Re: [Xen-devel] [PATCH] xen/arm: Let the IOMMU be accessible by Dom0 if forcibly disabled in Xen

[Julien Grall <julien.grall@xxxxxxx>]

Hi,

On 08/08/2019 14:36, Oleksandr wrote:
> Hi
>
>
> > > > > > On Thu, Aug 08, 2019 at 01:53:23PM +0300, Oleksandr Tyshchenko wrote:
> > > > > > > From: Oleksandr Tyshchenko <oleksandr_tyshchenko@xxxxxxxx>
> > > > > > >
> > > > > > > Don't skip IOMMU nodes when creating DT for Dom0 if IOMMU has been
> > > > > > > forcibly disabled in bootargs (e.g. "iommu=0") in order to let
> > > > > > > the IOMMU be accessible by DOM0.
> > > > I don't think your code is doing what you expect... If iommu=0, then Xen 
> > > > will not lookup for IOMMUs (iommu_hardware_setup() will not be called). So 
> > > > none of the device will have DEVICE_IOMMU set and hence they are already 
> > > > given to dom0.
> > > > But I think it is wrong to give the IOMMUs to Dom0 when iommu=0. This is not 
> > > > the goal of this option. If you want to passthrough the IOMMU to Dom0, then 
> > > > you should use the parameter iommu_hwdom_passthrough.
> > > > However, I agree with Roger that giving the IOMMU to dom0 is a pretty bad 
> > > > idea. So this should be fixed.
> > >
> > > I fully agree with the arguments provided that it is a bad idea. So, please 
> > > consider that patch as not relevant.
> > >
> > > But, I am not sure I follow the last sentence:
> > >
> > >  >>> If iommu=0, then Xen will not lookup for IOMMUs (iommu_hardware_setup() 
> > > will not be called). So none of the device will have DEVICE_IOMMU set and 
> > > hence they are already given to dom0.
> > > I can see that devices have DEVICE_IOMMU set. Although, the IOMMU driver is 
> > > not in use, it is present and compatible matches. So, even if iommu=0, the 
> > > IOMMU devices are not given to Dom0, because of skipped. Or I missed something?
> > I can't see how iommu_hardware_setup() can be called on staging when iommu=0 
> > as this is protected by a if ( iommu_enable ).
> > Can you please give a stack trace how this is called and the version you use? 
> > WARN() should do it for you.
>
> iommu_hardware_setup() is not called. But, devices have DEVICE_IOMMU set, even 
> if "iommu=0". I am based on "7d1460c xen/arm: optee: fix compilation with GCC 
> 4.8" + Stefano's reserved-memory series + my IPMMU series.
Hmmm, what you mean by set is "device_get_class() return DEVICE_IOMMU". This is 
were I got confused and I mixed up with dt_device_set_{protected, used_by}() 
function.
device_get_class() will just lookup for a match and return the class associated. 
So you are right and the node will be skipped in that case. Nothing to modify in 
the current code.
Cheers,

--
Julien Grall

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel