Re: [PATCH] arch/x86/hvm: Bind xen-created event channels to building domain

[Keir Fraser <keir@xen.org>]

On 03/01/2013 18:28, "Daniel De Graaf" <dgdegra@tycho.nsa.gov> wrote:

> Instead of using a hardcoded domain 0 as the endpoint for the event
> channels created in hvm_vcpu_initialise, use the domain ID of the
> building domain so that a domain builder in a domain other than dom0 has
> the expected access to the event channels.

This should be done by setting HVM_PARAM_DM_DOMAIN appropriately after the
guest is created. As it is, HVM_PARAM_DM_DOMAIN can be out of sync with the
event-channel bindings after this patch.

I see this patch is already applied, and I propose to revert it. Is that
okay with you, Jan?

 -- Keir

> Signed-off-by: Daniel De Graaf <dgdegra@tycho.nsa.gov>
> Cc: Keir Fraser <keir@xen.org>
> Cc: Jan Beulich <jbeulich@suse.com>
> ---
>  xen/arch/x86/hvm/hvm.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/xen/arch/x86/hvm/hvm.c b/xen/arch/x86/hvm/hvm.c
> index 40c1ab2..682d934 100644
> --- a/xen/arch/x86/hvm/hvm.c
> +++ b/xen/arch/x86/hvm/hvm.c
> @@ -1071,7 +1071,7 @@ int hvm_vcpu_initialise(struct vcpu *v)
>          goto fail3;
>  
>      /* Create ioreq event channel. */
> -    rc = alloc_unbound_xen_event_channel(v, 0, NULL);
> +    rc = alloc_unbound_xen_event_channel(v, current->domain->domain_id,
> NULL);
>      if ( rc < 0 )
>          goto fail4;
>  
> @@ -1081,7 +1081,7 @@ int hvm_vcpu_initialise(struct vcpu *v)
>      if ( v->vcpu_id == 0 )
>      {
>          /* Create bufioreq event channel. */
> -        rc = alloc_unbound_xen_event_channel(v, 0, NULL);
> +        rc = alloc_unbound_xen_event_channel(v, current->domain->domain_id,
> NULL);
>          if ( rc < 0 )
>              goto fail2;
>          v->domain->arch.hvm_domain.params[HVM_PARAM_BUFIOREQ_EVTCHN] = rc;