Re: [PATCH] iommu: fix amd_iommu=force_isolation

[Yu Zhao]

On Wed, Dec 05, 2018 at 05:09:55PM +0100, Joerg Roedel wrote:
> On Tue, Dec 04, 2018 at 03:37:16PM -0700, Yu Zhao wrote:
> > The parameter is still there but it's ignored. We need to check its
> > value before deciding to go into passthrough mode for AMD IOMMU.
> > 
> > Fixes: aafd8ba0ca74 ("iommu/amd: Implement add_device and remove_device")
> > 
> > Signed-off-by: Yu Zhao <yuzhao@xxxxxxxxxx>
> > ---
> >  drivers/iommu/amd_iommu.c | 3 ++-
> >  1 file changed, 2 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/iommu/amd_iommu.c b/drivers/iommu/amd_iommu.c
> > index 1167ff0416cf..3e4219e6cff0 100644
> > --- a/drivers/iommu/amd_iommu.c
> > +++ b/drivers/iommu/amd_iommu.c
> > @@ -2195,7 +2195,8 @@ static int amd_iommu_add_device(struct device *dev)
> >  
> >  	BUG_ON(!dev_data);
> >  
> > -	if (iommu_pass_through || dev_data->iommu_v2)
> > +	if (iommu_pass_through ||
> > +	    (!amd_iommu_force_isolation && dev_data->iommu_v2))
> 
> This breaks the iommu_v2 use-case, as it needs a direct mapping for the
> devices that support it.

Actually this is what we want. We occasionally use this parameter
to isolate v2 device when debugging memory corruption that we
suspect is caused by DMA. It helped us before. Now with the patch,
we caught another v2 device doing DMA write to memory areas that
it's not supposed to.

> I think the force_isolation parameter does not make sense anymore today
> and should be removed.

Could you please suggest an alternative? Otherwise, we won't be able
to force v2 device out of direct mapping for the reason mentioned
above.

Thank you.