On Wed, 22 Sep 2021, Rahul Singh wrote:
> ARM architecture does not implement I/O ports. Ignore this call on ARM
> to avoid the overhead of making a hypercall just for Xen to return
> -ENOSYS.
>
> Signed-off-by: Rahul Singh <rahul.singh@arm.com>
> ---
> Change in v2:
> - Instead of returning success in XEN, ignored the call in xl.
> ---
> tools/libs/ctrl/xc_domain.c | 9 +++++++++
> 1 file changed, 9 insertions(+)
>
> diff --git a/tools/libs/ctrl/xc_domain.c b/tools/libs/ctrl/xc_domain.c
> index 23322b70b5..25c95f6596 100644
> --- a/tools/libs/ctrl/xc_domain.c
> +++ b/tools/libs/ctrl/xc_domain.c
> @@ -1348,6 +1348,14 @@ int xc_domain_ioport_permission(xc_interface *xch,
> uint32_t nr_ports,
> uint32_t allow_access)
> {
> +#if defined(__arm__) || defined(__aarch64__)
> + /*
> + * The ARM architecture does not implement I/O ports.
> + * Avoid the overhead of making a hypercall just for Xen to return -ENOSYS.
> + * It is safe to ignore this call on ARM so we just return 0.
> + */
> + return 0;
I think this is fine from a code change point of view.
I wonder if we want to return 0 or return an error here, but I am
thinking that 0 is OK because there is really nothing to do anyway as
they end up ignored one way or the other.
Based on that reasoning:
Reviewed-by: Stefano Stabellini <sstabellini@kernel.org>
> +#else
> DECLARE_DOMCTL;
>
> domctl.cmd = XEN_DOMCTL_ioport_permission;
> @@ -1357,6 +1365,7 @@ int xc_domain_ioport_permission(xc_interface *xch,
> domctl.u.ioport_permission.allow_access = allow_access;
>
> return do_domctl(xch, &domctl);
> +#endif
> }
>
> int xc_availheap(xc_interface *xch,
> --
> 2.17.1
>
>