Re: [Xen-devel] [PATCH v2 2/5] xen: Provide XEN_DMOP_add_to_physmap

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 23.10.17 at 11:05, <ross.lagerwall@xxxxxxxxxx> wrote:

First of all, instead of xen: please consider using something more
specific, like x86/hvm:.

> --- a/xen/include/public/hvm/dm_op.h
> +++ b/xen/include/public/hvm/dm_op.h
> @@ -368,6 +368,22 @@ struct xen_dm_op_remote_shutdown {
>                             /* (Other reason values are not blocked) */
>  };
>  
> +/*
> + * XEN_DMOP_add_to_physmap : Sets the GPFNs at which a page range appears in
> + *                           the specified guest's pseudophysical address

Talking of "pseudophysical" is at least confusing for HVM guests. So
far it was my understanding that such exists for PV guests only.

> + *                           space. Identical to XENMEM_add_to_physmap with
> + *                           space == XENMAPSPACE_gmfn_range.
> + */
> +#define XEN_DMOP_add_to_physmap 17
> +
> +struct xen_dm_op_add_to_physmap {
> +    uint16_t size;         /* Number of GMFNs to process. */

Why limit this to 16 bits?

> +    uint16_t pad0;
> +    uint32_t pad1;
> +    uint64_aligned_t idx;  /* Index into GMFN space. */

Why would you call this "idx"? The other interface and its naming
should have no significance here. So perhaps "src_gfn" and ...

> +    uint64_aligned_t gpfn; /* Starting GPFN where the GMFNs should appear. */

... "dst_gfn"?

Jan


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