[PATCH] iommu/io-pgtable-arm: Remove split on unmap behavior

[Will Deacon]

Hi Jason,

On Fri, Oct 18, 2024 at 02:19:26PM -0300, Jason Gunthorpe wrote:
> Of the page table implementations (AMD v1/2, VT-D SS, ARM32, DART)
> arm_lpae is unique in how it handles partial unmap of large IOPTEs.
> 
> All other drivers will unmap the large IOPTE and return it's length.  For
> example if a 2M IOPTE is present and the first 4K is requested to be
> unmapped then unmap will remove the whole 2M and report 2M as the result.
> 
> arm_lpae instead replaces the IOPTE with a table of smaller IOPTEs, unmaps
> the 4K and returns 4k. This is actually an illegal/non-hitless operation
> on at least SMMUv3 because of the BBM level 0 rules.
> 
> Long ago VFIO could trigger a path like this, today I know of no user of
> this functionality.
> 
> Given it doesn't work fully correctly on SMMUv3 and would create
> portability problems if any user depends on it, remove the unique support
> in arm_lpae and align with the expected iommu interface.
> 
> Outside the iommu users, this will potentially effect io_pgtable users of
> ARM_32_LPAE_S1, ARM_32_LPAE_S2, ARM_64_LPAE_S1, ARM_64_LPAE_S2, and
> ARM_MALI_LPAE formats.
> 
> Cc: Boris Brezillon <boris.brezillon at collabora.com>
> Cc: Steven Price <steven.price at arm.com>
> Cc: Liviu Dudau <liviu.dudau at arm.com>
> Cc: dri-devel at lists.freedesktop.org
> Signed-off-by: Jason Gunthorpe <jgg at nvidia.com>
> ---
>  drivers/iommu/io-pgtable-arm.c | 72 +++-------------------------------
>  1 file changed, 6 insertions(+), 66 deletions(-)

I'd love to drop this, but I'm sure it was needed when I added it :/

My recollection is hazy, but I seem to remember VFIO using the largest
page sizes in the IOMMU 'pgsize_bitmap' for map() requests but then
using the smallest page size for unmap() requests, so you'd end up
cracking block mappings when tearing down a VM with assigne devices.

Is this what you're referring to when you say?

  > Long ago VFIO could trigger a path like this, today I know of no user of
  > this functionality.

If so, please can you provide a reference to the patch that moved VFIO
off that problematic behaviour?

Thanks!

Will