[PATCH v2 25/25] iommu: Convert remaining simple drivers to domain_alloc_paging()

Jason Gunthorpe jgg at nvidia.com
Fri Jun 9 12:26:59 PDT 2023


On Thu, Jun 01, 2023 at 08:47:28PM +0100, Robin Murphy wrote:
> > diff --git a/drivers/iommu/fsl_pamu_domain.c b/drivers/iommu/fsl_pamu_domain.c
> > index ca4f5ebf028783..8d5d6a3acf9dfd 100644
> > --- a/drivers/iommu/fsl_pamu_domain.c
> > +++ b/drivers/iommu/fsl_pamu_domain.c
> > @@ -192,13 +192,10 @@ static void fsl_pamu_domain_free(struct iommu_domain *domain)
> >   	kmem_cache_free(fsl_pamu_domain_cache, dma_domain);
> >   }
> > -static struct iommu_domain *fsl_pamu_domain_alloc(unsigned type)
> > +static struct iommu_domain *fsl_pamu_domain_alloc_paging(struct device *dev)
> 
> This isn't a paging domain - it doesn't support map/unmap, and AFAICT all it
> has ever been intended to do is "isolate" accesses to within an aperture
> which is never set to anything less than the entire physical address space
> :/
> 
> I hate to imagine what the VFIO userspace applications looked
> like...

After looking at this some more I don't think there is any VFIO
userspace..

There is a VFIO bus driver drivers/vfio/fsl-mc/ but from what I can
tell that is for the ARM version of this platform (from 2020, not
2014) and it doesn't use this driver.

So, really, the only thing this driver does is setup the identity
domain at boot and do something special for
drivers/soc/fsl/qbman/qman_portal.c :\

I wonder if we should just delete it, any chance the power systems
need the code to switch to identity at boot?

Michael do you have an opinion?

Jason



More information about the Linux-mediatek mailing list