Re: [Xen-devel] [RFC for-4.5 06/12] xen/passthrough: rework dom0_pvh_reqs to use it also on ARM

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 10.02.14 at 18:42, Julien Grall <julien.grall@xxxxxxxxxx> wrote:
> On 02/10/2014 04:35 PM, Jan Beulich wrote:
>>>>> On 10.02.14 at 17:10, Julien Grall <julien.grall@xxxxxxxxxx> wrote:
>>> On 02/07/2014 05:43 PM, Julien Grall wrote:
>>>> DOM0 on ARM will have the same requirements as DOM0 PVH when iommu is 
>>> enabled.
>>>> Both PVH and ARM guest has paging mode translate enabled, so Xen can use it
>>>> to know if it needs to check the requirements.
>>>>
>>>> Rename the function and remove "pvh" word in the commit message.
>>>>
>>>> Signed-off-by: Julien Grall <julien.grall@xxxxxxxxxx>
>>>> Cc: Xiantao Zhang <xiantao.zhang@xxxxxxxxx>
>>>> Cc: Jan Beulich <jbeulich@xxxxxxxx>
>>>> ---
>>>>  xen/drivers/passthrough/iommu.c |   14 +++++++++-----
>>>>  1 file changed, 9 insertions(+), 5 deletions(-)
>>>>
>>>> diff --git a/xen/drivers/passthrough/iommu.c 
>>>> b/xen/drivers/passthrough/iommu.c
>>>> index 19b0e23..26a5d91 100644
>>>> --- a/xen/drivers/passthrough/iommu.c
>>>> +++ b/xen/drivers/passthrough/iommu.c
>>>> @@ -130,13 +130,18 @@ int iommu_domain_init(struct domain *d)
>>>>      return hd->platform_ops->init(d);
>>>>  }
>>>>  
>>>> -static __init void check_dom0_pvh_reqs(struct domain *d)
>>>> +static __init void check_dom0_reqs(struct domain *d)
>>>>  {
>>>> +    if ( !paging_mode_translate(d) )
>>>> +        return;
>>>> +
>>>>      if ( !iommu_enabled )
>>>> -        panic("Presently, iommu must be enabled for pvh dom0\n");
>>>> +        panic("Presently, iommu must be enabled to use dom0 with 
>>>> translate 
>>> "
>>>> +              "paging mode\n");
>>>
>>> Hmmm... this change is wrong. I forgot that iommu doesn't exist on some
>>> ARM platform (for instance Arndale).
>>>
>>> Do we really this check for PVH? If yes, I will add replace the check
>>> by: is_pvh_domain(d) && !iommu_enabled.
>> 
>> Of course we need it: How would PVH Dom0 be able to do any kind
>> of DMA without an IOMMU?
> 
> Right, on ARM we have the 1:1 memory mapping to avoid this issue.
> 
> I will fix it. Can I keep your ack on this patch?

For this simple an adjustment - sure.

Jan


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