[Intel-gfx] [RFC PATCH v2] drm/i915: Never allow userptr into the new mapping types

[Chris Wilson]

Quoting Janusz Krzysztofik (2020-01-31 15:31:23)
> diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
> index 9ddcf17230e6..334a578ce85f 100644
> --- a/drivers/gpu/drm/i915/i915_gem.c
> +++ b/drivers/gpu/drm/i915/i915_gem.c
> @@ -923,7 +923,7 @@ i915_gem_object_ggtt_pin(struct drm_i915_gem_object *obj,
>         struct i915_vma *vma;
>         int ret;
>  
> -       if (i915_gem_object_never_bind_ggtt(obj))
> +       if (i915_gem_object_never_mmap(obj))
>                 return ERR_PTR(-ENODEV);

I think never_mmap() is more strictly correct than bind_ggtt() and it
was a mistake to prevent ggtt_pin() itself. Though I'll need to check
the revocation to make sure.

Right we should only revoke_mmap if i915_vma_has_userfault, which will
never be the case.
-Chris