drm/vc4: Expose performance counters to userspace

The V3D engine has various hardware counters which might be interesting
to userspace performance analysis tools.

Expose new ioctls to create/destroy a performance monitor object and
query the counter values of this perfmance monitor.

Note that a perfomance monitor is given an ID that is only valid on the
file descriptor it has been allocated from. A performance monitor can be
attached to a CL submission and the driver will enable HW counters for
this request and update the performance monitor values at the end of the
job.

Signed-off-by: Boris Brezillon <boris.brezillon@free-electrons.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Eric Anholt <eric@anholt.net>
Link: https://patchwork.freedesktop.org/patch/msgid/20180112090926.12538-1-boris.brezillon@free-electrons.com

1 files changed, 37 insertions, 3 deletions

diff --git a/drivers/gpu/drm/vc4/vc4_irq.c b/drivers/gpu/drm/vc4/vc4_irq.c
index 3dd62d75f531..4cd2ccfe15f4 100644
--- a/drivers/gpu/drm/vc4/vc4_irq.c
+++ b/drivers/gpu/drm/vc4/vc4_irq.c
@@ -104,13 +104,20 @@ static void
 vc4_irq_finish_bin_job(struct drm_device *dev)
 {
 	struct vc4_dev *vc4 = to_vc4_dev(dev);
-	struct vc4_exec_info *exec = vc4_first_bin_job(vc4);
+	struct vc4_exec_info *next, *exec = vc4_first_bin_job(vc4);
 
 	if (!exec)
 		return;
 
 	vc4_move_job_to_render(dev, exec);
-	vc4_submit_next_bin_job(dev);
+	next = vc4_first_bin_job(vc4);
+
+	/* Only submit the next job in the bin list if it matches the perfmon
+	 * attached to the one that just finished (or if both jobs don't have
+	 * perfmon attached to them).
+	 */
+	if (next && next->perfmon == exec->perfmon)
+		vc4_submit_next_bin_job(dev);
 }
 
 static void
@@ -122,6 +129,10 @@ vc4_cancel_bin_job(struct drm_device *dev)
 	if (!exec)
 		return;
 
+	/* Stop the perfmon so that the next bin job can be started. */
+	if (exec->perfmon)
+		vc4_perfmon_stop(vc4, exec->perfmon, false);
+
 	list_move_tail(&exec->head, &vc4->bin_job_list);
 	vc4_submit_next_bin_job(dev);
 }
@@ -131,18 +142,41 @@ vc4_irq_finish_render_job(struct drm_device *dev)
 {
 	struct vc4_dev *vc4 = to_vc4_dev(dev);
 	struct vc4_exec_info *exec = vc4_first_render_job(vc4);
+	struct vc4_exec_info *nextbin, *nextrender;
 
 	if (!exec)
 		return;
 
 	vc4->finished_seqno++;
 	list_move_tail(&exec->head, &vc4->job_done_list);
+
+	nextbin = vc4_first_bin_job(vc4);
+	nextrender = vc4_first_render_job(vc4);
+
+	/* Only stop the perfmon if following jobs in the queue don't expect it
+	 * to be enabled.
+	 */
+	if (exec->perfmon && !nextrender &&
+	    (!nextbin || nextbin->perfmon != exec->perfmon))
+		vc4_perfmon_stop(vc4, exec->perfmon, true);
+
+	/* If there's a render job waiting, start it. If this is not the case
+	 * we may have to unblock the binner if it's been stalled because of
+	 * perfmon (this can be checked by comparing the perfmon attached to
+	 * the finished renderjob to the one attached to the next bin job: if
+	 * they don't match, this means the binner is stalled and should be
+	 * restarted).
+	 */
+	if (nextrender)
+		vc4_submit_next_render_job(dev);
+	else if (nextbin && nextbin->perfmon != exec->perfmon)
+		vc4_submit_next_bin_job(dev);
+
 	if (exec->fence) {
 		dma_fence_signal_locked(exec->fence);
 		dma_fence_put(exec->fence);
 		exec->fence = NULL;
 	}
-	vc4_submit_next_render_job(dev);
 
 	wake_up_all(&vc4->job_wait_queue);
 	schedule_work(&vc4->job_done_work);