arm.acpi, 0);
> > > > > > > > > > > > +    }
> > > > > > > > We cannot share the existing code to parse the acpi paramter because
> > > > > > > > that is saved in b_info->u.hvm.acpi, right?
> > > > > > Yes.
> > > > > > > > It's a pity. I wonder if we
> > > > > > > > could refactor the existing code so that we can actually share the acpi
> > > > > > > > parameter between x86 and arm.
> > > > > > I have no idea about this since I'm not familiar with this. But is there
> > > > > > any downsides of current way? Because for x86, it will use
> > > > > > b_info->u.hvm.acpi and for ARM it will use b_info->arch_arm.acpi. I
> > > > > > think they don't conflict even though we store it at two places.
> > > > Yes, there is a downside.  Toolstack, such as libvirt, would need to have
> > > > separate code for x86 and ARM in order to enable/disable ACPI.
> > > >
> > > > I would introduce a new generic acpi parameters, deprecate
> > > > b_info->u.hvm.acpi. Ian, Stefano, Wei, any opinions?
> > Yeah, we can deprecate that field. But we need to take care to not break
> > users of the old field.
> Ok, what name would you suggest?
I would suggest b_info->u.acpi

Regards,

--
Julien Grall

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel