[PATCH 3/6] active-ref

[Chris Wilson]

---
 drivers/gpu/drm/i915/i915_active.c | 103 +++++++++++++++++++++++++++--
 drivers/gpu/drm/i915/i915_active.h |   5 ++
 2 files changed, 101 insertions(+), 7 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_active.c b/drivers/gpu/drm/i915/i915_active.c
index c4048628188a..fcd875c78b1b 100644
--- a/drivers/gpu/drm/i915/i915_active.c
+++ b/drivers/gpu/drm/i915/i915_active.c
@@ -217,11 +217,10 @@ excl_retire(struct dma_fence *fence, struct dma_fence_cb *cb)
 }
 
 static struct i915_active_fence *
-active_instance(struct i915_active *ref, struct intel_timeline *tl)
+active_instance(struct i915_active *ref, u64 idx)
 {
 	struct active_node *node, *prealloc;
 	struct rb_node **p, *parent;
-	u64 idx = tl->fence_context;
 
 	/*
 	 * We track the most recently used timeline to skip a rbtree search
@@ -367,7 +366,7 @@ int i915_active_ref(struct i915_active *ref,
 	if (err)
 		return err;
 
-	active = active_instance(ref, tl);
+	active = active_instance(ref, tl->fence_context);
 	if (!active) {
 		err = -ENOMEM;
 		goto out;
@@ -391,25 +390,97 @@ int i915_active_ref(struct i915_active *ref,
 	return err;
 }
 
-struct dma_fence *
-i915_active_set_exclusive(struct i915_active *ref, struct dma_fence *f)
+static struct dma_fence *
+__i915_active_set_fence(struct i915_active *ref,
+			struct i915_active_fence *active,
+			struct dma_fence *fence)
 {
 	struct dma_fence *prev;
 
 	/* We expect the caller to manage the exclusive timeline ordering */
 	GEM_BUG_ON(i915_active_is_idle(ref));
 
+	if (is_barrier(active)) { /* proto-node used by our idle barrier */
+		/*
+		 * This request is on the kernel_context timeline, and so
+		 * we can use it to substitute for the pending idle-barrer
+		 * request that we want to emit on the kernel_context.
+		 */
+		__active_del_barrier(ref, node_from_active(active));
+		RCU_INIT_POINTER(active->fence, NULL);
+		atomic_dec(&ref->count);
+	}
+
 	rcu_read_lock();
-	prev = __i915_active_fence_set(&ref->excl, f);
+	prev = __i915_active_fence_set(active, fence);
 	if (prev)
 		prev = dma_fence_get_rcu(prev);
 	else
-		atomic_inc(&ref->count);
+		__i915_active_acquire(ref);
 	rcu_read_unlock();
 
 	return prev;
 }
 
+static struct i915_active_fence *
+__active_lookup(struct i915_active *ref, u64 idx)
+{
+	struct active_node *node;
+	struct rb_node *p;
+
+	/* Like active_instance() but with no malloc */
+
+	node = READ_ONCE(ref->cache);
+	if (node && node->timeline == idx)
+		return &node->base;
+
+	spin_lock_irq(&ref->tree_lock);
+	GEM_BUG_ON(i915_active_is_idle(ref));
+
+	p = ref->tree.rb_node;
+	while (p) {
+		node = rb_entry(p, struct active_node, node);
+		if (node->timeline == idx) {
+			ref->cache = node;
+			spin_unlock_irq(&ref->tree_lock);
+			return &node->base;
+		}
+
+		if (node->timeline < idx)
+			p = p->rb_right;
+		else
+			p = p->rb_left;
+	}
+
+	spin_unlock_irq(&ref->tree_lock);
+
+	return NULL;
+}
+
+struct dma_fence *
+__i915_active_ref(struct i915_active *ref, u64 idx, struct dma_fence *fence)
+{
+	struct dma_fence *prev = ERR_PTR(-ENOENT);
+	struct i915_active_fence *active;
+
+	if (!i915_active_acquire_if_busy(ref))
+		return ERR_PTR(-EINVAL);
+
+	active = __active_lookup(ref, idx);
+	if (active)
+		prev = __i915_active_set_fence(ref, active, fence);
+
+	i915_active_release(ref);
+	return prev;
+}
+
+struct dma_fence *
+i915_active_set_exclusive(struct i915_active *ref, struct dma_fence *f)
+{
+	/* We expect the caller to manage the exclusive timeline ordering */
+	return __i915_active_set_fence(ref, &ref->excl, f);
+}
+
 bool i915_active_acquire_if_busy(struct i915_active *ref)
 {
 	debug_active_assert(ref);
@@ -443,6 +514,24 @@ int i915_active_acquire(struct i915_active *ref)
 	return err;
 }
 
+int i915_active_acquire_for_context(struct i915_active *ref, u64 idx)
+{
+	struct i915_active_fence *active;
+	int err;
+
+	err = i915_active_acquire(ref);
+	if (err)
+		return err;
+
+	active = active_instance(ref, idx);
+	if (!active) {
+		i915_active_release(ref);
+		return -ENOMEM;
+	}
+
+	return 0; /* return with active ref */
+}
+
 void i915_active_release(struct i915_active *ref)
 {
 	debug_active_assert(ref);
diff --git a/drivers/gpu/drm/i915/i915_active.h b/drivers/gpu/drm/i915/i915_active.h
index b3282ae7913c..65a9c560483b 100644
--- a/drivers/gpu/drm/i915/i915_active.h
+++ b/drivers/gpu/drm/i915/i915_active.h
@@ -163,6 +163,9 @@ void __i915_active_init(struct i915_active *ref,
 	__i915_active_init(ref, active, retire, &__mkey, &__wkey);	\
 } while (0)
 
+struct dma_fence *
+__i915_active_ref(struct i915_active *ref, u64 idx, struct dma_fence *fence);
+
 int i915_active_ref(struct i915_active *ref,
 		    struct intel_timeline *tl,
 		    struct dma_fence *fence);
@@ -192,7 +195,9 @@ int i915_request_await_active(struct i915_request *rq,
 #define I915_ACTIVE_AWAIT_ALL BIT(0)
 
 int i915_active_acquire(struct i915_active *ref);
+int i915_active_acquire_for_context(struct i915_active *ref, u64 idx);
 bool i915_active_acquire_if_busy(struct i915_active *ref);
+
 void i915_active_release(struct i915_active *ref);
 
 static inline void __i915_active_acquire(struct i915_active *ref)
-- 
2.20.1