On Thu, 2018-08-30 at 10:52 +0100, Chris Wilson wrote:

> +static int igt_gem_suspend(void *arg)
[...]
> +     if (i915_gem_suspend(i915)) {
> +             pr_err("i915_gem_suspend failed\n");
> +             err = -EINVAL;
> +             goto out;
> +     }
> +
> +     i915_gem_suspend_late(i915);

Shouldn't we also do i915_gem_suspend_gtt_mappings after
i915_gem_suspend (usually it's in drm_suspend, which is called by both
pm_suspend and pm_freeze)? It's later restored in i915_gem_resume.

> +static int igt_gem_hibernate(void *arg)

Most of this function is same as igt_gem_suspend and that is probably
not going to change since both suspend and hibernation normally go
through pm_prepare and pm_resume_early/resume, maybe some prepare and
resume helpers for readability?

- Jakub

Attachment: smime.p7s
Description: S/MIME cryptographic signature

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

Reply via email to