Merge branch 'drm-core-next' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6

* 'drm-core-next' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6: (390 commits)
  drm/radeon/kms: disable underscan by default
  drm/radeon/kms: only enable hdmi features if the monitor supports audio
  drm: Restore the old_fb upon modeset failure
  drm/nouveau: fix hwmon device binding
  radeon: consolidate asic-specific function decls for pre-r600
  vga_switcheroo: comparing too few characters in strncmp()
  drm/radeon/kms: add NI pci ids
  drm/radeon/kms: don't enable pcie gen2 on NI yet
  drm/radeon/kms: add radeon_asic struct for NI asics
  drm/radeon/kms/ni: load default sclk/mclk/vddc at pm init
  drm/radeon/kms: add ucode loader for NI
  drm/radeon/kms: add support for DCE5 display LUTs
  drm/radeon/kms: add ni_reg.h
  drm/radeon/kms: add bo blit support for NI
  drm/radeon/kms: always use writeback/events for fences on NI
  drm/radeon/kms: adjust default clock/vddc tracking for pm on DCE5
  drm/radeon/kms: add backend map workaround for barts
  drm/radeon/kms: fill gpu init for NI asics
  drm/radeon/kms: add disabled vbios accessor for NI asics
  drm/radeon/kms: handle NI thermal controller
  ...

4 files changed, 381 insertions, 111 deletions

diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 934a96a78540..af61fc29e843 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -169,7 +169,7 @@ int ttm_bo_wait_unreserved(struct ttm_buffer_object *bo, bool interruptible)
 }
 EXPORT_SYMBOL(ttm_bo_wait_unreserved);
 
-static void ttm_bo_add_to_lru(struct ttm_buffer_object *bo)
+void ttm_bo_add_to_lru(struct ttm_buffer_object *bo)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
 	struct ttm_mem_type_manager *man;
@@ -191,11 +191,7 @@ static void ttm_bo_add_to_lru(struct ttm_buffer_object *bo)
 	}
 }
 
-/**
- * Call with the lru_lock held.
- */
-
-static int ttm_bo_del_from_lru(struct ttm_buffer_object *bo)
+int ttm_bo_del_from_lru(struct ttm_buffer_object *bo)
 {
 	int put_count = 0;
 
@@ -227,9 +223,18 @@ int ttm_bo_reserve_locked(struct ttm_buffer_object *bo,
 		/**
 		 * Deadlock avoidance for multi-bo reserving.
 		 */
-		if (use_sequence && bo->seq_valid &&
-			(sequence - bo->val_seq < (1 << 31))) {
-			return -EAGAIN;
+		if (use_sequence && bo->seq_valid) {
+			/**
+			 * We've already reserved this one.
+			 */
+			if (unlikely(sequence == bo->val_seq))
+				return -EDEADLK;
+			/**
+			 * Already reserved by a thread that will not back
+			 * off for us. We need to back off.
+			 */
+			if (unlikely(sequence - bo->val_seq < (1 << 31)))
+				return -EAGAIN;
 		}
 
 		if (no_wait)
@@ -267,6 +272,13 @@ static void ttm_bo_ref_bug(struct kref *list_kref)
 	BUG();
 }
 
+void ttm_bo_list_ref_sub(struct ttm_buffer_object *bo, int count,
+			 bool never_free)
+{
+	kref_sub(&bo->list_kref, count,
+		 (never_free) ? ttm_bo_ref_bug : ttm_bo_release_list);
+}
+
 int ttm_bo_reserve(struct ttm_buffer_object *bo,
 		   bool interruptible,
 		   bool no_wait, bool use_sequence, uint32_t sequence)
@@ -282,20 +294,24 @@ int ttm_bo_reserve(struct ttm_buffer_object *bo,
 		put_count = ttm_bo_del_from_lru(bo);
 	spin_unlock(&glob->lru_lock);
 
-	while (put_count--)
-		kref_put(&bo->list_kref, ttm_bo_ref_bug);
+	ttm_bo_list_ref_sub(bo, put_count, true);
 
 	return ret;
 }
 
+void ttm_bo_unreserve_locked(struct ttm_buffer_object *bo)
+{
+	ttm_bo_add_to_lru(bo);
+	atomic_set(&bo->reserved, 0);
+	wake_up_all(&bo->event_queue);
+}
+
 void ttm_bo_unreserve(struct ttm_buffer_object *bo)
 {
 	struct ttm_bo_global *glob = bo->glob;
 
 	spin_lock(&glob->lru_lock);
-	ttm_bo_add_to_lru(bo);
-	atomic_set(&bo->reserved, 0);
-	wake_up_all(&bo->event_queue);
+	ttm_bo_unreserve_locked(bo);
 	spin_unlock(&glob->lru_lock);
 }
 EXPORT_SYMBOL(ttm_bo_unreserve);
@@ -362,8 +378,13 @@ static int ttm_bo_handle_move_mem(struct ttm_buffer_object *bo,
 	int ret = 0;
 
 	if (old_is_pci || new_is_pci ||
-	    ((mem->placement & bo->mem.placement & TTM_PL_MASK_CACHING) == 0))
-		ttm_bo_unmap_virtual(bo);
+	    ((mem->placement & bo->mem.placement & TTM_PL_MASK_CACHING) == 0)) {
+		ret = ttm_mem_io_lock(old_man, true);
+		if (unlikely(ret != 0))
+			goto out_err;
+		ttm_bo_unmap_virtual_locked(bo);
+		ttm_mem_io_unlock(old_man);
+	}
 
 	/*
 	 * Create and bind a ttm if required.
@@ -416,11 +437,9 @@ moved:
 	}
 
 	if (bo->mem.mm_node) {
-		spin_lock(&bo->lock);
 		bo->offset = (bo->mem.start << PAGE_SHIFT) +
 		    bdev->man[bo->mem.mem_type].gpu_offset;
 		bo->cur_placement = bo->mem.placement;
-		spin_unlock(&bo->lock);
 	} else
 		bo->offset = 0;
 
@@ -452,7 +471,6 @@ static void ttm_bo_cleanup_memtype_use(struct ttm_buffer_object *bo)
 		ttm_tt_destroy(bo->ttm);
 		bo->ttm = NULL;
 	}
-
 	ttm_bo_mem_put(bo, &bo->mem);
 
 	atomic_set(&bo->reserved, 0);
@@ -474,14 +492,14 @@ static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo)
 	int put_count;
 	int ret;
 
-	spin_lock(&bo->lock);
+	spin_lock(&bdev->fence_lock);
 	(void) ttm_bo_wait(bo, false, false, true);
 	if (!bo->sync_obj) {
 
 		spin_lock(&glob->lru_lock);
 
 		/**
-		 * Lock inversion between bo::reserve and bo::lock here,
+		 * Lock inversion between bo:reserve and bdev::fence_lock here,
 		 * but that's OK, since we're only trylocking.
 		 */
 
@@ -490,14 +508,13 @@ static void ttm_bo_cleanup_refs_or_queue(struct ttm_buffer_object *bo)
 		if (unlikely(ret == -EBUSY))
 			goto queue;
 
-		spin_unlock(&bo->lock);
+		spin_unlock(&bdev->fence_lock);
 		put_count = ttm_bo_del_from_lru(bo);
 
 		spin_unlock(&glob->lru_lock);
 		ttm_bo_cleanup_memtype_use(bo);
 
-		while (put_count--)
-			kref_put(&bo->list_kref, ttm_bo_ref_bug);
+		ttm_bo_list_ref_sub(bo, put_count, true);
 
 		return;
 	} else {
@@ -512,7 +529,7 @@ queue:
 	kref_get(&bo->list_kref);
 	list_add_tail(&bo->ddestroy, &bdev->ddestroy);
 	spin_unlock(&glob->lru_lock);
-	spin_unlock(&bo->lock);
+	spin_unlock(&bdev->fence_lock);
 
 	if (sync_obj) {
 		driver->sync_obj_flush(sync_obj, sync_obj_arg);
@@ -537,14 +554,15 @@ static int ttm_bo_cleanup_refs(struct ttm_buffer_object *bo,
 			       bool no_wait_reserve,
 			       bool no_wait_gpu)
 {
+	struct ttm_bo_device *bdev = bo->bdev;
 	struct ttm_bo_global *glob = bo->glob;
 	int put_count;
 	int ret = 0;
 
 retry:
-	spin_lock(&bo->lock);
+	spin_lock(&bdev->fence_lock);
 	ret = ttm_bo_wait(bo, false, interruptible, no_wait_gpu);
-	spin_unlock(&bo->lock);
+	spin_unlock(&bdev->fence_lock);
 
 	if (unlikely(ret != 0))
 		return ret;
@@ -580,8 +598,7 @@ retry:
 	spin_unlock(&glob->lru_lock);
 	ttm_bo_cleanup_memtype_use(bo);
 
-	while (put_count--)
-		kref_put(&bo->list_kref, ttm_bo_ref_bug);
+	ttm_bo_list_ref_sub(bo, put_count, true);
 
 	return 0;
 }
@@ -652,6 +669,7 @@ static void ttm_bo_release(struct kref *kref)
 	struct ttm_buffer_object *bo =
 	    container_of(kref, struct ttm_buffer_object, kref);
 	struct ttm_bo_device *bdev = bo->bdev;
+	struct ttm_mem_type_manager *man = &bdev->man[bo->mem.mem_type];
 
 	if (likely(bo->vm_node != NULL)) {
 		rb_erase(&bo->vm_rb, &bdev->addr_space_rb);
@@ -659,6 +677,9 @@ static void ttm_bo_release(struct kref *kref)
 		bo->vm_node = NULL;
 	}
 	write_unlock(&bdev->vm_lock);
+	ttm_mem_io_lock(man, false);
+	ttm_mem_io_free_vm(bo);
+	ttm_mem_io_unlock(man);
 	ttm_bo_cleanup_refs_or_queue(bo);
 	kref_put(&bo->list_kref, ttm_bo_release_list);
 	write_lock(&bdev->vm_lock);
@@ -698,9 +719,9 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
 	struct ttm_placement placement;
 	int ret = 0;
 
-	spin_lock(&bo->lock);
+	spin_lock(&bdev->fence_lock);
 	ret = ttm_bo_wait(bo, false, interruptible, no_wait_gpu);
-	spin_unlock(&bo->lock);
+	spin_unlock(&bdev->fence_lock);
 
 	if (unlikely(ret != 0)) {
 		if (ret != -ERESTARTSYS) {
@@ -715,7 +736,8 @@ static int ttm_bo_evict(struct ttm_buffer_object *bo, bool interruptible,
 
 	evict_mem = bo->mem;
 	evict_mem.mm_node = NULL;
-	evict_mem.bus.io_reserved = false;
+	evict_mem.bus.io_reserved_vm = false;
+	evict_mem.bus.io_reserved_count = 0;
 
 	placement.fpfn = 0;
 	placement.lpfn = 0;
@@ -802,8 +824,7 @@ retry:
 
 	BUG_ON(ret != 0);
 
-	while (put_count--)
-		kref_put(&bo->list_kref, ttm_bo_ref_bug);
+	ttm_bo_list_ref_sub(bo, put_count, true);
 
 	ret = ttm_bo_evict(bo, interruptible, no_wait_reserve, no_wait_gpu);
 	ttm_bo_unreserve(bo);
@@ -1036,6 +1057,7 @@ int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
 {
 	int ret = 0;
 	struct ttm_mem_reg mem;
+	struct ttm_bo_device *bdev = bo->bdev;
 
 	BUG_ON(!atomic_read(&bo->reserved));
 
@@ -1044,15 +1066,16 @@ int ttm_bo_move_buffer(struct ttm_buffer_object *bo,
 	 * Have the driver move function wait for idle when necessary,
 	 * instead of doing it here.
 	 */
-	spin_lock(&bo->lock);
+	spin_lock(&bdev->fence_lock);
 	ret = ttm_bo_wait(bo, false, interruptible, no_wait_gpu);
-	spin_unlock(&bo->lock);
+	spin_unlock(&bdev->fence_lock);
 	if (ret)
 		return ret;
 	mem.num_pages = bo->num_pages;
 	mem.size = mem.num_pages << PAGE_SHIFT;
 	mem.page_alignment = bo->mem.page_alignment;
-	mem.bus.io_reserved = false;
+	mem.bus.io_reserved_vm = false;
+	mem.bus.io_reserved_count = 0;
 	/*
 	 * Determine where to move the buffer.
 	 */
@@ -1163,7 +1186,6 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 	}
 	bo->destroy = destroy;
 
-	spin_lock_init(&bo->lock);
 	kref_init(&bo->kref);
 	kref_init(&bo->list_kref);
 	atomic_set(&bo->cpu_writers, 0);
@@ -1172,6 +1194,7 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 	INIT_LIST_HEAD(&bo->lru);
 	INIT_LIST_HEAD(&bo->ddestroy);
 	INIT_LIST_HEAD(&bo->swap);
+	INIT_LIST_HEAD(&bo->io_reserve_lru);
 	bo->bdev = bdev;
 	bo->glob = bdev->glob;
 	bo->type = type;
@@ -1181,7 +1204,8 @@ int ttm_bo_init(struct ttm_bo_device *bdev,
 	bo->mem.num_pages = bo->num_pages;
 	bo->mem.mm_node = NULL;
 	bo->mem.page_alignment = page_alignment;
-	bo->mem.bus.io_reserved = false;
+	bo->mem.bus.io_reserved_vm = false;
+	bo->mem.bus.io_reserved_count = 0;
 	bo->buffer_start = buffer_start & PAGE_MASK;
 	bo->priv_flags = 0;
 	bo->mem.placement = (TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED);
@@ -1355,6 +1379,10 @@ int ttm_bo_init_mm(struct ttm_bo_device *bdev, unsigned type,
 	BUG_ON(type >= TTM_NUM_MEM_TYPES);
 	man = &bdev->man[type];
 	BUG_ON(man->has_type);
+	man->io_reserve_fastpath = true;
+	man->use_io_reserve_lru = false;
+	mutex_init(&man->io_reserve_mutex);
+	INIT_LIST_HEAD(&man->io_reserve_lru);
 
 	ret = bdev->driver->init_mem_type(bdev, type, man);
 	if (ret)
@@ -1526,7 +1554,8 @@ int ttm_bo_device_init(struct ttm_bo_device *bdev,
 	bdev->dev_mapping = NULL;
 	bdev->glob = glob;
 	bdev->need_dma32 = need_dma32;
-
+	bdev->val_seq = 0;
+	spin_lock_init(&bdev->fence_lock);
 	mutex_lock(&glob->device_list_mutex);
 	list_add_tail(&bdev->device_list, &glob->device_list);
 	mutex_unlock(&glob->device_list_mutex);
@@ -1560,7 +1589,7 @@ bool ttm_mem_reg_is_pci(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem)
 	return true;
 }
 
-void ttm_bo_unmap_virtual(struct ttm_buffer_object *bo)
+void ttm_bo_unmap_virtual_locked(struct ttm_buffer_object *bo)
 {
 	struct ttm_bo_device *bdev = bo->bdev;
 	loff_t offset = (loff_t) bo->addr_space_offset;
@@ -1569,8 +1598,20 @@ void ttm_bo_unmap_virtual(struct ttm_buffer_object *bo)
 	if (!bdev->dev_mapping)
 		return;
 	unmap_mapping_range(bdev->dev_mapping, offset, holelen, 1);
-	ttm_mem_io_free(bdev, &bo->mem);
+	ttm_mem_io_free_vm(bo);
+}
+
+void ttm_bo_unmap_virtual(struct ttm_buffer_object *bo)
+{
+	struct ttm_bo_device *bdev = bo->bdev;
+	struct ttm_mem_type_manager *man = &bdev->man[bo->mem.mem_type];
+
+	ttm_mem_io_lock(man, false);
+	ttm_bo_unmap_virtual_locked(bo);
+	ttm_mem_io_unlock(man);
 }
+
+
 EXPORT_SYMBOL(ttm_bo_unmap_virtual);
 
 static void ttm_bo_vm_insert_rb(struct ttm_buffer_object *bo)
@@ -1650,6 +1691,7 @@ int ttm_bo_wait(struct ttm_buffer_object *bo,
 		bool lazy, bool interruptible, bool no_wait)
 {
 	struct ttm_bo_driver *driver = bo->bdev->driver;
+	struct ttm_bo_device *bdev = bo->bdev;
 	void *sync_obj;
 	void *sync_obj_arg;
 	int ret = 0;
@@ -1663,9 +1705,9 @@ int ttm_bo_wait(struct ttm_buffer_object *bo,
 			void *tmp_obj = bo->sync_obj;
 			bo->sync_obj = NULL;
 			clear_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags);
-			spin_unlock(&bo->lock);
+			spin_unlock(&bdev->fence_lock);
 			driver->sync_obj_unref(&tmp_obj);
-			spin_lock(&bo->lock);
+			spin_lock(&bdev->fence_lock);
 			continue;
 		}
 
@@ -1674,29 +1716,29 @@ int ttm_bo_wait(struct ttm_buffer_object *bo,
 
 		sync_obj = driver->sync_obj_ref(bo->sync_obj);
 		sync_obj_arg = bo->sync_obj_arg;
-		spin_unlock(&bo->lock);
+		spin_unlock(&bdev->fence_lock);
 		ret = driver->sync_obj_wait(sync_obj, sync_obj_arg,
 					    lazy, interruptible);
 		if (unlikely(ret != 0)) {
 			driver->sync_obj_unref(&sync_obj);
-			spin_lock(&bo->lock);
+			spin_lock(&bdev->fence_lock);
 			return ret;
 		}
-		spin_lock(&bo->lock);
+		spin_lock(&bdev->fence_lock);
 		if (likely(bo->sync_obj == sync_obj &&
 			   bo->sync_obj_arg == sync_obj_arg)) {
 			void *tmp_obj = bo->sync_obj;
 			bo->sync_obj = NULL;
 			clear_bit(TTM_BO_PRIV_FLAG_MOVING,
 				  &bo->priv_flags);
-			spin_unlock(&bo->lock);
+			spin_unlock(&bdev->fence_lock);
 			driver->sync_obj_unref(&sync_obj);
 			driver->sync_obj_unref(&tmp_obj);
-			spin_lock(&bo->lock);
+			spin_lock(&bdev->fence_lock);
 		} else {
-			spin_unlock(&bo->lock);
+			spin_unlock(&bdev->fence_lock);
 			driver->sync_obj_unref(&sync_obj);
-			spin_lock(&bo->lock);
+			spin_lock(&bdev->fence_lock);
 		}
 	}
 	return 0;
@@ -1705,6 +1747,7 @@ EXPORT_SYMBOL(ttm_bo_wait);
 
 int ttm_bo_synccpu_write_grab(struct ttm_buffer_object *bo, bool no_wait)
 {
+	struct ttm_bo_device *bdev = bo->bdev;
 	int ret = 0;
 
 	/*
@@ -1714,9 +1757,9 @@ int ttm_bo_synccpu_write_grab(struct ttm_buffer_object *bo, bool no_wait)
 	ret = ttm_bo_reserve(bo, true, no_wait, false, 0);
 	if (unlikely(ret != 0))
 		return ret;
-	spin_lock(&bo->lock);
+	spin_lock(&bdev->fence_lock);
 	ret = ttm_bo_wait(bo, false, true, no_wait);
-	spin_unlock(&bo->lock);
+	spin_unlock(&bdev->fence_lock);
 	if (likely(ret == 0))
 		atomic_inc(&bo->cpu_writers);
 	ttm_bo_unreserve(bo);
@@ -1782,16 +1825,15 @@ static int ttm_bo_swapout(struct ttm_mem_shrink *shrink)
 	put_count = ttm_bo_del_from_lru(bo);
 	spin_unlock(&glob->lru_lock);
 
-	while (put_count--)
-		kref_put(&bo->list_kref, ttm_bo_ref_bug);
+	ttm_bo_list_ref_sub(bo, put_count, true);
 
 	/**
 	 * Wait for GPU, then move to system cached.
 	 */
 
-	spin_lock(&bo->lock);
+	spin_lock(&bo->bdev->fence_lock);
 	ret = ttm_bo_wait(bo, false, false, false);
-	spin_unlock(&bo->lock);
+	spin_unlock(&bo->bdev->fence_lock);
 
 	if (unlikely(ret != 0))
 		goto out;
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index 3106d5bcce32..77dbf408c0d0 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -75,37 +75,123 @@ int ttm_bo_move_ttm(struct ttm_buffer_object *bo,
 }
 EXPORT_SYMBOL(ttm_bo_move_ttm);
 
-int ttm_mem_io_reserve(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem)
+int ttm_mem_io_lock(struct ttm_mem_type_manager *man, bool interruptible)
 {
-	int ret;
+	if (likely(man->io_reserve_fastpath))
+		return 0;
+
+	if (interruptible)
+		return mutex_lock_interruptible(&man->io_reserve_mutex);
+
+	mutex_lock(&man->io_reserve_mutex);
+	return 0;
+}
 
-	if (!mem->bus.io_reserved) {
-		mem->bus.io_reserved = true;
+void ttm_mem_io_unlock(struct ttm_mem_type_manager *man)
+{
+	if (likely(man->io_reserve_fastpath))
+		return;
+
+	mutex_unlock(&man->io_reserve_mutex);
+}
+
+static int ttm_mem_io_evict(struct ttm_mem_type_manager *man)
+{
+	struct ttm_buffer_object *bo;
+
+	if (!man->use_io_reserve_lru || list_empty(&man->io_reserve_lru))
+		return -EAGAIN;
+
+	bo = list_first_entry(&man->io_reserve_lru,
+			      struct ttm_buffer_object,
+			      io_reserve_lru);
+	list_del_init(&bo->io_reserve_lru);
+	ttm_bo_unmap_virtual_locked(bo);
+
+	return 0;
+}
+
+static int ttm_mem_io_reserve(struct ttm_bo_device *bdev,
+			      struct ttm_mem_reg *mem)
+{
+	struct ttm_mem_type_manager *man = &bdev->man[mem->mem_type];
+	int ret = 0;
+
+	if (!bdev->driver->io_mem_reserve)
+		return 0;
+	if (likely(man->io_reserve_fastpath))
+		return bdev->driver->io_mem_reserve(bdev, mem);
+
+	if (bdev->driver->io_mem_reserve &&
+	    mem->bus.io_reserved_count++ == 0) {
+retry:
 		ret = bdev->driver->io_mem_reserve(bdev, mem);
+		if (ret == -EAGAIN) {
+			ret = ttm_mem_io_evict(man);
+			if (ret == 0)
+				goto retry;
+		}
+	}
+	return ret;
+}
+
+static void ttm_mem_io_free(struct ttm_bo_device *bdev,
+			    struct ttm_mem_reg *mem)
+{
+	struct ttm_mem_type_manager *man = &bdev->man[mem->mem_type];
+
+	if (likely(man->io_reserve_fastpath))
+		return;
+
+	if (bdev->driver->io_mem_reserve &&
+	    --mem->bus.io_reserved_count == 0 &&
+	    bdev->driver->io_mem_free)
+		bdev->driver->io_mem_free(bdev, mem);
+
+}
+
+int ttm_mem_io_reserve_vm(struct ttm_buffer_object *bo)
+{
+	struct ttm_mem_reg *mem = &bo->mem;
+	int ret;
+
+	if (!mem->bus.io_reserved_vm) {
+		struct ttm_mem_type_manager *man =
+			&bo->bdev->man[mem->mem_type];
+
+		ret = ttm_mem_io_reserve(bo->bdev, mem);
 		if (unlikely(ret != 0))
 			return ret;
+		mem->bus.io_reserved_vm = true;
+		if (man->use_io_reserve_lru)
+			list_add_tail(&bo->io_reserve_lru,
+				      &man->io_reserve_lru);
 	}
 	return 0;
 }
 
-void ttm_mem_io_free(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem)
+void ttm_mem_io_free_vm(struct ttm_buffer_object *bo)
 {
-	if (bdev->driver->io_mem_reserve) {
-		if (mem->bus.io_reserved) {
-			mem->bus.io_reserved = false;
-			bdev->driver->io_mem_free(bdev, mem);
-		}
+	struct ttm_mem_reg *mem = &bo->mem;
+
+	if (mem->bus.io_reserved_vm) {
+		mem->bus.io_reserved_vm = false;
+		list_del_init(&bo->io_reserve_lru);
+		ttm_mem_io_free(bo->bdev, mem);
 	}
 }
 
 int ttm_mem_reg_ioremap(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem,
 			void **virtual)
 {
+	struct ttm_mem_type_manager *man = &bdev->man[mem->mem_type];
 	int ret;
 	void *addr;
 
 	*virtual = NULL;
+	(void) ttm_mem_io_lock(man, false);
 	ret = ttm_mem_io_reserve(bdev, mem);
+	ttm_mem_io_unlock(man);
 	if (ret || !mem->bus.is_iomem)
 		return ret;
 
@@ -117,7 +203,9 @@ int ttm_mem_reg_ioremap(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem,
 		else
 			addr = ioremap_nocache(mem->bus.base + mem->bus.offset, mem->bus.size);
 		if (!addr) {
+			(void) ttm_mem_io_lock(man, false);
 			ttm_mem_io_free(bdev, mem);
+			ttm_mem_io_unlock(man);
 			return -ENOMEM;
 		}
 	}
@@ -134,7 +222,9 @@ void ttm_mem_reg_iounmap(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem,
 
 	if (virtual && mem->bus.addr == NULL)
 		iounmap(virtual);
+	(void) ttm_mem_io_lock(man, false);
 	ttm_mem_io_free(bdev, mem);
+	ttm_mem_io_unlock(man);
 }
 
 static int ttm_copy_io_page(void *dst, void *src, unsigned long page)
@@ -231,7 +321,7 @@ int ttm_bo_move_memcpy(struct ttm_buffer_object *bo,
 	struct ttm_mem_type_manager *man = &bdev->man[new_mem->mem_type];
 	struct ttm_tt *ttm = bo->ttm;
 	struct ttm_mem_reg *old_mem = &bo->mem;
-	struct ttm_mem_reg old_copy = *old_mem;
+	struct ttm_mem_reg old_copy;
 	void *old_iomap;
 	void *new_iomap;
 	int ret;
@@ -280,8 +370,7 @@ int ttm_bo_move_memcpy(struct ttm_buffer_object *bo,
 	}
 	mb();
 out2:
-	ttm_bo_free_old_node(bo);
-
+	old_copy = *old_mem;
 	*old_mem = *new_mem;
 	new_mem->mm_node = NULL;
 
@@ -292,9 +381,10 @@ out2:
 	}
 
 out1:
-	ttm_mem_reg_iounmap(bdev, new_mem, new_iomap);
+	ttm_mem_reg_iounmap(bdev, old_mem, new_iomap);
 out:
 	ttm_mem_reg_iounmap(bdev, &old_copy, old_iomap);
+	ttm_bo_mem_put(bo, &old_copy);
 	return ret;
 }
 EXPORT_SYMBOL(ttm_bo_move_memcpy);
@@ -337,11 +427,11 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 	 * TODO: Explicit member copy would probably be better here.
 	 */
 
-	spin_lock_init(&fbo->lock);
 	init_waitqueue_head(&fbo->event_queue);
 	INIT_LIST_HEAD(&fbo->ddestroy);
 	INIT_LIST_HEAD(&fbo->lru);
 	INIT_LIST_HEAD(&fbo->swap);
+	INIT_LIST_HEAD(&fbo->io_reserve_lru);
 	fbo->vm_node = NULL;
 	atomic_set(&fbo->cpu_writers, 0);
 
@@ -453,6 +543,8 @@ int ttm_bo_kmap(struct ttm_buffer_object *bo,
 		unsigned long start_page, unsigned long num_pages,
 		struct ttm_bo_kmap_obj *map)
 {
+	struct ttm_mem_type_manager *man =
+		&bo->bdev->man[bo->mem.mem_type];
 	unsigned long offset, size;
 	int ret;
 
@@ -467,7 +559,9 @@ int ttm_bo_kmap(struct ttm_buffer_object *bo,
 	if (num_pages > 1 && !DRM_SUSER(DRM_CURPROC))
 		return -EPERM;
 #endif
+	(void) ttm_mem_io_lock(man, false);
 	ret = ttm_mem_io_reserve(bo->bdev, &bo->mem);
+	ttm_mem_io_unlock(man);
 	if (ret)
 		return ret;
 	if (!bo->mem.bus.is_iomem) {
@@ -482,12 +576,15 @@ EXPORT_SYMBOL(ttm_bo_kmap);
 
 void ttm_bo_kunmap(struct ttm_bo_kmap_obj *map)
 {
+	struct ttm_buffer_object *bo = map->bo;
+	struct ttm_mem_type_manager *man =
+		&bo->bdev->man[bo->mem.mem_type];
+
 	if (!map->virtual)
 		return;
 	switch (map->bo_kmap_type) {
 	case ttm_bo_map_iomap:
 		iounmap(map->virtual);
-		ttm_mem_io_free(map->bo->bdev, &map->bo->mem);
 		break;
 	case ttm_bo_map_vmap:
 		vunmap(map->virtual);
@@ -500,6 +597,9 @@ void ttm_bo_kunmap(struct ttm_bo_kmap_obj *map)
 	default:
 		BUG();
 	}
+	(void) ttm_mem_io_lock(man, false);
+	ttm_mem_io_free(map->bo->bdev, &map->bo->mem);
+	ttm_mem_io_unlock(man);
 	map->virtual = NULL;
 	map->page = NULL;
 }
@@ -520,7 +620,7 @@ int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo,
 	struct ttm_buffer_object *ghost_obj;
 	void *tmp_obj = NULL;
 
-	spin_lock(&bo->lock);
+	spin_lock(&bdev->fence_lock);
 	if (bo->sync_obj) {
 		tmp_obj = bo->sync_obj;
 		bo->sync_obj = NULL;
@@ -529,7 +629,7 @@ int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo,
 	bo->sync_obj_arg = sync_obj_arg;
 	if (evict) {
 		ret = ttm_bo_wait(bo, false, false, false);
-		spin_unlock(&bo->lock);
+		spin_unlock(&bdev->fence_lock);
 		if (tmp_obj)
 			driver->sync_obj_unref(&tmp_obj);
 		if (ret)
@@ -552,7 +652,7 @@ int ttm_bo_move_accel_cleanup(struct ttm_buffer_object *bo,
 		 */
 
 		set_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags);
-		spin_unlock(&bo->lock);
+		spin_unlock(&bdev->fence_lock);
 		if (tmp_obj)
 			driver->sync_obj_unref(&tmp_obj);
 
diff --git a/drivers/gpu/drm/ttm/ttm_bo_vm.c b/drivers/gpu/drm/ttm/ttm_bo_vm.c
index fe6cb77899f4..221b924acebe 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_vm.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_vm.c
@@ -83,6 +83,8 @@ static int ttm_bo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 	int i;
 	unsigned long address = (unsigned long)vmf->virtual_address;
 	int retval = VM_FAULT_NOPAGE;
+	struct ttm_mem_type_manager *man =
+		&bdev->man[bo->mem.mem_type];
 
 	/*
 	 * Work around locking order reversal in fault / nopfn
@@ -118,24 +120,28 @@ static int ttm_bo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 	 * move.
 	 */
 
-	spin_lock(&bo->lock);
+	spin_lock(&bdev->fence_lock);
 	if (test_bit(TTM_BO_PRIV_FLAG_MOVING, &bo->priv_flags)) {
 		ret = ttm_bo_wait(bo, false, true, false);
-		spin_unlock(&bo->lock);
+		spin_unlock(&bdev->fence_lock);
 		if (unlikely(ret != 0)) {
 			retval = (ret != -ERESTARTSYS) ?
 			    VM_FAULT_SIGBUS : VM_FAULT_NOPAGE;
 			goto out_unlock;
 		}
 	} else
-		spin_unlock(&bo->lock);
+		spin_unlock(&bdev->fence_lock);
 
-
-	ret = ttm_mem_io_reserve(bdev, &bo->mem);
-	if (ret) {
-		retval = VM_FAULT_SIGBUS;
+	ret = ttm_mem_io_lock(man, true);
+	if (unlikely(ret != 0)) {
+		retval = VM_FAULT_NOPAGE;
 		goto out_unlock;
 	}
+	ret = ttm_mem_io_reserve_vm(bo);
+	if (unlikely(ret != 0)) {
+		retval = VM_FAULT_SIGBUS;
+		goto out_io_unlock;
+	}
 
 	page_offset = ((address - vma->vm_start) >> PAGE_SHIFT) +
 	    bo->vm_node->start - vma->vm_pgoff;
@@ -144,7 +150,7 @@ static int ttm_bo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 
 	if (unlikely(page_offset >= bo->num_pages)) {
 		retval = VM_FAULT_SIGBUS;
-		goto out_unlock;
+		goto out_io_unlock;
 	}
 
 	/*
@@ -182,7 +188,7 @@ static int ttm_bo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 			page = ttm_tt_get_page(ttm, page_offset);
 			if (unlikely(!page && i == 0)) {
 				retval = VM_FAULT_OOM;
-				goto out_unlock;
+				goto out_io_unlock;
 			} else if (unlikely(!page)) {
 				break;
 			}
@@ -200,14 +206,15 @@ static int ttm_bo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
 		else if (unlikely(ret != 0)) {
 			retval =
 			    (ret == -ENOMEM) ? VM_FAULT_OOM : VM_FAULT_SIGBUS;
-			goto out_unlock;
+			goto out_io_unlock;
 		}
 
 		address += PAGE_SIZE;
 		if (unlikely(++page_offset >= page_last))
 			break;
 	}
-
+out_io_unlock:
+	ttm_mem_io_unlock(man);
 out_unlock:
 	ttm_bo_unreserve(bo);
 	return retval;
diff --git a/drivers/gpu/drm/ttm/ttm_execbuf_util.c b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
index c285c2902d15..3832fe10b4df 100644
--- a/drivers/gpu/drm/ttm/ttm_execbuf_util.c
+++ b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
@@ -32,7 +32,7 @@
 #include <linux/sched.h>
 #include <linux/module.h>
 
-void ttm_eu_backoff_reservation(struct list_head *list)
+static void ttm_eu_backoff_reservation_locked(struct list_head *list)
 {
 	struct ttm_validate_buffer *entry;
 
@@ -41,10 +41,77 @@ void ttm_eu_backoff_reservation(struct list_head *list)
 		if (!entry->reserved)
 			continue;
 
+		if (entry->removed) {
+			ttm_bo_add_to_lru(bo);
+			entry->removed = false;
+
+		}
 		entry->reserved = false;
-		ttm_bo_unreserve(bo);
+		atomic_set(&bo->reserved, 0);
+		wake_up_all(&bo->event_queue);
+	}
+}
+
+static void ttm_eu_del_from_lru_locked(struct list_head *list)
+{
+	struct ttm_validate_buffer *entry;
+
+	list_for_each_entry(entry, list, head) {
+		struct ttm_buffer_object *bo = entry->bo;
+		if (!entry->reserved)
+			continue;
+
+		if (!entry->removed) {
+			entry->put_count = ttm_bo_del_from_lru(bo);
+			entry->removed = true;
+		}
 	}
 }
+
+static void ttm_eu_list_ref_sub(struct list_head *list)
+{
+	struct ttm_validate_buffer *entry;
+
+	list_for_each_entry(entry, list, head) {
+		struct ttm_buffer_object *bo = entry->bo;
+
+		if (entry->put_count) {
+			ttm_bo_list_ref_sub(bo, entry->put_count, true);
+			entry->put_count = 0;
+		}
+	}
+}
+
+static int ttm_eu_wait_unreserved_locked(struct list_head *list,
+					 struct ttm_buffer_object *bo)
+{
+	struct ttm_bo_global *glob = bo->glob;
+	int ret;
+
+	ttm_eu_del_from_lru_locked(list);
+	spin_unlock(&glob->lru_lock);
+	ret = ttm_bo_wait_unreserved(bo, true);
+	spin_lock(&glob->lru_lock);
+	if (unlikely(ret != 0))
+		ttm_eu_backoff_reservation_locked(list);
+	return ret;
+}
+
+
+void ttm_eu_backoff_reservation(struct list_head *list)
+{
+	struct ttm_validate_buffer *entry;
+	struct ttm_bo_global *glob;
+
+	if (list_empty(list))
+		return;
+
+	entry = list_first_entry(list, struct ttm_validate_buffer, head);
+	glob = entry->bo->glob;
+	spin_lock(&glob->lru_lock);
+	ttm_eu_backoff_reservation_locked(list);
+	spin_unlock(&glob->lru_lock);
+}
 EXPORT_SYMBOL(ttm_eu_backoff_reservation);
 
 /*
@@ -59,37 +126,76 @@ EXPORT_SYMBOL(ttm_eu_backoff_reservation);
  * buffers in different orders.
  */
 
-int ttm_eu_reserve_buffers(struct list_head *list, uint32_t val_seq)
+int ttm_eu_reserve_buffers(struct list_head *list)
 {
+	struct ttm_bo_global *glob;
 	struct ttm_validate_buffer *entry;
 	int ret;
+	uint32_t val_seq;
+
+	if (list_empty(list))
+		return 0;
+
+	list_for_each_entry(entry, list, head) {
+		entry->reserved = false;
+		entry->put_count = 0;
+		entry->removed = false;
+	}
+
+	entry = list_first_entry(list, struct ttm_validate_buffer, head);
+	glob = entry->bo->glob;
 
 retry:
+	spin_lock(&glob->lru_lock);
+	val_seq = entry->bo->bdev->val_seq++;
+
 	list_for_each_entry(entry, list, head) {
 		struct ttm_buffer_object *bo = entry->bo;
 
-		entry->reserved = false;
-		ret = ttm_bo_reserve(bo, true, false, true, val_seq);
-		if (ret != 0) {
-			ttm_eu_backoff_reservation(list);
-			if (ret == -EAGAIN) {
-				ret = ttm_bo_wait_unreserved(bo, true);
-				if (unlikely(ret != 0))
-					return ret;
-				goto retry;
-			} else
+retry_this_bo:
+		ret = ttm_bo_reserve_locked(bo, true, true, true, val_seq);
+		switch (ret) {
+		case 0:
+			break;
+		case -EBUSY:
+			ret = ttm_eu_wait_unreserved_locked(list, bo);
+			if (unlikely(ret != 0)) {
+				spin_unlock(&glob->lru_lock);
+				ttm_eu_list_ref_sub(list);
 				return ret;
+			}
+			goto retry_this_bo;
+		case -EAGAIN:
+			ttm_eu_backoff_reservation_locked(list);
+			spin_unlock(&glob->lru_lock);
+			ttm_eu_list_ref_sub(list);
+			ret = ttm_bo_wait_unreserved(bo, true);
+			if (unlikely(ret != 0))
+				return ret;
+			goto retry;
+		default:
+			ttm_eu_backoff_reservation_locked(list);
+			spin_unlock(&glob->lru_lock);
+			ttm_eu_list_ref_sub(list);
+			return ret;
 		}
 
 		entry->reserved = true;
 		if (unlikely(atomic_read(&bo->cpu_writers) > 0)) {
-			ttm_eu_backoff_reservation(list);
+			ttm_eu_backoff_reservation_locked(list);
+			spin_unlock(&glob->lru_lock);
+			ttm_eu_list_ref_sub(list);
 			ret = ttm_bo_wait_cpu(bo, false);
 			if (ret)
 				return ret;
 			goto retry;
 		}
 	}
+
+	ttm_eu_del_from_lru_locked(list);
+	spin_unlock(&glob->lru_lock);
+	ttm_eu_list_ref_sub(list);
+
 	return 0;
 }
 EXPORT_SYMBOL(ttm_eu_reserve_buffers);
@@ -97,21 +203,36 @@ EXPORT_SYMBOL(ttm_eu_reserve_buffers);
 void ttm_eu_fence_buffer_objects(struct list_head *list, void *sync_obj)
 {
 	struct ttm_validate_buffer *entry;
+	struct ttm_buffer_object *bo;
+	struct ttm_bo_global *glob;
+	struct ttm_bo_device *bdev;
+	struct ttm_bo_driver *driver;
 
-	list_for_each_entry(entry, list, head) {
-		struct ttm_buffer_object *bo = entry->bo;
-		struct ttm_bo_driver *driver = bo->bdev->driver;
-		void *old_sync_obj;
+	if (list_empty(list))
+		return;
+
+	bo = list_first_entry(list, struct ttm_validate_buffer, head)->bo;
+	bdev = bo->bdev;
+	driver = bdev->driver;
+	glob = bo->glob;
 
-		spin_lock(&bo->lock);
-		old_sync_obj = bo->sync_obj;
+	spin_lock(&bdev->fence_lock);
+	spin_lock(&glob->lru_lock);
+
+	list_for_each_entry(entry, list, head) {
+		bo = entry->bo;
+		entry->old_sync_obj = bo->sync_obj;
 		bo->sync_obj = driver->sync_obj_ref(sync_obj);
 		bo->sync_obj_arg = entry->new_sync_obj_arg;
-		spin_unlock(&bo->lock);
-		ttm_bo_unreserve(bo);
+		ttm_bo_unreserve_locked(bo);
 		entry->reserved = false;
-		if (old_sync_obj)
-			driver->sync_obj_unref(&old_sync_obj);
+	}
+	spin_unlock(&glob->lru_lock);
+	spin_unlock(&bdev->fence_lock);
+
+	list_for_each_entry(entry, list, head) {
+		if (entry->old_sync_obj)
+			driver->sync_obj_unref(&entry->old_sync_obj);
 	}
 }
 EXPORT_SYMBOL(ttm_eu_fence_buffer_objects);