Re: [Xen-devel] [PATCH 20/21] xen: replace XEN_GUEST_HANDLE with XEN_GUEST_HANDLE_PARAM when appropriate

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 05.10.12 at 13:30, Ian Campbell <Ian.Campbell@xxxxxxxxxx> wrote:
> This probably ought to be folded into the original patch rather than
> kept separate.

Think so, yes.

> --- a/xen/include/public/tmem.h
> +++ b/xen/include/public/tmem.h
> @@ -96,7 +96,8 @@
>  
>  #ifndef __ASSEMBLY__
>  typedef xen_pfn_t tmem_cli_mfn_t;
> -typedef XEN_GUEST_HANDLE_PARAM(char) tmem_cli_va_t;
> +typedef XEN_GUEST_HANDLE(char) tmem_cli_va_t;
> +typedef XEN_GUEST_HANDLE_PARAM(char) tmem_cli_va_param_t;

This doesn't belong here - there's no use of tmem_cli_va_param_t
anywhere in the public interface afaict. I didn't check, but if there
are other uses of XEN_GUEST_HANDLE_PARAM() in the public
headers, I would suspect them to be wrong too - at the interface
layer, there shouldn't be any need for them.

> --- a/xen/include/xen/compat.h
> +++ b/xen/include/xen/compat.h
> @@ -21,7 +21,9 @@
>      __DEFINE_COMPAT_HANDLE(name, name); \
>      __DEFINE_COMPAT_HANDLE(const_ ## name, const name)
>  #define COMPAT_HANDLE(name)          __compat_handle_ ## name
> -
> +/* NB: it is assumed that if an arch uses the compat layer it does not
> + * distinguish handles from parameter handles. */
> +#define COMPAT_HANDLE_PARAM(name)    __compat_handle_ ## name
>  /* Is the compat handle a NULL reference? */
>  #define compat_handle_is_null(hnd)        ((hnd).c == 0)
 
This seems acceptable to me (minus the dropped newline).

Jan


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