Re: [PATCH v3 1/7] drm/xe/mmio_gem: forbid VMA split

[Matthew Auld <matthew.auld@intel.com>]

On 23/07/2026 17:18, Ilia Levi wrote:
> The fault handler assumes it always operates on a VMA spanning the entire
> GEM object. This does not hold when the VMA has been split, e.g. by a
> partial munmap or mprotect. In that case the handler may map wrong
> physical pages or cause SIGBUS.
> 
> Handle this by forbidding VMA split, as partial unmaps are not deemed
> useful for MMIO GEMs.
> 
> Suggested-by: Matthew Auld <matthew.auld@intel.com>
> Signed-off-by: Ilia Levi <ilia.levi@intel.com>
> Fixes: 1ffcf8b8ae8a ("drm/xe: Support for mmap-ing mmio regions")

Reviewed-by: Matthew Auld <matthew.auld@intel.com>

> ---
>   drivers/gpu/drm/xe/xe_mmio_gem.c | 10 ++++++++++
>   1 file changed, 10 insertions(+)
> 
> diff --git a/drivers/gpu/drm/xe/xe_mmio_gem.c b/drivers/gpu/drm/xe/xe_mmio_gem.c
> index 8c803ef233cc..f15a6a84af15 100644
> --- a/drivers/gpu/drm/xe/xe_mmio_gem.c
> +++ b/drivers/gpu/drm/xe/xe_mmio_gem.c
> @@ -39,10 +39,20 @@ struct xe_mmio_gem {
>   	phys_addr_t phys_addr;
>   };
>   
> +static int xe_mmio_gem_vm_may_split(struct vm_area_struct *area, unsigned long addr)
> +{
> +	/*
> +	 * Forbid splitting. Together with VM_DONTEXPAND, this keeps the VMA
> +	 * matching the GEM object exactly.
> +	 */
> +	return -EINVAL;
> +}
> +
>   static const struct vm_operations_struct vm_ops = {
>   	.open = drm_gem_vm_open,
>   	.close = drm_gem_vm_close,
>   	.fault = xe_mmio_gem_vm_fault,
> +	.may_split = xe_mmio_gem_vm_may_split,
>   };
>   
>   static const struct drm_gem_object_funcs xe_mmio_gem_funcs = {