drm/i915/bdw: Emission of requests with logical rings

On a previous iteration of this patch, I created an Execlists
version of __i915_add_request and asbtracted it away as a
vfunc. Daniel Vetter wondered then why that was needed:

"with the clean split in command submission I expect every
function to know wether it'll submit to an lrc (everything in
intel_lrc.c) or wether it'll submit to a legacy ring (existing
code), so I don't see a need for an add_request vfunc."

The honest, hairy truth is that this patch is the glue keeping
the whole logical ring puzzle together:

- i915_add_request is used by intel_ring_idle, which in turn is
  used by i915_gpu_idle, which in turn is used in several places
  inside the eviction and gtt codes.
- Also, it is used by i915_gem_check_olr, which is littered all
  over i915_gem.c
- ...

If I were to duplicate all the code that directly or indirectly
uses __i915_add_request, I'll end up creating a separate driver.

To show the differences between the existing legacy version and
the new Execlists one, this time I have special-cased
__i915_add_request instead of adding an add_request vfunc. I
hope this helps to untangle this Gordian knot.

Signed-off-by: Oscar Mateo <oscar.mateo@intel.com>
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
[danvet: Adjust to ringbuf->FIXME_lrc_ctx per the discussion with
Thomas Daniel.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drivers/gpu/drm/i915/i915_gem.c

@@ -2311,10 +2311,21 @@ int __i915_add_request(struct intel_engine_cs *ring,
 {
 	struct drm_i915_private *dev_priv = ring->dev->dev_private;
 	struct drm_i915_gem_request *request;
+	struct intel_ringbuffer *ringbuf;
 	u32 request_ring_position, request_start;
 	int ret;
 
-	request_start = intel_ring_get_tail(ring->buffer);
+	request = ring->preallocated_lazy_request;
+	if (WARN_ON(request == NULL))
+		return -ENOMEM;
+
+	if (i915.enable_execlists) {
+		struct intel_context *ctx = request->ctx;
+		ringbuf = ctx->engine[ring->id].ringbuf;
+	} else
+		ringbuf = ring->buffer;
+
+	request_start = intel_ring_get_tail(ringbuf);
 	/*
 	 * Emit any outstanding flushes - execbuf can fail to emit the flush
 	 * after having emitted the batchbuffer command. Hence we need to fix
@@ -2322,24 +2333,32 @@ int __i915_add_request(struct intel_engine_cs *ring,
 	 * is that the flush _must_ happen before the next request, no matter
 	 * what.
 	 */
-	ret = intel_ring_flush_all_caches(ring);
+	if (i915.enable_execlists) {
-	if (ret)
+		ret = logical_ring_flush_all_caches(ringbuf);
-		return ret;
+		if (ret)
-
+			return ret;
-	request = ring->preallocated_lazy_request;
+	} else {
-	if (WARN_ON(request == NULL))
+		ret = intel_ring_flush_all_caches(ring);
-		return -ENOMEM;
+		if (ret)
+			return ret;
+	}
 
 	/* Record the position of the start of the request so that
 	 * should we detect the updated seqno part-way through the
 	 * GPU processing the request, we never over-estimate the
 	 * position of the head.
 	 */
-	request_ring_position = intel_ring_get_tail(ring->buffer);
+	request_ring_position = intel_ring_get_tail(ringbuf);
 
-	ret = ring->add_request(ring);
+	if (i915.enable_execlists) {
-	if (ret)
+		ret = ring->emit_request(ringbuf);
-		return ret;
+		if (ret)
+			return ret;
+	} else {
+		ret = ring->add_request(ring);
+		if (ret)
+			return ret;
+	}
 
 	request->seqno = intel_ring_get_seqno(ring);
 	request->ring = ring;
@@ -2354,12 +2373,14 @@ int __i915_add_request(struct intel_engine_cs *ring,
 	 */
 	request->batch_obj = obj;
 
-	/* Hold a reference to the current context so that we can inspect
+	if (!i915.enable_execlists) {
-	 * it later in case a hangcheck error event fires.
+		/* Hold a reference to the current context so that we can inspect
-	 */
+		 * it later in case a hangcheck error event fires.
-	request->ctx = ring->last_context;
+		 */
-	if (request->ctx)
+		request->ctx = ring->last_context;
-		i915_gem_context_reference(request->ctx);
+		if (request->ctx)
+			i915_gem_context_reference(request->ctx);
+	}
 
 	request->emitted_jiffies = jiffies;
 	list_add_tail(&request->list, &ring->request_list);
@@ -2614,6 +2635,7 @@ i915_gem_retire_requests_ring(struct intel_engine_cs *ring)
 
 	while (!list_empty(&ring->request_list)) {
 		struct drm_i915_gem_request *request;
+		struct intel_ringbuffer *ringbuf;
 
 		request = list_first_entry(&ring->request_list,
 					   struct drm_i915_gem_request,
@@ -2623,12 +2645,24 @@ i915_gem_retire_requests_ring(struct intel_engine_cs *ring)
 			break;
 
 		trace_i915_gem_request_retire(ring, request->seqno);
+
+		/* This is one of the few common intersection points
+		 * between legacy ringbuffer submission and execlists:
+		 * we need to tell them apart in order to find the correct
+		 * ringbuffer to which the request belongs to.
+		 */
+		if (i915.enable_execlists) {
+			struct intel_context *ctx = request->ctx;
+			ringbuf = ctx->engine[ring->id].ringbuf;
+		} else
+			ringbuf = ring->buffer;
+
 		/* We know the GPU must have read the request to have
 		 * sent us the seqno + interrupt, so use the position
 		 * of tail of the request to update the last known position
 		 * of the GPU head.
 		 */
-		ring->buffer->last_retired_head = request->tail;
+		ringbuf->last_retired_head = request->tail;
 
 		i915_gem_free_request(request);
 	}

drivers/gpu/drm/i915/intel_lrc.c

@@ -252,6 +252,22 @@ void intel_logical_ring_stop(struct intel_engine_cs *ring)
 	I915_WRITE_MODE(ring, _MASKED_BIT_DISABLE(STOP_RING));
 }
 
+int logical_ring_flush_all_caches(struct intel_ringbuffer *ringbuf)
+{
+	struct intel_engine_cs *ring = ringbuf->ring;
+	int ret;
+
+	if (!ring->gpu_caches_dirty)
+		return 0;
+
+	ret = ring->emit_flush(ringbuf, 0, I915_GEM_GPU_DOMAINS);
+	if (ret)
+		return ret;
+
+	ring->gpu_caches_dirty = false;
+	return 0;
+}
+
 void intel_logical_ring_advance_and_submit(struct intel_ringbuffer *ringbuf)
 {
 	intel_logical_ring_advance(ringbuf);
@@ -262,7 +278,8 @@ void intel_logical_ring_advance_and_submit(struct intel_ringbuffer *ringbuf)
 	/* TODO: how to submit a context to the ELSP is not here yet */
 }
 
-static int logical_ring_alloc_seqno(struct intel_engine_cs *ring)
+static int logical_ring_alloc_seqno(struct intel_engine_cs *ring,
+				    struct intel_context *ctx)
 {
 	if (ring->outstanding_lazy_seqno)
 		return 0;
@@ -274,6 +291,13 @@ static int logical_ring_alloc_seqno(struct intel_engine_cs *ring)
 		if (request == NULL)
 			return -ENOMEM;
 
+		/* Hold a reference to the context this request belongs to
+		 * (we will need it when the time comes to emit/retire the
+		 * request).
+		 */
+		request->ctx = ctx;
+		i915_gem_context_reference(request->ctx);
+
 		ring->preallocated_lazy_request = request;
 	}
 
@@ -312,8 +336,6 @@ static int logical_ring_wait_request(struct intel_ringbuffer *ringbuf,
 	if (ret)
 		return ret;
 
-	/* TODO: make sure we update the right ringbuffer's last_retired_head
-	 * when retiring requests */
 	i915_gem_retire_requests_ring(ring);
 	ringbuf->head = ringbuf->last_retired_head;
 	ringbuf->last_retired_head = -1;
@@ -433,7 +455,7 @@ int intel_logical_ring_begin(struct intel_ringbuffer *ringbuf, int num_dwords)
 		return ret;
 
 	/* Preallocate the olr before touching the ring */
-	ret = logical_ring_alloc_seqno(ring);
+	ret = logical_ring_alloc_seqno(ring, ringbuf->FIXME_lrc_ctx);
 	if (ret)
 		return ret;
 

drivers/gpu/drm/i915/intel_lrc.h

@@ -29,6 +29,7 @@ void intel_logical_ring_stop(struct intel_engine_cs *ring);
 void intel_logical_ring_cleanup(struct intel_engine_cs *ring);
 int intel_logical_rings_init(struct drm_device *dev);
 
+int logical_ring_flush_all_caches(struct intel_ringbuffer *ringbuf);
 void intel_logical_ring_advance_and_submit(struct intel_ringbuffer *ringbuf);
 static inline void intel_logical_ring_advance(struct intel_ringbuffer *ringbuf)
 {