1 files changed, 93 insertions, 107 deletions

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 28f91df2604d..3e0cabe9b544 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -1134,7 +1134,7 @@ static bool can_wait_boost(struct drm_i915_file_private *file_priv)
 }
 
 /**
- * __wait_seqno - wait until execution of seqno has finished
+ * __i915_wait_seqno - wait until execution of seqno has finished
  * @ring: the ring expected to report seqno
  * @seqno: duh!
  * @reset_counter: reset sequence associated with the given seqno
@@ -1151,7 +1151,7 @@ static bool can_wait_boost(struct drm_i915_file_private *file_priv)
  * Returns 0 if the seqno was found within the alloted time. Else returns the
  * errno with remaining time filled in timeout argument.
  */
-static int __wait_seqno(struct intel_engine_cs *ring, u32 seqno,
+int __i915_wait_seqno(struct intel_engine_cs *ring, u32 seqno,
 			unsigned reset_counter,
 			bool interruptible,
 			s64 *timeout,
@@ -1262,6 +1262,7 @@ i915_wait_seqno(struct intel_engine_cs *ring, uint32_t seqno)
 	struct drm_device *dev = ring->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	bool interruptible = dev_priv->mm.interruptible;
+	unsigned reset_counter;
 	int ret;
 
 	BUG_ON(!mutex_is_locked(&dev->struct_mutex));
@@ -1275,14 +1276,13 @@ i915_wait_seqno(struct intel_engine_cs *ring, uint32_t seqno)
 	if (ret)
 		return ret;
 
-	return __wait_seqno(ring, seqno,
-			    atomic_read(&dev_priv->gpu_error.reset_counter),
-			    interruptible, NULL, NULL);
+	reset_counter = atomic_read(&dev_priv->gpu_error.reset_counter);
+	return __i915_wait_seqno(ring, seqno, reset_counter, interruptible,
+				 NULL, NULL);
 }
 
 static int
-i915_gem_object_wait_rendering__tail(struct drm_i915_gem_object *obj,
-				     struct intel_engine_cs *ring)
+i915_gem_object_wait_rendering__tail(struct drm_i915_gem_object *obj)
 {
 	if (!obj->active)
 		return 0;
@@ -1319,7 +1319,7 @@ i915_gem_object_wait_rendering(struct drm_i915_gem_object *obj,
 	if (ret)
 		return ret;
 
-	return i915_gem_object_wait_rendering__tail(obj, ring);
+	return i915_gem_object_wait_rendering__tail(obj);
 }
 
 /* A nonblocking variant of the above wait. This is a highly dangerous routine
@@ -1354,12 +1354,13 @@ i915_gem_object_wait_rendering__nonblocking(struct drm_i915_gem_object *obj,
 
 	reset_counter = atomic_read(&dev_priv->gpu_error.reset_counter);
 	mutex_unlock(&dev->struct_mutex);
-	ret = __wait_seqno(ring, seqno, reset_counter, true, NULL, file_priv);
+	ret = __i915_wait_seqno(ring, seqno, reset_counter, true, NULL,
+				file_priv);
 	mutex_lock(&dev->struct_mutex);
 	if (ret)
 		return ret;
 
-	return i915_gem_object_wait_rendering__tail(obj, ring);
+	return i915_gem_object_wait_rendering__tail(obj);
 }
 
 /**
@@ -1466,6 +1467,16 @@ unlock:
  *
  * While the mapping holds a reference on the contents of the object, it doesn't
  * imply a ref on the object itself.
+ *
+ * IMPORTANT:
+ *
+ * DRM driver writers who look a this function as an example for how to do GEM
+ * mmap support, please don't implement mmap support like here. The modern way
+ * to implement DRM mmap support is with an mmap offset ioctl (like
+ * i915_gem_mmap_gtt) and then using the mmap syscall on the DRM fd directly.
+ * That way debug tooling like valgrind will understand what's going on, hiding
+ * the mmap call in a driver private ioctl will break that. The i915 driver only
+ * does cpu mmaps this way because we didn't know better.
  */
 int
 i915_gem_mmap_ioctl(struct drm_device *dev, void *data,
@@ -1945,7 +1956,14 @@ unsigned long
 i915_gem_shrink(struct drm_i915_private *dev_priv,
 		long target, unsigned flags)
 {
-	const bool purgeable_only = flags & I915_SHRINK_PURGEABLE;
+	const struct {
+		struct list_head *list;
+		unsigned int bit;
+	} phases[] = {
+		{ &dev_priv->mm.unbound_list, I915_SHRINK_UNBOUND },
+		{ &dev_priv->mm.bound_list, I915_SHRINK_BOUND },
+		{ NULL, 0 },
+	}, *phase;
 	unsigned long count = 0;
 
 	/*
@@ -1967,48 +1985,30 @@ i915_gem_shrink(struct drm_i915_private *dev_priv,
 	 * dev->struct_mutex and so we won't ever be able to observe an
 	 * object on the bound_list with a reference count equals 0.
 	 */
-	if (flags & I915_SHRINK_UNBOUND) {
+	for (phase = phases; phase->list; phase++) {
 		struct list_head still_in_list;
 
-		INIT_LIST_HEAD(&still_in_list);
-		while (count < target && !list_empty(&dev_priv->mm.unbound_list)) {
-			struct drm_i915_gem_object *obj;
-
-			obj = list_first_entry(&dev_priv->mm.unbound_list,
-					       typeof(*obj), global_list);
-			list_move_tail(&obj->global_list, &still_in_list);
-
-			if (!i915_gem_object_is_purgeable(obj) && purgeable_only)
-				continue;
-
-			drm_gem_object_reference(&obj->base);
-
-			if (i915_gem_object_put_pages(obj) == 0)
-				count += obj->base.size >> PAGE_SHIFT;
-
-			drm_gem_object_unreference(&obj->base);
-		}
-		list_splice(&still_in_list, &dev_priv->mm.unbound_list);
-	}
-
-	if (flags & I915_SHRINK_BOUND) {
-		struct list_head still_in_list;
+		if ((flags & phase->bit) == 0)
+			continue;
 
 		INIT_LIST_HEAD(&still_in_list);
-		while (count < target && !list_empty(&dev_priv->mm.bound_list)) {
+		while (count < target && !list_empty(phase->list)) {
 			struct drm_i915_gem_object *obj;
 			struct i915_vma *vma, *v;
 
-			obj = list_first_entry(&dev_priv->mm.bound_list,
+			obj = list_first_entry(phase->list,
 					       typeof(*obj), global_list);
 			list_move_tail(&obj->global_list, &still_in_list);
 
-			if (!i915_gem_object_is_purgeable(obj) && purgeable_only)
+			if (flags & I915_SHRINK_PURGEABLE &&
+			    !i915_gem_object_is_purgeable(obj))
 				continue;
 
 			drm_gem_object_reference(&obj->base);
 
-			list_for_each_entry_safe(vma, v, &obj->vma_list, vma_link)
+			/* For the unbound phase, this should be a no-op! */
+			list_for_each_entry_safe(vma, v,
+						 &obj->vma_list, vma_link)
 				if (i915_vma_unbind(vma))
 					break;
 
@@ -2017,7 +2017,7 @@ i915_gem_shrink(struct drm_i915_private *dev_priv,
 
 			drm_gem_object_unreference(&obj->base);
 		}
-		list_splice(&still_in_list, &dev_priv->mm.bound_list);
+		list_splice(&still_in_list, phase->list);
 	}
 
 	return count;
@@ -2811,6 +2811,9 @@ i915_gem_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
 	u32 seqno = 0;
 	int ret = 0;
 
+	if (args->flags != 0)
+		return -EINVAL;
+
 	ret = i915_mutex_lock_interruptible(dev);
 	if (ret)
 		return ret;
@@ -2846,8 +2849,8 @@ i915_gem_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
 	reset_counter = atomic_read(&dev_priv->gpu_error.reset_counter);
 	mutex_unlock(&dev->struct_mutex);
 
-	return __wait_seqno(ring, seqno, reset_counter, true, &args->timeout_ns,
-			    file->driver_priv);
+	return __i915_wait_seqno(ring, seqno, reset_counter, true,
+				 &args->timeout_ns, file->driver_priv);
 
 out:
 	drm_gem_object_unreference(&obj->base);
@@ -3166,6 +3169,7 @@ static void i915_gem_write_fence(struct drm_device *dev, int reg,
 	     obj->stride, obj->tiling_mode);
 
 	switch (INTEL_INFO(dev)->gen) {
+	case 9:
 	case 8:
 	case 7:
 	case 6:
@@ -3384,46 +3388,6 @@ static bool i915_gem_valid_gtt_space(struct i915_vma *vma,
 	return true;
 }
 
-static void i915_gem_verify_gtt(struct drm_device *dev)
-{
-#if WATCH_GTT
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_gem_object *obj;
-	int err = 0;
-
-	list_for_each_entry(obj, &dev_priv->mm.gtt_list, global_list) {
-		if (obj->gtt_space == NULL) {
-			printk(KERN_ERR "object found on GTT list with no space reserved\n");
-			err++;
-			continue;
-		}
-
-		if (obj->cache_level != obj->gtt_space->color) {
-			printk(KERN_ERR "object reserved space [%08lx, %08lx] with wrong color, cache_level=%x, color=%lx\n",
-			       i915_gem_obj_ggtt_offset(obj),
-			       i915_gem_obj_ggtt_offset(obj) + i915_gem_obj_ggtt_size(obj),
-			       obj->cache_level,
-			       obj->gtt_space->color);
-			err++;
-			continue;
-		}
-
-		if (!i915_gem_valid_gtt_space(dev,
-					      obj->gtt_space,
-					      obj->cache_level)) {
-			printk(KERN_ERR "invalid GTT space found at [%08lx, %08lx] - color=%x\n",
-			       i915_gem_obj_ggtt_offset(obj),
-			       i915_gem_obj_ggtt_offset(obj) + i915_gem_obj_ggtt_size(obj),
-			       obj->cache_level);
-			err++;
-			continue;
-		}
-	}
-
-	WARN_ON(err);
-#endif
-}
-
 /**
  * Finds free space in the GTT aperture and binds the object there.
  */
@@ -3514,25 +3478,10 @@ search_free:
 	list_move_tail(&obj->global_list, &dev_priv->mm.bound_list);
 	list_add_tail(&vma->mm_list, &vm->inactive_list);
 
-	if (i915_is_ggtt(vm)) {
-		bool mappable, fenceable;
-
-		fenceable = (vma->node.size == fence_size &&
-			     (vma->node.start & (fence_alignment - 1)) == 0);
-
-		mappable = (vma->node.start + obj->base.size <=
-			    dev_priv->gtt.mappable_end);
-
-		obj->map_and_fenceable = mappable && fenceable;
-	}
-
-	WARN_ON(flags & PIN_MAPPABLE && !obj->map_and_fenceable);
-
 	trace_i915_vma_bind(vma, flags);
 	vma->bind_vma(vma, obj->cache_level,
-		      flags & (PIN_MAPPABLE | PIN_GLOBAL) ? GLOBAL_BIND : 0);
+		      flags & PIN_GLOBAL ? GLOBAL_BIND : 0);
 
-	i915_gem_verify_gtt(dev);
 	return vma;
 
 err_remove_node:
@@ -3739,7 +3688,7 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
 		list_for_each_entry(vma, &obj->vma_list, vma_link)
 			if (drm_mm_node_allocated(&vma->node))
 				vma->bind_vma(vma, cache_level,
-					      obj->has_global_gtt_mapping ? GLOBAL_BIND : 0);
+						vma->bound & GLOBAL_BIND);
 	}
 
 	list_for_each_entry(vma, &obj->vma_list, vma_link)
@@ -3769,7 +3718,6 @@ int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
 						    old_write_domain);
 	}
 
-	i915_gem_verify_gtt(dev);
 	return 0;
 }
 
@@ -4067,7 +4015,7 @@ i915_gem_ring_throttle(struct drm_device *dev, struct drm_file *file)
 	if (seqno == 0)
 		return 0;
 
-	ret = __wait_seqno(ring, seqno, reset_counter, true, NULL, NULL);
+	ret = __i915_wait_seqno(ring, seqno, reset_counter, true, NULL, NULL);
 	if (ret == 0)
 		queue_delayed_work(dev_priv->wq, &dev_priv->mm.retire_work, 0);
 
@@ -4101,6 +4049,7 @@ i915_gem_object_pin(struct drm_i915_gem_object *obj,
 {
 	struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
 	struct i915_vma *vma;
+	unsigned bound;
 	int ret;
 
 	if (WARN_ON(vm == &dev_priv->mm.aliasing_ppgtt->base))
@@ -4109,6 +4058,9 @@ i915_gem_object_pin(struct drm_i915_gem_object *obj,
 	if (WARN_ON(flags & (PIN_GLOBAL | PIN_MAPPABLE) && !i915_is_ggtt(vm)))
 		return -EINVAL;
 
+	if (WARN_ON((flags & (PIN_MAPPABLE | PIN_GLOBAL)) == PIN_MAPPABLE))
+		return -EINVAL;
+
 	vma = i915_gem_obj_to_vma(obj, vm);
 	if (vma) {
 		if (WARN_ON(vma->pin_count == DRM_I915_GEM_OBJECT_MAX_PIN_COUNT))
@@ -4130,15 +4082,39 @@ i915_gem_object_pin(struct drm_i915_gem_object *obj,
 		}
 	}
 
+	bound = vma ? vma->bound : 0;
 	if (vma == NULL || !drm_mm_node_allocated(&vma->node)) {
 		vma = i915_gem_object_bind_to_vm(obj, vm, alignment, flags);
 		if (IS_ERR(vma))
 			return PTR_ERR(vma);
 	}
 
-	if (flags & PIN_GLOBAL && !obj->has_global_gtt_mapping)
+	if (flags & PIN_GLOBAL && !(vma->bound & GLOBAL_BIND))
 		vma->bind_vma(vma, obj->cache_level, GLOBAL_BIND);
 
+	if ((bound ^ vma->bound) & GLOBAL_BIND) {
+		bool mappable, fenceable;
+		u32 fence_size, fence_alignment;
+
+		fence_size = i915_gem_get_gtt_size(obj->base.dev,
+						   obj->base.size,
+						   obj->tiling_mode);
+		fence_alignment = i915_gem_get_gtt_alignment(obj->base.dev,
+							     obj->base.size,
+							     obj->tiling_mode,
+							     true);
+
+		fenceable = (vma->node.size == fence_size &&
+			     (vma->node.start & (fence_alignment - 1)) == 0);
+
+		mappable = (vma->node.start + obj->base.size <=
+			    dev_priv->gtt.mappable_end);
+
+		obj->map_and_fenceable = mappable && fenceable;
+	}
+
+	WARN_ON(flags & PIN_MAPPABLE && !obj->map_and_fenceable);
+
 	vma->pin_count++;
 	if (flags & PIN_MAPPABLE)
 		obj->pin_mappable |= true;
@@ -5119,6 +5095,15 @@ int i915_gem_open(struct drm_device *dev, struct drm_file *file)
 	return ret;
 }
 
+/**
+ * i915_gem_track_fb - update frontbuffer tracking
+ * old: current GEM buffer for the frontbuffer slots
+ * new: new GEM buffer for the frontbuffer slots
+ * frontbuffer_bits: bitmask of frontbuffer slots
+ *
+ * This updates the frontbuffer tracking bits @frontbuffer_bits by clearing them
+ * from @old and setting them in @new. Both @old and @new can be NULL.
+ */
 void i915_gem_track_fb(struct drm_i915_gem_object *old,
 		       struct drm_i915_gem_object *new,
 		       unsigned frontbuffer_bits)
@@ -5302,7 +5287,7 @@ i915_gem_shrinker_oom(struct notifier_block *nb, unsigned long event, void *ptr)
 	struct drm_device *dev = dev_priv->dev;
 	struct drm_i915_gem_object *obj;
 	unsigned long timeout = msecs_to_jiffies(5000) + 1;
-	unsigned long pinned, bound, unbound, freed;
+	unsigned long pinned, bound, unbound, freed_pages;
 	bool was_interruptible;
 	bool unlock;
 
@@ -5319,7 +5304,7 @@ i915_gem_shrinker_oom(struct notifier_block *nb, unsigned long event, void *ptr)
 	was_interruptible = dev_priv->mm.interruptible;
 	dev_priv->mm.interruptible = false;
 
-	freed = i915_gem_shrink_all(dev_priv);
+	freed_pages = i915_gem_shrink_all(dev_priv);
 
 	dev_priv->mm.interruptible = was_interruptible;
 
@@ -5350,14 +5335,15 @@ i915_gem_shrinker_oom(struct notifier_block *nb, unsigned long event, void *ptr)
 	if (unlock)
 		mutex_unlock(&dev->struct_mutex);
 
-	pr_info("Purging GPU memory, %lu bytes freed, %lu bytes still pinned.\n",
-		freed, pinned);
+	if (freed_pages || unbound || bound)
+		pr_info("Purging GPU memory, %lu bytes freed, %lu bytes still pinned.\n",
+			freed_pages << PAGE_SHIFT, pinned);
 	if (unbound || bound)
 		pr_err("%lu and %lu bytes still available in the "
 		       "bound and unbound GPU page lists.\n",
 		       bound, unbound);
 
-	*(unsigned long *)ptr += freed;
+	*(unsigned long *)ptr += freed_pages;
 	return NOTIFY_DONE;
 }