Re: [PATCH v2 1/3] drm/i915: Avoid lock inversion when pinning to GGTT

[Andi Shyti]

>  err_fence:
> -     if (work)
> -             dma_fence_work_commit_imm(&work->base);
> +     if (work) {
> +             /* don't risk lockdep splat against stop_machine() */

BTW, if we are going to apply the patch, we need a clearer
comment here to explain why do we need to check for concurrent
access.

Andi

> +             if (i915_vma_is_ggtt(vma) &&
> +                 intel_vm_no_concurrent_access_wa(vma->vm->i915))
> +                     dma_fence_work_commit(&work->base);
> +             else
> +                     dma_fence_work_commit_imm(&work->base);
> +     }