RE: [Xen-ia64-devel][PATCH][VTD][dom0] dom0 vtd patch
Isaku Yamahata wrote:
> On Thu, Oct 23, 2008 at 02:14:52PM +0800, Xu, Anthony wrote:
>> Isaku Yamahata wrote:
>>> On Wed, Oct 22, 2008 at 05:43:13PM +0800, Xu, Anthony wrote:
dom0/ia64/vtd patch
Signed-off-by: Anthony Xu <[EMAIL PROTECTED]>
>>>
>>>
>>> In XEN_DOMCTL_get_device_group case.
>>>
>>>
>>> struct xen_domctl_get_device_group {
>>> uint32_t machine_bdf; /* IN */
>>> uint32_t max_sdevs;/* IN */
>>> uint32_t num_sdevs;/* OUT */
>>> XEN_GUEST_HANDLE_64(uint32) sdev_array; /* OUT */ };
>>>
>>> XEN_DOMCTL_get_device_group is used by libxc.
>>> sdev_array needs ot be handled correctly.
>>
>> I guess you are talking about xencomm,
>> I don't know how to handle it.
>> Can you enlighten me?
>
> How about this?
That's good.
I understand how to handle it now.
Anthony
>
>
> [IA64] xencomm: support several domctls for VT-d.
>
> This patch adds xencomm support of several domctl hypercalls
> for VT-d.
>
> Signed-off-by: Anthony Xu <[EMAIL PROTECTED]>
> Signed-off-by: Isaku Yamahata <[EMAIL PROTECTED]>
>
> diff -r c87adc976013 arch/ia64/xen/xcom_privcmd.c
> --- a/arch/ia64/xen/xcom_privcmd.cMon Oct 20 15:29:07 2008 +0100
> +++ b/arch/ia64/xen/xcom_privcmd.cFri Oct 24 11:22:02 2008 +0900
> @@ -340,6 +340,20 @@
> return -ENOMEM;
> set_xen_guest_handle(kern_op.u.hvmcontext.buffer, (void*)desc);
> break;
> + case XEN_DOMCTL_get_device_group:
> + {
> + struct xen_domctl_get_device_group *get_device_group =
> + &kern_op.u.get_device_group;
> + desc = xencomm_map(
> + xen_guest_handle(get_device_group->sdev_array),
> + get_device_group->max_sdevs * sizeof(uint32_t));
> + if (xen_guest_handle(get_device_group->sdev_array) != NULL &&
> + get_device_group->max_sdevs > 0 && desc == NULL)
> + return -ENOMEM;
> + set_xen_guest_handle(kern_op.u.get_device_group.sdev_array,
> + (void*)desc);
> + break;
> + }
> case XEN_DOMCTL_max_vcpus:
> case XEN_DOMCTL_scheduler_op:
> case XEN_DOMCTL_setdomainhandle:
> @@ -354,6 +368,12 @@
> case XEN_DOMCTL_set_opt_feature:
> case XEN_DOMCTL_assign_device:
> case XEN_DOMCTL_subscribe:
> + case XEN_DOMCTL_test_assign_device:
> + case XEN_DOMCTL_deassign_device:
> + case XEN_DOMCTL_bind_pt_irq:
> + case XEN_DOMCTL_unbind_pt_irq:
> + case XEN_DOMCTL_memory_mapping:
> + case XEN_DOMCTL_ioport_mapping:
> break;
> case XEN_DOMCTL_pin_mem_cacheattr:
> return -ENOSYS;
___
Xen-ia64-devel mailing list
Xen-ia64-devel@lists.xensource.com
http://lists.xensource.com/xen-ia64-devel
Re: [Xen-ia64-devel][PATCH][VTD][dom0] dom0 vtd patch
On Thu, Oct 23, 2008 at 02:14:52PM +0800, Xu, Anthony wrote:
> Isaku Yamahata wrote:
> > On Wed, Oct 22, 2008 at 05:43:13PM +0800, Xu, Anthony wrote:
> >> dom0/ia64/vtd patch
> >>
> >> Signed-off-by: Anthony Xu <[EMAIL PROTECTED]>
> >
> >
> > In XEN_DOMCTL_get_device_group case.
> >
> >
> > struct xen_domctl_get_device_group {
> > uint32_t machine_bdf; /* IN */
> > uint32_t max_sdevs;/* IN */
> > uint32_t num_sdevs;/* OUT */
> > XEN_GUEST_HANDLE_64(uint32) sdev_array; /* OUT */
> > };
> >
> > XEN_DOMCTL_get_device_group is used by libxc.
> > sdev_array needs ot be handled correctly.
>
> I guess you are talking about xencomm,
> I don't know how to handle it.
> Can you enlighten me?
How about this?
[IA64] xencomm: support several domctls for VT-d.
This patch adds xencomm support of several domctl hypercalls
for VT-d.
Signed-off-by: Anthony Xu <[EMAIL PROTECTED]>
Signed-off-by: Isaku Yamahata <[EMAIL PROTECTED]>
diff -r c87adc976013 arch/ia64/xen/xcom_privcmd.c
--- a/arch/ia64/xen/xcom_privcmd.c Mon Oct 20 15:29:07 2008 +0100
+++ b/arch/ia64/xen/xcom_privcmd.c Fri Oct 24 11:22:02 2008 +0900
@@ -340,6 +340,20 @@
return -ENOMEM;
set_xen_guest_handle(kern_op.u.hvmcontext.buffer, (void*)desc);
break;
+ case XEN_DOMCTL_get_device_group:
+ {
+ struct xen_domctl_get_device_group *get_device_group =
+ &kern_op.u.get_device_group;
+ desc = xencomm_map(
+ xen_guest_handle(get_device_group->sdev_array),
+ get_device_group->max_sdevs * sizeof(uint32_t));
+ if (xen_guest_handle(get_device_group->sdev_array) != NULL &&
+ get_device_group->max_sdevs > 0 && desc == NULL)
+ return -ENOMEM;
+ set_xen_guest_handle(kern_op.u.get_device_group.sdev_array,
+(void*)desc);
+ break;
+ }
case XEN_DOMCTL_max_vcpus:
case XEN_DOMCTL_scheduler_op:
case XEN_DOMCTL_setdomainhandle:
@@ -354,6 +368,12 @@
case XEN_DOMCTL_set_opt_feature:
case XEN_DOMCTL_assign_device:
case XEN_DOMCTL_subscribe:
+ case XEN_DOMCTL_test_assign_device:
+ case XEN_DOMCTL_deassign_device:
+ case XEN_DOMCTL_bind_pt_irq:
+ case XEN_DOMCTL_unbind_pt_irq:
+ case XEN_DOMCTL_memory_mapping:
+ case XEN_DOMCTL_ioport_mapping:
break;
case XEN_DOMCTL_pin_mem_cacheattr:
return -ENOSYS;
--
yamahata
___
Xen-ia64-devel mailing list
Xen-ia64-devel@lists.xensource.com
http://lists.xensource.com/xen-ia64-devel
RE: [Xen-ia64-devel][PATCH][VTD][dom0] dom0 vtd patch
Isaku Yamahata wrote:
> On Wed, Oct 22, 2008 at 05:43:13PM +0800, Xu, Anthony wrote:
>> dom0/ia64/vtd patch
>>
>> Signed-off-by: Anthony Xu <[EMAIL PROTECTED]>
>
>
> In XEN_DOMCTL_get_device_group case.
>
>
> struct xen_domctl_get_device_group {
> uint32_t machine_bdf; /* IN */
> uint32_t max_sdevs;/* IN */
> uint32_t num_sdevs;/* OUT */
> XEN_GUEST_HANDLE_64(uint32) sdev_array; /* OUT */
> };
>
> XEN_DOMCTL_get_device_group is used by libxc.
> sdev_array needs ot be handled correctly.
I guess you are talking about xencomm,
I don't know how to handle it.
Can you enlighten me?
Thanks,
anthony
>
>
>> dom0/ia64/vtd patch
>>
>> Signed-off-by: Anthony Xu <[EMAIL PROTECTED]>
>>
>>
>> diff -r 16c549b3ba59 arch/ia64/xen/xcom_privcmd.c
>> --- a/arch/ia64/xen/xcom_privcmd.cTue Oct 14 14:58:08 2008 +0100
>> +++ b/arch/ia64/xen/xcom_privcmd.cWed Oct 22 17:15:56 2008 +0800
>> @@ -354,6 +354,14 @@ case XEN_DOMCTL_set_opt_feature:
>> case XEN_DOMCTL_assign_device:
>> case XEN_DOMCTL_subscribe:
>> + case XEN_DOMCTL_get_device_group:
>> + case XEN_DOMCTL_test_assign_device:
>> + case XEN_DOMCTL_deassign_device:
>> + case XEN_DOMCTL_bind_pt_irq:
>> + case XEN_DOMCTL_unbind_pt_irq:
>> + case XEN_DOMCTL_memory_mapping:
>> + case XEN_DOMCTL_ioport_mapping:
>> +
>> break;
>> case XEN_DOMCTL_pin_mem_cacheattr:
>> return -ENOSYS;
___
Xen-ia64-devel mailing list
Xen-ia64-devel@lists.xensource.com
http://lists.xensource.com/xen-ia64-devel
Re: [Xen-ia64-devel][PATCH][VTD][dom0] dom0 vtd patch
On Wed, Oct 22, 2008 at 05:43:13PM +0800, Xu, Anthony wrote:
> dom0/ia64/vtd patch
>
> Signed-off-by: Anthony Xu <[EMAIL PROTECTED]>
In XEN_DOMCTL_get_device_group case.
struct xen_domctl_get_device_group {
uint32_t machine_bdf; /* IN */
uint32_t max_sdevs;/* IN */
uint32_t num_sdevs;/* OUT */
XEN_GUEST_HANDLE_64(uint32) sdev_array; /* OUT */
};
XEN_DOMCTL_get_device_group is used by libxc.
sdev_array needs ot be handled correctly.
> dom0/ia64/vtd patch
>
> Signed-off-by: Anthony Xu <[EMAIL PROTECTED]>
>
>
> diff -r 16c549b3ba59 arch/ia64/xen/xcom_privcmd.c
> --- a/arch/ia64/xen/xcom_privcmd.cTue Oct 14 14:58:08 2008 +0100
> +++ b/arch/ia64/xen/xcom_privcmd.cWed Oct 22 17:15:56 2008 +0800
> @@ -354,6 +354,14 @@
> case XEN_DOMCTL_set_opt_feature:
> case XEN_DOMCTL_assign_device:
> case XEN_DOMCTL_subscribe:
> + case XEN_DOMCTL_get_device_group:
> + case XEN_DOMCTL_test_assign_device:
> + case XEN_DOMCTL_deassign_device:
> + case XEN_DOMCTL_bind_pt_irq:
> + case XEN_DOMCTL_unbind_pt_irq:
> + case XEN_DOMCTL_memory_mapping:
> + case XEN_DOMCTL_ioport_mapping:
> +
> break;
> case XEN_DOMCTL_pin_mem_cacheattr:
> return -ENOSYS;
--
yamahata
___
Xen-ia64-devel mailing list
Xen-ia64-devel@lists.xensource.com
http://lists.xensource.com/xen-ia64-devel
