[Intel-gfx] [RFC 4/7] drm/i915: Add mechanism for forwarding the data samples to userspace through Gen PMU perf interface

[Chris Wilson]

On Mon, Jun 22, 2015 at 03:25:06PM +0530, sourab.gupta at intel.com wrote:
> +void i915_gen_pmu_wait_gpu(struct drm_i915_private *dev_priv)
> +{
> +	struct drm_i915_ts_queue_header *hdr =
> +		(struct drm_i915_ts_queue_header *)
> +		dev_priv->gen_pmu.buffer.addr;
> +	struct drm_i915_ts_node *first_node, *node;
> +	int head, tail, num_nodes, ret;
> +	struct drm_i915_gem_request *req;
> +
> +	first_node = (struct drm_i915_ts_node *)
> +			((char *)hdr + hdr->data_offset);
> +	num_nodes = (hdr->size_in_bytes - hdr->data_offset) /
> +			sizeof(*node);
> +
> +	tail = hdr->node_count;
> +	head = dev_priv->gen_pmu.buffer.head;
> +
> +	/* wait for all requests to complete*/
> +	while ((head % num_nodes) != (tail % num_nodes)) {
> +		node = &first_node[head % num_nodes];
> +		req = node->node_info.req;
> +		if (req) {
> +			if (!i915_gem_request_completed(req, true)) {
> +				ret = i915_wait_request(req);
> +				if (ret)
> +					DRM_DEBUG_DRIVER(
> +					"gen pmu: failed to wait\n");
> +			}
> +			i915_gem_request_assign(&node->node_info.req, NULL);
> +		}
> +		head++;
> +	}

You can, sorry must, rewrite this to avoid the struct mutex, and handle
the *likely* errors correctly, and even remove the superfluous
i915_gem_request_completed.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre