Re: [PATCH v6 2/6] iommufd: Add IOMMU_HWPT_INVALIDATE

[Nicolin Chen <nicolinc@nvidia.com>]

On Tue, Nov 21, 2023 at 01:02:49PM +0800, Baolu Lu wrote:
> On 11/17/23 9:07 PM, Yi Liu wrote:
> > In nested translation, the stage-1 page table is user-managed but cached
> > by the IOMMU hardware, so an update on present page table entries in the
> > stage-1 page table should be followed with a cache invalidation.
> > 
> > Add an IOMMU_HWPT_INVALIDATE ioctl to support such a cache invalidation.
> > It takes hwpt_id to specify the iommu_domain, and a multi-entry array to
> > support multiple invalidation requests in one ioctl.
> > 
> > Check cache_invalidate_user op in the iommufd_hw_pagetable_alloc_nested,
> > since all nested domains need that.
> > 
> > Co-developed-by: Nicolin Chen<nicolinc@nvidia.com>
> > Signed-off-by: Nicolin Chen<nicolinc@nvidia.com>
> > Signed-off-by: Yi Liu<yi.l.liu@intel.com>
> > ---
> >   drivers/iommu/iommufd/hw_pagetable.c    | 35 +++++++++++++++++++++++++
> >   drivers/iommu/iommufd/iommufd_private.h |  9 +++++++
> >   drivers/iommu/iommufd/main.c            |  3 +++
> >   include/uapi/linux/iommufd.h            | 35 +++++++++++++++++++++++++
> >   4 files changed, 82 insertions(+)
> > 
> > diff --git a/drivers/iommu/iommufd/hw_pagetable.c b/drivers/iommu/iommufd/hw_pagetable.c
> > index 2abbeafdbd22..367459d92f69 100644
> > --- a/drivers/iommu/iommufd/hw_pagetable.c
> > +++ b/drivers/iommu/iommufd/hw_pagetable.c
> > @@ -238,6 +238,11 @@ iommufd_hwpt_nested_alloc(struct iommufd_ctx *ictx,
> >               rc = -EINVAL;
> >               goto out_abort;
> >       }
> > +     /* Driver is buggy by missing cache_invalidate_user in domain_ops */
> > +     if (WARN_ON_ONCE(!hwpt->domain->ops->cache_invalidate_user)) {
> > +             rc = -EINVAL;
> > +             goto out_abort;
> > +     }
> >       return hwpt_nested;
> 
> The WARN message here may cause kernel regression when users bisect
> issues. Till this patch, there are no drivers support the
> cache_invalidation_user callback yet.

Ah, this is an unintended consequence from our uAPI bisect to
merge the nesting alloc first...

Would removing the WARN_ON_ONCE be okay? Although having this
WARN is actually the point here...

Thanks
Nic