Re: [PATCH 0/5] mm,drm/i915,powerpc/spufs: vm_ops->access related fixes

["David Hildenbrand (Arm)" <david@kernel.org>]

On 11/12/26 23:08, David Hildenbrand (Arm) wrote:
> While working on vm_ops->access related fixes in core code for FOLL_ANON
> and VMA access permission + size handling, I stumbled over two driver
> issues. Also, I have to update one selftest that specifies PROT_WRITE
> but actually wants read-access.
> 
> Fixing it all in this series, because it's all intertwined.
> 
> In patch #5 we break a debugging use case for older ROCm versions, where
> the amdgpu driver hacked around it by exploiting broken access checks in
> core-mm.
> 
> Tested as good as I can with selftests. I cannot really test the
> i915 and powerpc/spufs stuff on real hardware.
> 
> Signed-off-by: David Hildenbrand (Arm) <david@kernel.org>
> ---
> Chris Mason (1):
>       mm: do not call vm_ops->access with FOLL_ANON
> 
> David Hildenbrand (Arm) (4):
>       drm/i915: consider vm_pgoff in vm_access()
>       drm/i915/selftests: request read access for mmap access tests
>       powerpc/spufs: consider vm_pgoff in spufs_mem_mmap_access()
>       mm: check permissions and bounds for vm_ops->access
> 
>  arch/powerpc/platforms/cell/spufs/file.c           |  3 ++-
>  drivers/gpu/drm/i915/gem/i915_gem_mman.c           |  4 +++-
>  drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c |  3 ++-
>  mm/memory.c                                        | 28 ++++++++++++++++++++--
>  4 files changed, 33 insertions(+), 5 deletions(-)
> 

We get some test failures regarding error handling: e.g.,
https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_178442v1/shard-glk6/igt@gem_mmap_offset@oob-read.html

It's all about error handling.

Before the correct handling in patch #5, we would pass a size that exceeds the
actual VMA to vm_access(), and that one would return -EIO.

In the test above 8 bytes whereby the VMA only allowed for 4 more bytes.

Now we properly handle that in core code and get a partial read completion on
the 4 bytes.

I think the test should be fixed to expect a partial read completion.

We won't hack around this in core-mm.

-- 
Cheers,

David