Re: [PATCH v3 07/12] xen: introduce domain-layout.h with common domain_use_host_layout()

[Oleksii Kurochko <oleksii.kurochko@xxxxxxxxx>]

On 4/21/26 11:20 AM, Jan Beulich wrote:
> On 10.04.2026 17:54, Oleksii Kurochko wrote:
> > --- /dev/null
> > +++ b/xen/include/xen/domain-layout.h
> > @@ -0,0 +1,28 @@
> > +#ifndef __XEN_DOMAIN_LAYOUT_H__
> > +#define __XEN_DOMAIN_LAYOUT_H__
> > +
> > +#include <xen/domain.h>
> > +#include <xen/paging.h>
> > +#include <xen/sched.h>
> > +
> > +/*
> > + * Is a domain using the host memory layout?
> > + *
> > + * domain_use_host_layout() is always False for PV guests.
> > + *
> > + * Direct-mapped domains (autotranslated domains with memory allocated
> > + * contiguously and mapped 1:1 so that GFN == MFN) are always using the
> > + * host memory layout to avoid address clashes.
> What is "to avoid address clashes" about? If GFN == MFN, how could there
> be clashes?
>
> > + * The hardware domain will use the host layout (regardless of
> > + * direct-mapped) because some OS may rely on specific address ranges
> > + * for the devices. PV Dom0, like any other PV guests, has
> > + * domain_use_host_layout() returning False.
> This last sentence is somewhat redundant and somewhat in conflict with
> what is said further up. If you did s/guests/domains (including Dom0)/
> there, imo this sentence would best be dropped from down here.
I will change the comment to:

/*
 * Is a domain using the host memory layout?
 *
 * domain_use_host_layout() is always False for PV domains (including 
Dom0).
 *
 * Direct-mapped domains (autotranslated domains with memory allocated
 * contiguously and mapped 1:1 so that GFN == MFN) must use the host
 * memory layout since GFN == MFN by definition.
 *
 * The hardware domain will use the host layout (regardless of
 * direct-mapped) because some OS may rely on specific address ranges
 * for the devices.
 */

> With these adjustments:
> Acked-by: Jan Beulich <jbeulich@xxxxxxxx>
Thanks.

~ Oleksii