drm/i915/bdw: Two-stage execlist submit process

Context switch (and execlist submission) should happen only when
other contexts are not active, otherwise pre-emption occurs.

To assure this, we place context switch requests in a queue and those
request are later consumed when the right context switch interrupt is
received (still TODO).

v2: Use a spinlock, do not remove the requests on unqueue (wait for
context switch completion).

Signed-off-by: Thomas Daniel <thomas.daniel@intel.com>

v3: Several rebases and code changes. Use unique ID.

v4:
- Move the queue/lock init to the late ring initialization.
- Damien's kmalloc review comments: check return, use sizeof(*req),
do not cast.

v5:
- Do not reuse drm_i915_gem_request. Instead, create our own.
- New namespace.

Signed-off-by: Michel Thierry <michel.thierry@intel.com> (v1)
Signed-off-by: Oscar Mateo <oscar.mateo@intel.com> (v2-v5)
Reviewed-by: Damien Lespiau <damien.lespiau@intel.com>
[davnet: Checkpatch + wash-up s/BUG_ON/WARN_ON/.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

drivers/gpu/drm/i915/intel_lrc.c

@@ -126,7 +126,8 @@ static uint64_t execlists_ctx_descriptor(struct drm_i915_gem_object *ctx_obj)
 {
 	uint64_t desc;
 	uint64_t lrca = i915_gem_obj_ggtt_offset(ctx_obj);
-	BUG_ON(lrca & 0xFFFFFFFF00000FFFULL);
+
+	WARN_ON(lrca & 0xFFFFFFFF00000FFFULL);
 
 	desc = GEN8_CTX_VALID;
 	desc |= LEGACY_CONTEXT << GEN8_CTX_MODE_SHIFT;
@@ -202,14 +203,14 @@ static int execlists_submit_context(struct intel_engine_cs *ring,
 
 	ctx_obj0 = to0->engine[ring->id].state;
 	BUG_ON(!ctx_obj0);
-	BUG_ON(!i915_gem_obj_is_pinned(ctx_obj0));
+	WARN_ON(!i915_gem_obj_is_pinned(ctx_obj0));
 
 	execlists_ctx_write_tail(ctx_obj0, tail0);
 
 	if (to1) {
 		ctx_obj1 = to1->engine[ring->id].state;
 		BUG_ON(!ctx_obj1);
-		BUG_ON(!i915_gem_obj_is_pinned(ctx_obj1));
+		WARN_ON(!i915_gem_obj_is_pinned(ctx_obj1));
 
 		execlists_ctx_write_tail(ctx_obj1, tail1);
 	}
@@ -219,6 +220,65 @@ static int execlists_submit_context(struct intel_engine_cs *ring,
 	return 0;
 }
 
+static void execlists_context_unqueue(struct intel_engine_cs *ring)
+{
+	struct intel_ctx_submit_request *req0 = NULL, *req1 = NULL;
+	struct intel_ctx_submit_request *cursor = NULL, *tmp = NULL;
+
+	if (list_empty(&ring->execlist_queue))
+		return;
+
+	/* Try to read in pairs */
+	list_for_each_entry_safe(cursor, tmp, &ring->execlist_queue,
+				 execlist_link) {
+		if (!req0) {
+			req0 = cursor;
+		} else if (req0->ctx == cursor->ctx) {
+			/* Same ctx: ignore first request, as second request
+			 * will update tail past first request's workload */
+			list_del(&req0->execlist_link);
+			i915_gem_context_unreference(req0->ctx);
+			kfree(req0);
+			req0 = cursor;
+		} else {
+			req1 = cursor;
+			break;
+		}
+	}
+
+	WARN_ON(execlists_submit_context(ring, req0->ctx, req0->tail,
+					 req1 ? req1->ctx : NULL,
+					 req1 ? req1->tail : 0));
+}
+
+static int execlists_context_queue(struct intel_engine_cs *ring,
+				   struct intel_context *to,
+				   u32 tail)
+{
+	struct intel_ctx_submit_request *req = NULL;
+	unsigned long flags;
+	bool was_empty;
+
+	req = kzalloc(sizeof(*req), GFP_KERNEL);
+	if (req == NULL)
+		return -ENOMEM;
+	req->ctx = to;
+	i915_gem_context_reference(req->ctx);
+	req->ring = ring;
+	req->tail = tail;
+
+	spin_lock_irqsave(&ring->execlist_lock, flags);
+
+	was_empty = list_empty(&ring->execlist_queue);
+	list_add_tail(&req->execlist_link, &ring->execlist_queue);
+	if (was_empty)
+		execlists_context_unqueue(ring);
+
+	spin_unlock_irqrestore(&ring->execlist_lock, flags);
+
+	return 0;
+}
+
 static int logical_ring_invalidate_all_caches(struct intel_ringbuffer *ringbuf)
 {
 	struct intel_engine_cs *ring = ringbuf->ring;
@@ -405,8 +465,7 @@ void intel_logical_ring_advance_and_submit(struct intel_ringbuffer *ringbuf)
 	if (intel_ring_stopped(ring))
 		return;
 
-	/* FIXME: too cheeky, we don't even check if the ELSP is ready */
-	execlists_submit_context(ring, ctx, ringbuf->tail, NULL, 0);
+	execlists_context_queue(ring, ctx, ringbuf->tail);
 }
 
 static int logical_ring_alloc_seqno(struct intel_engine_cs *ring,
@@ -846,6 +905,9 @@ static int logical_ring_init(struct drm_device *dev, struct intel_engine_cs *rin
 	INIT_LIST_HEAD(&ring->request_list);
 	init_waitqueue_head(&ring->irq_queue);
 
+	INIT_LIST_HEAD(&ring->execlist_queue);
+	spin_lock_init(&ring->execlist_lock);
+
 	ret = intel_lr_context_deferred_create(dctx, ring);
 	if (ret)
 		return ret;

drivers/gpu/drm/i915/intel_lrc.h

@@ -59,4 +59,12 @@ int intel_execlists_submission(struct drm_device *dev, struct drm_file *file,
 			       u64 exec_start, u32 flags);
 u32 intel_execlists_ctx_id(struct drm_i915_gem_object *ctx_obj);
 
+struct intel_ctx_submit_request {
+	struct intel_context *ctx;
+	struct intel_engine_cs *ring;
+	u32 tail;
+
+	struct list_head execlist_link;
+};
+
 #endif /* _INTEL_LRC_H_ */

drivers/gpu/drm/i915/intel_ringbuffer.h

@@ -231,6 +231,8 @@ struct  intel_engine_cs {
 	} semaphore;
 
 	/* Execlists */
+	spinlock_t execlist_lock;
+	struct list_head execlist_queue;
 	u32             irq_keep_mask; /* bitmask for interrupts that should not be masked */
 	int		(*emit_request)(struct intel_ringbuffer *ringbuf);
 	int		(*emit_flush)(struct intel_ringbuffer *ringbuf,