Re: [RESEND PATCH v2 1/3] xen: Define xen_domain_handle_t encoding and formatting

["Anthony PERARD" <anthony.perard@xxxxxxxxxx>]

On Fri, Aug 29, 2025 at 09:58:15AM +0000, Teddy Astie wrote:
> diff --git a/xen/include/public/xen.h b/xen/include/public/xen.h
> index 82b9c05a76..a219ef870f 100644
> --- a/xen/include/public/xen.h
> +++ b/xen/include/public/xen.h
> @@ -973,6 +973,13 @@ typedef struct dom0_vga_console_info {
>  #define xen_vga_console_info dom0_vga_console_info
>  #define xen_vga_console_info_t dom0_vga_console_info_t
>  
> +/*
> + * The domain handle is chosen by the toolstack, and intended to hold a UUID
> + * conforming to RFC 9562 (i.e. big endian).
> + *
> + * Certain cases (e.g. SMBios) transform it to a Microsoft GUID (little
> + * endian) for presentation to the guest.
> + */

So, there's already a definition about this, but it's a comment on
XENVER_guest_handle (in public/version.h):

> /* arg == xen_domain_handle_t.
>  *
>  * The toolstack fills it out for guest consumption. It is intended to hold
>  * the UUID of the guest.
>  */

We can add that the binary format is most significant byte or network
byte order, but that's already the format stated in RFC 4122 and its
successor RFC 9562.

Also, I don't think it is useful here to point out how that handle/UUID
might be used/transformed by guest.

>  typedef uint8_t xen_domain_handle_t[16];

Cheers,

-- 

--
Anthony Perard | Vates XCP-ng Developer

XCP-ng & Xen Orchestra - Vates solutions

web: https://vates.tech