Rob Herring <robh@kernel.org> writes:

> This adds the initial driver for panfrost which supports Arm Mali
> Midgard and Bifrost family of GPUs. Currently, only the T860 and
> T760 Midgard GPUs have been tested.

> +static int panfrost_ioctl_get_bo_offset(struct drm_device *dev, void *data,
> +			    struct drm_file *file_priv)
> +{
> +	struct drm_panfrost_get_bo_offset *args = data;
> +	struct drm_gem_object *gem_obj;
> +	struct panfrost_gem_object *bo;
> +

Missing check for pad == 0.  With that fixed,

Reviewed-by: Eric Anholt <eric@anholt.net>

> +	gem_obj = drm_gem_object_lookup(file_priv, args->handle);
> +	if (!gem_obj) {
> +		DRM_DEBUG("Failed to look up GEM BO %d\n", args->handle);
> +		return -ENOENT;
> +	}
> +	bo = to_panfrost_bo(gem_obj);
> +
> +	args->offset = bo->node.start << PAGE_SHIFT;
> +
> +	drm_gem_object_put_unlocked(gem_obj);
> +	return 0;
> +}


> +static void panfrost_job_timedout(struct drm_sched_job *sched_job)
> +{
> +	struct panfrost_job *job = to_panfrost_job(sched_job);
> +	struct panfrost_device *pfdev = job->pfdev;
> +	int js = panfrost_job_get_slot(job);
> +	int i;
> +
> +	/*
> +	 * If the GPU managed to complete this jobs fence, the timeout is
> +	 * spurious. Bail out.
> +	 */
> +	if (dma_fence_is_signaled(job->done_fence))
> +		return;

Note: The scheduler calls cancel_delayed_work_sync() in
s_job->finish_work, so this is just reducing the race for the job
successfully completing near the timeout but finish_work being in the
workqueue across the timeout boundary.  I dropped it from v3d.