Merge tag 'drm-intel-next-2016-08-08' of git://anongit.freedesktop.org/drm-intel into drm-next

- refactor ddi buffer programming a bit (Ville)
- large-scale renaming to untangle naming in the gem code (Chris)
- rework vma/active tracking for accurately reaping idle mappings of shared
  objects (Chris)
- misc dp sst/mst probing corner case fixes (Ville)
- tons of cleanup&tunings all around in gem
- lockless (rcu-protected) request lookup, plus use it everywhere for
  non(b)locking waits (Chris)
- pipe crc debugfs fixes (Rodrigo)
- random fixes all over

* tag 'drm-intel-next-2016-08-08' of git://anongit.freedesktop.org/drm-intel: (222 commits)
  drm/i915: Update DRIVER_DATE to 20160808
  drm/i915: fix aliasing_ppgtt leak
  drm/i915: Update comment before i915_spin_request
  drm/i915: Use drm official vblank_no_hw_counter callback.
  drm/i915: Fix copy_to_user usage for pipe_crc
  Revert "drm/i915: Track active streams also for DP SST"
  drm/i915: fix WaInsertDummyPushConstPs
  drm/i915: Assert that the request hasn't been retired
  drm/i915: Repack fence tiling mode and stride into a single integer
  drm/i915: Document and reject invalid tiling modes
  drm/i915: Remove locking for get_tiling
  drm/i915: Remove pinned check from madvise ioctl
  drm/i915: Reduce locking inside swfinish ioctl
  drm/i915: Remove (struct_mutex) locking for busy-ioctl
  drm/i915: Remove (struct_mutex) locking for wait-ioctl
  drm/i915: Do a nonblocking wait first in pread/pwrite
  drm/i915: Remove unused no-shrinker-steal
  drm/i915: Tidy generation of the GTT mmap offset
  drm/i915/shrinker: Wait before acquiring struct_mutex under oom
  drm/i915: Simplify do_idling() (Ironlake vt-d w/a)
  ...

Documentation/gpu/i915.rst

@@ -70,6 +70,9 @@ Frontbuffer Tracking
 .. kernel-doc:: drivers/gpu/drm/i915/intel_frontbuffer.c
    :doc: frontbuffer tracking
 
+.. kernel-doc:: drivers/gpu/drm/i915/intel_frontbuffer.h
+   :internal:
+
 .. kernel-doc:: drivers/gpu/drm/i915/intel_frontbuffer.c
    :internal:
 

drivers/gpu/drm/i915/Makefile

@@ -25,7 +25,6 @@ i915-$(CONFIG_DEBUG_FS) += i915_debugfs.o
 i915-y += i915_cmd_parser.o \
 	  i915_gem_batch_pool.o \
 	  i915_gem_context.o \
-	  i915_gem_debug.o \
 	  i915_gem_dmabuf.o \
 	  i915_gem_evict.o \
 	  i915_gem_execbuffer.o \
@@ -33,6 +32,7 @@ i915-y += i915_cmd_parser.o \
 	  i915_gem_gtt.o \
 	  i915_gem.o \
 	  i915_gem_render_state.o \
+	  i915_gem_request.o \
 	  i915_gem_shrinker.o \
 	  i915_gem_stolen.o \
 	  i915_gem_tiling.o \
@@ -40,6 +40,7 @@ i915-y += i915_cmd_parser.o \
 	  i915_gpu_error.o \
 	  i915_trace_points.o \
 	  intel_breadcrumbs.o \
+	  intel_engine_cs.o \
 	  intel_lrc.o \
 	  intel_mocs.o \
 	  intel_ringbuffer.o \

drivers/gpu/drm/i915/i915_cmd_parser.c

@@ -62,23 +62,23 @@
  * The parser always rejects such commands.
  *
  * The majority of the problematic commands fall in the MI_* range, with only a
- * few specific commands on each ring (e.g. PIPE_CONTROL and MI_FLUSH_DW).
+ * few specific commands on each engine (e.g. PIPE_CONTROL and MI_FLUSH_DW).
  *
  * Implementation:
- * Each ring maintains tables of commands and registers which the parser uses in
- * scanning batch buffers submitted to that ring.
+ * Each engine maintains tables of commands and registers which the parser
+ * uses in scanning batch buffers submitted to that engine.
  *
  * Since the set of commands that the parser must check for is significantly
  * smaller than the number of commands supported, the parser tables contain only
  * those commands required by the parser. This generally works because command
  * opcode ranges have standard command length encodings. So for commands that
  * the parser does not need to check, it can easily skip them. This is
- * implemented via a per-ring length decoding vfunc.
+ * implemented via a per-engine length decoding vfunc.
  *
  * Unfortunately, there are a number of commands that do not follow the standard
  * length encoding for their opcode range, primarily amongst the MI_* commands.
  * To handle this, the parser provides a way to define explicit "skip" entries
- * in the per-ring command tables.
+ * in the per-engine command tables.
  *
  * Other command table entries map fairly directly to high level categories
  * mentioned above: rejected, master-only, register whitelist. The parser
@@ -603,7 +603,7 @@ static u32 gen7_blt_get_cmd_length_mask(u32 cmd_header)
 	return 0;
 }
 
-static bool validate_cmds_sorted(struct intel_engine_cs *engine,
+static bool validate_cmds_sorted(const struct intel_engine_cs *engine,
 				 const struct drm_i915_cmd_table *cmd_tables,
 				 int cmd_table_count)
 {
@@ -624,8 +624,10 @@ static bool validate_cmds_sorted(struct intel_engine_cs *engine,
 			u32 curr = desc->cmd.value & desc->cmd.mask;
 
 			if (curr < previous) {
-				DRM_ERROR("CMD: table not sorted ring=%d table=%d entry=%d cmd=0x%08X prev=0x%08X\n",
-					  engine->id, i, j, curr, previous);
+				DRM_ERROR("CMD: %s [%d] command table not sorted: "
+					  "table=%d entry=%d cmd=0x%08X prev=0x%08X\n",
+					  engine->name, engine->id,
+					  i, j, curr, previous);
 				ret = false;
 			}
 
@@ -636,7 +638,7 @@ static bool validate_cmds_sorted(struct intel_engine_cs *engine,
 	return ret;
 }
 
-static bool check_sorted(int ring_id,
+static bool check_sorted(const struct intel_engine_cs *engine,
 			 const struct drm_i915_reg_descriptor *reg_table,
 			 int reg_count)
 {
@@ -648,8 +650,10 @@ static bool check_sorted(int ring_id,
 		u32 curr = i915_mmio_reg_offset(reg_table[i].addr);
 
 		if (curr < previous) {
-			DRM_ERROR("CMD: table not sorted ring=%d entry=%d reg=0x%08X prev=0x%08X\n",
-				  ring_id, i, curr, previous);
+			DRM_ERROR("CMD: %s [%d] register table not sorted: "
+				  "entry=%d reg=0x%08X prev=0x%08X\n",
+				  engine->name, engine->id,
+				  i, curr, previous);
 			ret = false;
 		}
 
@@ -666,7 +670,7 @@ static bool validate_regs_sorted(struct intel_engine_cs *engine)
 
 	for (i = 0; i < engine->reg_table_count; i++) {
 		table = &engine->reg_tables[i];
-		if (!check_sorted(engine->id, table->regs, table->num_regs))
+		if (!check_sorted(engine, table->regs, table->num_regs))
 			return false;
 	}
 
@@ -736,7 +740,7 @@ static void fini_hash_table(struct intel_engine_cs *engine)
 }
 
 /**
- * i915_cmd_parser_init_ring() - set cmd parser related fields for a ringbuffer
+ * intel_engine_init_cmd_parser() - set cmd parser related fields for an engine
  * @engine: the engine to initialize
  *
  * Optionally initializes fields related to batch buffer command parsing in the
@@ -745,7 +749,7 @@ static void fini_hash_table(struct intel_engine_cs *engine)
  *
  * Return: non-zero if initialization fails
  */
-int i915_cmd_parser_init_ring(struct intel_engine_cs *engine)
+int intel_engine_init_cmd_parser(struct intel_engine_cs *engine)
 {
 	const struct drm_i915_cmd_table *cmd_tables;
 	int cmd_table_count;
@@ -806,8 +810,7 @@ int i915_cmd_parser_init_ring(struct intel_engine_cs *engine)
 		engine->get_cmd_length_mask = gen7_bsd_get_cmd_length_mask;
 		break;
 	default:
-		DRM_ERROR("CMD: cmd_parser_init with unknown ring: %d\n",
-			  engine->id);
+		MISSING_CASE(engine->id);
 		BUG();
 	}
 
@@ -829,13 +832,13 @@ int i915_cmd_parser_init_ring(struct intel_engine_cs *engine)
 }
 
 /**
- * i915_cmd_parser_fini_ring() - clean up cmd parser related fields
+ * intel_engine_cleanup_cmd_parser() - clean up cmd parser related fields
  * @engine: the engine to clean up
  *
  * Releases any resources related to command parsing that may have been
- * initialized for the specified ring.
+ * initialized for the specified engine.
  */
-void i915_cmd_parser_fini_ring(struct intel_engine_cs *engine)
+void intel_engine_cleanup_cmd_parser(struct intel_engine_cs *engine)
 {
 	if (!engine->needs_cmd_parser)
 		return;
@@ -866,9 +869,9 @@ find_cmd_in_table(struct intel_engine_cs *engine,
  * Returns a pointer to a descriptor for the command specified by cmd_header.
  *
  * The caller must supply space for a default descriptor via the default_desc
- * parameter. If no descriptor for the specified command exists in the ring's
+ * parameter. If no descriptor for the specified command exists in the engine's
  * command parser tables, this function fills in default_desc based on the
- * ring's default length encoding and returns default_desc.
+ * engine's default length encoding and returns default_desc.
  */
 static const struct drm_i915_cmd_descriptor*
 find_cmd(struct intel_engine_cs *engine,
@@ -1023,15 +1026,16 @@ static u32 *copy_batch(struct drm_i915_gem_object *dest_obj,
 }
 
 /**
- * i915_needs_cmd_parser() - should a given ring use software command parsing?
+ * intel_engine_needs_cmd_parser() - should a given engine use software
+ *                                   command parsing?
  * @engine: the engine in question
  *
  * Only certain platforms require software batch buffer command parsing, and
  * only when enabled via module parameter.
  *
- * Return: true if the ring requires software command parsing
+ * Return: true if the engine requires software command parsing
  */
-bool i915_needs_cmd_parser(struct intel_engine_cs *engine)
+bool intel_engine_needs_cmd_parser(struct intel_engine_cs *engine)
 {
 	if (!engine->needs_cmd_parser)
 		return false;
@@ -1078,8 +1082,8 @@ static bool check_cmd(const struct intel_engine_cs *engine,
 						   reg_addr);
 
 			if (!reg) {
-				DRM_DEBUG_DRIVER("CMD: Rejected register 0x%08X in command: 0x%08X (ring=%d)\n",
-						 reg_addr, *cmd, engine->id);
+				DRM_DEBUG_DRIVER("CMD: Rejected register 0x%08X in command: 0x%08X (exec_id=%d)\n",
+						 reg_addr, *cmd, engine->exec_id);
 				return false;
 			}
 
@@ -1159,11 +1163,11 @@ static bool check_cmd(const struct intel_engine_cs *engine,
 				desc->bits[i].mask;
 
 			if (dword != desc->bits[i].expected) {
-				DRM_DEBUG_DRIVER("CMD: Rejected command 0x%08X for bitmask 0x%08X (exp=0x%08X act=0x%08X) (ring=%d)\n",
+				DRM_DEBUG_DRIVER("CMD: Rejected command 0x%08X for bitmask 0x%08X (exp=0x%08X act=0x%08X) (exec_id=%d)\n",
 						 *cmd,
 						 desc->bits[i].mask,
 						 desc->bits[i].expected,
-						 dword, engine->id);
+						 dword, engine->exec_id);
 				return false;
 			}
 		}
@@ -1189,12 +1193,12 @@ static bool check_cmd(const struct intel_engine_cs *engine,
  * Return: non-zero if the parser finds violations or otherwise fails; -EACCES
  * if the batch appears legal but should use hardware parsing
  */
-int i915_parse_cmds(struct intel_engine_cs *engine,
-		    struct drm_i915_gem_object *batch_obj,
-		    struct drm_i915_gem_object *shadow_batch_obj,
-		    u32 batch_start_offset,
-		    u32 batch_len,
-		    bool is_master)
+int intel_engine_cmd_parser(struct intel_engine_cs *engine,
+			    struct drm_i915_gem_object *batch_obj,
+			    struct drm_i915_gem_object *shadow_batch_obj,
+			    u32 batch_start_offset,
+			    u32 batch_len,
+			    bool is_master)
 {
 	u32 *cmd, *batch_base, *batch_end;
 	struct drm_i915_cmd_descriptor default_desc = { 0 };
@@ -1295,7 +1299,7 @@ int i915_cmd_parser_get_version(struct drm_i915_private *dev_priv)
 
 	/* If the command parser is not enabled, report 0 - unsupported */
 	for_each_engine(engine, dev_priv) {
-		if (i915_needs_cmd_parser(engine)) {
+		if (intel_engine_needs_cmd_parser(engine)) {
 			active = true;
 			break;
 		}

drivers/gpu/drm/i915/i915_drv.c

@@ -228,27 +228,6 @@ static void intel_detect_pch(struct drm_device *dev)
 	pci_dev_put(pch);
 }
 
-bool i915_semaphore_is_enabled(struct drm_i915_private *dev_priv)
-{
-	if (INTEL_GEN(dev_priv) < 6)
-		return false;
-
-	if (i915.semaphores >= 0)
-		return i915.semaphores;
-
-	/* TODO: make semaphores and Execlists play nicely together */
-	if (i915.enable_execlists)
-		return false;
-
-#ifdef CONFIG_INTEL_IOMMU
-	/* Enable semaphores on SNB when IO remapping is off */
-	if (IS_GEN6(dev_priv) && intel_iommu_gfx_mapped)
-		return false;
-#endif
-
-	return true;
-}
-
 static int i915_getparam(struct drm_device *dev, void *data,
 			 struct drm_file *file_priv)
 {
@@ -324,7 +303,7 @@ static int i915_getparam(struct drm_device *dev, void *data,
 		value = 1;
 		break;
 	case I915_PARAM_HAS_SEMAPHORES:
-		value = i915_semaphore_is_enabled(dev_priv);
+		value = i915.semaphores;
 		break;
 	case I915_PARAM_HAS_PRIME_VMAP_FLUSH:
 		value = 1;
@@ -999,6 +978,9 @@ static void intel_sanitize_options(struct drm_i915_private *dev_priv)
 	i915.enable_ppgtt =
 		intel_sanitize_enable_ppgtt(dev_priv, i915.enable_ppgtt);
 	DRM_DEBUG_DRIVER("ppgtt mode: %i\n", i915.enable_ppgtt);
+
+	i915.semaphores = intel_sanitize_semaphores(dev_priv, i915.semaphores);
+	DRM_DEBUG_DRIVER("use GPU sempahores? %s\n", yesno(i915.semaphores));
 }
 
 /**
@@ -1011,8 +993,6 @@ static void intel_sanitize_options(struct drm_i915_private *dev_priv)
 static int i915_driver_init_hw(struct drm_i915_private *dev_priv)
 {
 	struct drm_device *dev = &dev_priv->drm;
-	struct i915_ggtt *ggtt = &dev_priv->ggtt;
-	uint32_t aperture_size;
 	int ret;
 
 	if (i915_inject_load_failure())
@@ -1022,16 +1002,10 @@ static int i915_driver_init_hw(struct drm_i915_private *dev_priv)
 
 	intel_sanitize_options(dev_priv);
 
-	ret = i915_ggtt_init_hw(dev);
+	ret = i915_ggtt_probe_hw(dev_priv);
 	if (ret)
 		return ret;
 
-	ret = i915_ggtt_enable_hw(dev);
-	if (ret) {
-		DRM_ERROR("failed to enable GGTT\n");
-		goto out_ggtt;
-	}
-
 	/* WARNING: Apparently we must kick fbdev drivers before vgacon,
 	 * otherwise the vga fbdev driver falls over. */
 	ret = i915_kick_out_firmware_fb(dev_priv);
@@ -1046,6 +1020,16 @@ static int i915_driver_init_hw(struct drm_i915_private *dev_priv)
 		goto out_ggtt;
 	}
 
+	ret = i915_ggtt_init_hw(dev_priv);
+	if (ret)
+		return ret;
+
+	ret = i915_ggtt_enable_hw(dev_priv);
+	if (ret) {
+		DRM_ERROR("failed to enable GGTT\n");
+		goto out_ggtt;
+	}
+
 	pci_set_master(dev->pdev);
 
 	/* overlay on gen2 is broken and can't address above 1G */
@@ -1058,7 +1042,6 @@ static int i915_driver_init_hw(struct drm_i915_private *dev_priv)
 		}
 	}
 
-
 	/* 965GM sometimes incorrectly writes to hardware status page (HWS)
 	 * using 32bit addressing, overwriting memory if HWS is located
 	 * above 4GB.
@@ -1077,19 +1060,6 @@ static int i915_driver_init_hw(struct drm_i915_private *dev_priv)
 		}
 	}
 
-	aperture_size = ggtt->mappable_end;
-
-	ggtt->mappable =
-		io_mapping_create_wc(ggtt->mappable_base,
-				     aperture_size);
-	if (!ggtt->mappable) {
-		ret = -EIO;
-		goto out_ggtt;
-	}
-
-	ggtt->mtrr = arch_phys_wc_add(ggtt->mappable_base,
-					      aperture_size);
-
 	pm_qos_add_request(&dev_priv->pm_qos, PM_QOS_CPU_DMA_LATENCY,
 			   PM_QOS_DEFAULT_VALUE);
 
@@ -1118,7 +1088,7 @@ static int i915_driver_init_hw(struct drm_i915_private *dev_priv)
 	return 0;
 
 out_ggtt:
-	i915_ggtt_cleanup_hw(dev);
+	i915_ggtt_cleanup_hw(dev_priv);
 
 	return ret;
 }
@@ -1130,15 +1100,12 @@ static int i915_driver_init_hw(struct drm_i915_private *dev_priv)
 static void i915_driver_cleanup_hw(struct drm_i915_private *dev_priv)
 {
 	struct drm_device *dev = &dev_priv->drm;
-	struct i915_ggtt *ggtt = &dev_priv->ggtt;
 
 	if (dev->pdev->msi_enabled)
 		pci_disable_msi(dev->pdev);
 
 	pm_qos_remove_request(&dev_priv->pm_qos);
-	arch_phys_wc_del(ggtt->mtrr);
-	io_mapping_free(ggtt->mappable);
-	i915_ggtt_cleanup_hw(dev);
+	i915_ggtt_cleanup_hw(dev_priv);
 }
 
 /**
@@ -1343,7 +1310,7 @@ void i915_driver_unload(struct drm_device *dev)
 	i915_destroy_error_state(dev);
 
 	/* Flush any outstanding unpin_work. */
-	flush_workqueue(dev_priv->wq);
+	drain_workqueue(dev_priv->wq);
 
 	intel_guc_fini(dev);
 	i915_gem_fini(dev);
@@ -1458,8 +1425,6 @@ static int i915_drm_suspend(struct drm_device *dev)
 
 	intel_guc_suspend(dev);
 
-	intel_suspend_gt_powersave(dev_priv);
-
 	intel_display_suspend(dev);
 
 	intel_dp_mst_suspend(dev);
@@ -1586,15 +1551,13 @@ static int i915_drm_resume(struct drm_device *dev)
 
 	disable_rpm_wakeref_asserts(dev_priv);
 
-	ret = i915_ggtt_enable_hw(dev);
+	ret = i915_ggtt_enable_hw(dev_priv);
 	if (ret)
 		DRM_ERROR("failed to re-enable GGTT\n");
 
 	intel_csr_ucode_resume(dev_priv);
 
-	mutex_lock(&dev->struct_mutex);
-	i915_gem_restore_gtt_mappings(dev);
-	mutex_unlock(&dev->struct_mutex);
+	i915_gem_resume(dev);
 
 	i915_restore_state(dev);
 	intel_opregion_setup(dev_priv);
@@ -1652,6 +1615,7 @@ static int i915_drm_resume(struct drm_device *dev)
 
 	intel_opregion_notify_adapter(dev_priv, PCI_D0);
 
+	intel_autoenable_gt_powersave(dev_priv);
 	drm_kms_helper_poll_enable(dev);
 
 	enable_rpm_wakeref_asserts(dev_priv);
@@ -1778,8 +1742,6 @@ int i915_reset(struct drm_i915_private *dev_priv)
 	unsigned reset_counter;
 	int ret;
 
-	intel_reset_gt_powersave(dev_priv);
-
 	mutex_lock(&dev->struct_mutex);
 
 	/* Clear any previous failed attempts at recovery. Time to try again. */
@@ -1835,8 +1797,7 @@ int i915_reset(struct drm_i915_private *dev_priv)
 	 * previous concerns that it doesn't respond well to some forms
 	 * of re-init after reset.
 	 */
-	if (INTEL_INFO(dev)->gen > 5)
-		intel_enable_gt_powersave(dev_priv);
+	intel_autoenable_gt_powersave(dev_priv);
 
 	return 0;
 
@@ -2462,7 +2423,6 @@ static int intel_runtime_resume(struct device *device)
 	 * we can do is to hope that things will still work (and disable RPM).
 	 */
 	i915_gem_init_swizzling(dev);
-	gen6_update_ring_freq(dev_priv);
 
 	intel_runtime_pm_enable_interrupts(dev_priv);
 
@@ -2618,6 +2578,7 @@ static struct drm_driver driver = {
 	.postclose = i915_driver_postclose,
 	.set_busid = drm_pci_set_busid,
 
+	.gem_close_object = i915_gem_close_object,
 	.gem_free_object = i915_gem_free_object,
 	.gem_vm_ops = &i915_gem_vm_ops,
 

drivers/gpu/drm/i915/i915_gem_batch_pool.c

@@ -41,15 +41,15 @@
 
 /**
  * i915_gem_batch_pool_init() - initialize a batch buffer pool
- * @dev: the drm device
+ * @engine: the associated request submission engine
  * @pool: the batch buffer pool
  */
-void i915_gem_batch_pool_init(struct drm_device *dev,
+void i915_gem_batch_pool_init(struct intel_engine_cs *engine,
 			      struct i915_gem_batch_pool *pool)
 {
 	int n;
 
-	pool->dev = dev;
+	pool->engine = engine;
 
 	for (n = 0; n < ARRAY_SIZE(pool->cache_list); n++)
 		INIT_LIST_HEAD(&pool->cache_list[n]);
@@ -65,18 +65,17 @@ void i915_gem_batch_pool_fini(struct i915_gem_batch_pool *pool)
 {
 	int n;
 
-	WARN_ON(!mutex_is_locked(&pool->dev->struct_mutex));
+	lockdep_assert_held(&pool->engine->i915->drm.struct_mutex);
 
 	for (n = 0; n < ARRAY_SIZE(pool->cache_list); n++) {
-		while (!list_empty(&pool->cache_list[n])) {
-			struct drm_i915_gem_object *obj =
-				list_first_entry(&pool->cache_list[n],
-						 struct drm_i915_gem_object,
-						 batch_pool_link);
-
-			list_del(&obj->batch_pool_link);
-			drm_gem_object_unreference(&obj->base);
-		}
+		struct drm_i915_gem_object *obj, *next;
+
+		list_for_each_entry_safe(obj, next,
+					 &pool->cache_list[n],
+					 batch_pool_link)
+			i915_gem_object_put(obj);
+
+		INIT_LIST_HEAD(&pool->cache_list[n]);
 	}
 }
 
@@ -102,7 +101,7 @@ i915_gem_batch_pool_get(struct i915_gem_batch_pool *pool,
 	struct list_head *list;
 	int n;
 
-	WARN_ON(!mutex_is_locked(&pool->dev->struct_mutex));
+	lockdep_assert_held(&pool->engine->i915->drm.struct_mutex);
 
 	/* Compute a power-of-two bucket, but throw everything greater than
 	 * 16KiB into the same bucket: i.e. the the buckets hold objects of
@@ -115,13 +114,14 @@ i915_gem_batch_pool_get(struct i915_gem_batch_pool *pool,
 
 	list_for_each_entry_safe(tmp, next, list, batch_pool_link) {
 		/* The batches are strictly LRU ordered */
-		if (tmp->active)
+		if (!i915_gem_active_is_idle(&tmp->last_read[pool->engine->id],
+					     &tmp->base.dev->struct_mutex))
 			break;
 
 		/* While we're looping, do some clean up */
 		if (tmp->madv == __I915_MADV_PURGED) {
 			list_del(&tmp->batch_pool_link);
-			drm_gem_object_unreference(&tmp->base);
+			i915_gem_object_put(tmp);
 			continue;
 		}
 
@@ -134,7 +134,7 @@ i915_gem_batch_pool_get(struct i915_gem_batch_pool *pool,
 	if (obj == NULL) {
 		int ret;
 
-		obj = i915_gem_object_create(pool->dev, size);
+		obj = i915_gem_object_create(&pool->engine->i915->drm, size);
 		if (IS_ERR(obj))
 			return obj;
 

drivers/gpu/drm/i915/i915_gem_batch_pool.h

@@ -27,13 +27,15 @@
 
 #include "i915_drv.h"
 
+struct intel_engine_cs;
+
 struct i915_gem_batch_pool {
-	struct drm_device *dev;
+	struct intel_engine_cs *engine;
 	struct list_head cache_list[4];
 };
 
 /* i915_gem_batch_pool.c */
-void i915_gem_batch_pool_init(struct drm_device *dev,
+void i915_gem_batch_pool_init(struct intel_engine_cs *engine,
 			      struct i915_gem_batch_pool *pool);
 void i915_gem_batch_pool_fini(struct i915_gem_batch_pool *pool);
 struct drm_i915_gem_object*

drivers/gpu/drm/i915/i915_gem_dmabuf.c

@@ -23,9 +23,13 @@
  * Authors:
  *	Dave Airlie <airlied@redhat.com>
  */
+
+#include <linux/dma-buf.h>
+#include <linux/reservation.h>
+
 #include <drm/drmP.h>
+
 #include "i915_drv.h"
-#include <linux/dma-buf.h>
 
 static struct drm_i915_gem_object *dma_buf_to_obj(struct dma_buf *buf)
 {
@@ -218,25 +222,73 @@ static const struct dma_buf_ops i915_dmabuf_ops =  {
 	.end_cpu_access = i915_gem_end_cpu_access,
 };
 
+static void export_fences(struct drm_i915_gem_object *obj,
+			  struct dma_buf *dma_buf)
+{
+	struct reservation_object *resv = dma_buf->resv;
+	struct drm_i915_gem_request *req;
+	unsigned long active;
+	int idx;
+
+	active = __I915_BO_ACTIVE(obj);
+	if (!active)
+		return;
+
+	/* Serialise with execbuf to prevent concurrent fence-loops */
+	mutex_lock(&obj->base.dev->struct_mutex);
+
+	/* Mark the object for future fences before racily adding old fences */
+	obj->base.dma_buf = dma_buf;
+
+	ww_mutex_lock(&resv->lock, NULL);
+
+	for_each_active(active, idx) {
+		req = i915_gem_active_get(&obj->last_read[idx],
+					  &obj->base.dev->struct_mutex);
+		if (!req)
+			continue;
+
+		if (reservation_object_reserve_shared(resv) == 0)
+			reservation_object_add_shared_fence(resv, &req->fence);
+
+		i915_gem_request_put(req);
+	}
+
+	req = i915_gem_active_get(&obj->last_write,
+				  &obj->base.dev->struct_mutex);
+	if (req) {
+		reservation_object_add_excl_fence(resv, &req->fence);
+		i915_gem_request_put(req);
+	}
+
+	ww_mutex_unlock(&resv->lock);
+	mutex_unlock(&obj->base.dev->struct_mutex);
+}
+
 struct dma_buf *i915_gem_prime_export(struct drm_device *dev,
 				      struct drm_gem_object *gem_obj, int flags)
 {
 	struct drm_i915_gem_object *obj = to_intel_bo(gem_obj);
 	DEFINE_DMA_BUF_EXPORT_INFO(exp_info);
+	struct dma_buf *dma_buf;
 
 	exp_info.ops = &i915_dmabuf_ops;
 	exp_info.size = gem_obj->size;
 	exp_info.flags = flags;
 	exp_info.priv = gem_obj;
 
-
 	if (obj->ops->dmabuf_export) {
 		int ret = obj->ops->dmabuf_export(obj);
 		if (ret)
 			return ERR_PTR(ret);
 	}
 
-	return dma_buf_export(&exp_info);
+	dma_buf = dma_buf_export(&exp_info);
+	if (IS_ERR(dma_buf))
+		return dma_buf;
+
+	export_fences(obj, dma_buf);
+	return dma_buf;
 }
 
 static int i915_gem_object_get_pages_dmabuf(struct drm_i915_gem_object *obj)
@@ -278,8 +330,7 @@ struct drm_gem_object *i915_gem_prime_import(struct drm_device *dev,
 			 * Importing dmabuf exported from out own gem increases
 			 * refcount on gem itself instead of f_count of dmabuf.
 			 */
-			drm_gem_object_reference(&obj->base);
-			return &obj->base;
+			return &i915_gem_object_get(obj)->base;
 		}
 	}
 
@@ -300,6 +351,16 @@ struct drm_gem_object *i915_gem_prime_import(struct drm_device *dev,
 	i915_gem_object_init(obj, &i915_gem_object_dmabuf_ops);
 	obj->base.import_attach = attach;
 
+	/* We use GTT as shorthand for a coherent domain, one that is
+	 * neither in the GPU cache nor in the CPU cache, where all
+	 * writes are immediately visible in memory. (That's not strictly
+	 * true, but it's close! There are internal buffers such as the
+	 * write-combined buffer or a delay through the chipset for GTT
+	 * writes that do require us to treat GTT as a separate cache domain.)
+	 */
+	obj->base.read_domains = I915_GEM_DOMAIN_GTT;
+	obj->base.write_domain = 0;
+
 	return &obj->base;
 
 fail_detach:

drivers/gpu/drm/i915/i915_gem_evict.c

@@ -33,41 +33,23 @@
 #include "intel_drv.h"
 #include "i915_trace.h"
 
-static int switch_to_pinned_context(struct drm_i915_private *dev_priv)
+static bool
+gpu_is_idle(struct drm_i915_private *dev_priv)
 {
 	struct intel_engine_cs *engine;
 
-	if (i915.enable_execlists)
-		return 0;
-
 	for_each_engine(engine, dev_priv) {
-		struct drm_i915_gem_request *req;
-		int ret;
-
-		if (engine->last_context == NULL)
-			continue;
-
-		if (engine->last_context == dev_priv->kernel_context)
-			continue;
-
-		req = i915_gem_request_alloc(engine, dev_priv->kernel_context);
-		if (IS_ERR(req))
-			return PTR_ERR(req);
-
-		ret = i915_switch_context(req);
-		i915_add_request_no_flush(req);
-		if (ret)
-			return ret;
+		if (intel_engine_is_active(engine))
+			return false;
 	}
 
-	return 0;
+	return true;
 }
 
-
 static bool
 mark_free(struct i915_vma *vma, struct list_head *unwind)
 {
-	if (vma->pin_count)
+	if (i915_vma_is_pinned(vma))
 		return false;
 
 	if (WARN_ON(!list_empty(&vma->exec_list)))
@@ -79,7 +61,6 @@ mark_free(struct i915_vma *vma, struct list_head *unwind)
 
 /**
  * i915_gem_evict_something - Evict vmas to make room for binding a new one
- * @dev: drm_device
  * @vm: address space to evict from
  * @min_size: size of the desired free space
  * @alignment: alignment constraint of the desired free space
@@ -102,42 +83,37 @@ mark_free(struct i915_vma *vma, struct list_head *unwind)
  * memory in e.g. the shrinker.
  */
 int
-i915_gem_evict_something(struct drm_device *dev, struct i915_address_space *vm,
-			 int min_size, unsigned alignment, unsigned cache_level,
-			 unsigned long start, unsigned long end,
+i915_gem_evict_something(struct i915_address_space *vm,
+			 u64 min_size, u64 alignment,
+			 unsigned cache_level,
+			 u64 start, u64 end,
 			 unsigned flags)
 {
-	struct list_head eviction_list, unwind_list;
-	struct i915_vma *vma;
-	int ret = 0;
-	int pass = 0;
+	struct drm_i915_private *dev_priv = to_i915(vm->dev);
+	struct list_head eviction_list;
+	struct list_head *phases[] = {
+		&vm->inactive_list,
+		&vm->active_list,
+		NULL,
+	}, **phase;
+	struct i915_vma *vma, *next;
+	int ret;
 
-	trace_i915_gem_evict(dev, min_size, alignment, flags);
+	trace_i915_gem_evict(vm, min_size, alignment, flags);
 
 	/*
 	 * The goal is to evict objects and amalgamate space in LRU order.
 	 * The oldest idle objects reside on the inactive list, which is in
-	 * retirement order. The next objects to retire are those on the (per
-	 * ring) active list that do not have an outstanding flush. Once the
-	 * hardware reports completion (the seqno is updated after the
-	 * batchbuffer has been finished) the clean buffer objects would
-	 * be retired to the inactive list. Any dirty objects would be added
-	 * to the tail of the flushing list. So after processing the clean
-	 * active objects we need to emit a MI_FLUSH to retire the flushing
-	 * list, hence the retirement order of the flushing list is in
-	 * advance of the dirty objects on the active lists.
+	 * retirement order. The next objects to retire are those in flight,
+	 * on the active list, again in retirement order.
 	 *
 	 * The retirement sequence is thus:
 	 *   1. Inactive objects (already retired)
-	 *   2. Clean active objects
-	 *   3. Flushing list
-	 *   4. Dirty active objects.
+	 *   2. Active objects (will stall on unbinding)
 	 *
 	 * On each list, the oldest objects lie at the HEAD with the freshest
 	 * object on the TAIL.
 	 */
-
-	INIT_LIST_HEAD(&unwind_list);
 	if (start != 0 || end != vm->total) {
 		drm_mm_init_scan_with_range(&vm->mm, min_size,
 					    alignment, cache_level,
@@ -145,96 +121,84 @@ i915_gem_evict_something(struct drm_device *dev, struct i915_address_space *vm,
 	} else
 		drm_mm_init_scan(&vm->mm, min_size, alignment, cache_level);
 
-search_again:
-	/* First see if there is a large enough contiguous idle region... */
-	list_for_each_entry(vma, &vm->inactive_list, vm_link) {
-		if (mark_free(vma, &unwind_list))
-			goto found;
-	}
-
 	if (flags & PIN_NONBLOCK)
-		goto none;
+		phases[1] = NULL;
 
-	/* Now merge in the soon-to-be-expired objects... */
-	list_for_each_entry(vma, &vm->active_list, vm_link) {
-		if (mark_free(vma, &unwind_list))
-			goto found;
-	}
+search_again:
+	INIT_LIST_HEAD(&eviction_list);
+	phase = phases;
+	do {
+		list_for_each_entry(vma, *phase, vm_link)
+			if (mark_free(vma, &eviction_list))
+				goto found;
+	} while (*++phase);
 
-none:
 	/* Nothing found, clean up and bail out! */
-	while (!list_empty(&unwind_list)) {
-		vma = list_first_entry(&unwind_list,
-				       struct i915_vma,
-				       exec_list);
+	list_for_each_entry_safe(vma, next, &eviction_list, exec_list) {
 		ret = drm_mm_scan_remove_block(&vma->node);
 		BUG_ON(ret);
 
-		list_del_init(&vma->exec_list);
+		INIT_LIST_HEAD(&vma->exec_list);
 	}
 
 	/* Can we unpin some objects such as idle hw contents,
-	 * or pending flips?
+	 * or pending flips? But since only the GGTT has global entries
+	 * such as scanouts, rinbuffers and contexts, we can skip the
+	 * purge when inspecting per-process local address spaces.
 	 */
-	if (flags & PIN_NONBLOCK)
+	if (!i915_is_ggtt(vm) || flags & PIN_NONBLOCK)
 		return -ENOSPC;
 
-	/* Only idle the GPU and repeat the search once */
-	if (pass++ == 0) {
-		struct drm_i915_private *dev_priv = to_i915(dev);
-
-		if (i915_is_ggtt(vm)) {
-			ret = switch_to_pinned_context(dev_priv);
-			if (ret)
-				return ret;
-		}
-
-		ret = i915_gem_wait_for_idle(dev_priv);
-		if (ret)
-			return ret;
-
-		i915_gem_retire_requests(dev_priv);
-		goto search_again;
+	if (gpu_is_idle(dev_priv)) {
+		/* If we still have pending pageflip completions, drop
+		 * back to userspace to give our workqueues time to
+		 * acquire our locks and unpin the old scanouts.
+		 */
+		return intel_has_pending_fb_unpin(vm->dev) ? -EAGAIN : -ENOSPC;
 	}
 
-	/* If we still have pending pageflip completions, drop
-	 * back to userspace to give our workqueues time to
-	 * acquire our locks and unpin the old scanouts.
+	/* Not everything in the GGTT is tracked via vma (otherwise we
+	 * could evict as required with minimal stalling) so we are forced
+	 * to idle the GPU and explicitly retire outstanding requests in
+	 * the hopes that we can then remove contexts and the like only
+	 * bound by their active reference.
 	 */
-	return intel_has_pending_fb_unpin(dev) ? -EAGAIN : -ENOSPC;
+	ret = i915_gem_switch_to_kernel_context(dev_priv);
+	if (ret)
+		return ret;
+
+	ret = i915_gem_wait_for_idle(dev_priv, true);
+	if (ret)
+		return ret;
+
+	i915_gem_retire_requests(dev_priv);
+	goto search_again;
 
 found:
 	/* drm_mm doesn't allow any other other operations while
-	 * scanning, therefore store to be evicted objects on a
-	 * temporary list. */
-	INIT_LIST_HEAD(&eviction_list);
-	while (!list_empty(&unwind_list)) {
-		vma = list_first_entry(&unwind_list,
-				       struct i915_vma,
-				       exec_list);
-		if (drm_mm_scan_remove_block(&vma->node)) {
-			list_move(&vma->exec_list, &eviction_list);
-			drm_gem_object_reference(&vma->obj->base);
-			continue;
-		}
-		list_del_init(&vma->exec_list);
+	 * scanning, therefore store to-be-evicted objects on a
+	 * temporary list and take a reference for all before
+	 * calling unbind (which may remove the active reference
+	 * of any of our objects, thus corrupting the list).
+	 */
+	list_for_each_entry_safe(vma, next, &eviction_list, exec_list) {
+		if (drm_mm_scan_remove_block(&vma->node))
+			__i915_vma_pin(vma);
+		else
+			list_del_init(&vma->exec_list);
 	}
 
 	/* Unbinding will emit any required flushes */
 	while (!list_empty(&eviction_list)) {
-		struct drm_gem_object *obj;
 		vma = list_first_entry(&eviction_list,
 				       struct i915_vma,
 				       exec_list);
 
-		obj =  &vma->obj->base;
 		list_del_init(&vma->exec_list);
+		__i915_vma_unpin(vma);
 		if (ret == 0)
 			ret = i915_vma_unbind(vma);
-
-		drm_gem_object_unreference(obj);
 	}
-
 	return ret;
 }
 
@@ -256,8 +220,8 @@ i915_gem_evict_for_vma(struct i915_vma *target)
 
 		vma = container_of(node, typeof(*vma), node);
 
-		if (vma->pin_count) {
-			if (!vma->exec_entry || (vma->pin_count > 1))
+		if (i915_vma_is_pinned(vma)) {
+			if (!vma->exec_entry || i915_vma_pin_count(vma) > 1)
 				/* Object is pinned for some other use */
 				return -EBUSY;
 
@@ -303,22 +267,21 @@ int i915_gem_evict_vm(struct i915_address_space *vm, bool do_idle)
 		struct drm_i915_private *dev_priv = to_i915(vm->dev);
 
 		if (i915_is_ggtt(vm)) {
-			ret = switch_to_pinned_context(dev_priv);
+			ret = i915_gem_switch_to_kernel_context(dev_priv);
 			if (ret)
 				return ret;
 		}
 
-		ret = i915_gem_wait_for_idle(dev_priv);
+		ret = i915_gem_wait_for_idle(dev_priv, true);
 		if (ret)
 			return ret;
 
 		i915_gem_retire_requests(dev_priv);
-
 		WARN_ON(!list_empty(&vm->active_list));
 	}
 
 	list_for_each_entry_safe(vma, next, &vm->inactive_list, vm_link)
-		if (vma->pin_count == 0)
+		if (!i915_vma_is_pinned(vma))
 			WARN_ON(i915_vma_unbind(vma));
 
 	return 0;

drivers/gpu/drm/i915/i915_gem_fence.c

@@ -86,20 +86,22 @@ static void i965_write_fence_reg(struct drm_device *dev, int reg,
 
 	if (obj) {
 		u32 size = i915_gem_obj_ggtt_size(obj);
+		unsigned int tiling = i915_gem_object_get_tiling(obj);
+		unsigned int stride = i915_gem_object_get_stride(obj);
 		uint64_t val;
 
 		/* Adjust fence size to match tiled area */
-		if (obj->tiling_mode != I915_TILING_NONE) {
-			uint32_t row_size = obj->stride *
-				(obj->tiling_mode == I915_TILING_Y ? 32 : 8);
+		if (tiling != I915_TILING_NONE) {
+			uint32_t row_size = stride *
+				(tiling == I915_TILING_Y ? 32 : 8);
 			size = (size / row_size) * row_size;
 		}
 
 		val = (uint64_t)((i915_gem_obj_ggtt_offset(obj) + size - 4096) &
 				 0xfffff000) << 32;
 		val |= i915_gem_obj_ggtt_offset(obj) & 0xfffff000;
-		val |= (uint64_t)((obj->stride / 128) - 1) << fence_pitch_shift;
-		if (obj->tiling_mode == I915_TILING_Y)
+		val |= (uint64_t)((stride / 128) - 1) << fence_pitch_shift;
+		if (tiling == I915_TILING_Y)
 			val |= 1 << I965_FENCE_TILING_Y_SHIFT;
 		val |= I965_FENCE_REG_VALID;
 
@@ -122,6 +124,8 @@ static void i915_write_fence_reg(struct drm_device *dev, int reg,
 
 	if (obj) {
 		u32 size = i915_gem_obj_ggtt_size(obj);
+		unsigned int tiling = i915_gem_object_get_tiling(obj);
+		unsigned int stride = i915_gem_object_get_stride(obj);
 		int pitch_val;
 		int tile_width;
 
@@ -131,17 +135,17 @@ static void i915_write_fence_reg(struct drm_device *dev, int reg,
 		     "object 0x%08llx [fenceable? %d] not 1M or pot-size (0x%08x) aligned\n",
 		     i915_gem_obj_ggtt_offset(obj), obj->map_and_fenceable, size);
 
-		if (obj->tiling_mode == I915_TILING_Y && HAS_128_BYTE_Y_TILING(dev))
+		if (tiling == I915_TILING_Y && HAS_128_BYTE_Y_TILING(dev))
 			tile_width = 128;
 		else
 			tile_width = 512;
 
 		/* Note: pitch better be a power of two tile widths */
-		pitch_val = obj->stride / tile_width;
+		pitch_val = stride / tile_width;
 		pitch_val = ffs(pitch_val) - 1;
 
 		val = i915_gem_obj_ggtt_offset(obj);
-		if (obj->tiling_mode == I915_TILING_Y)
+		if (tiling == I915_TILING_Y)
 			val |= 1 << I830_FENCE_TILING_Y_SHIFT;
 		val |= I915_FENCE_SIZE_BITS(size);
 		val |= pitch_val << I830_FENCE_PITCH_SHIFT;
@@ -161,6 +165,8 @@ static void i830_write_fence_reg(struct drm_device *dev, int reg,
 
 	if (obj) {
 		u32 size = i915_gem_obj_ggtt_size(obj);
+		unsigned int tiling = i915_gem_object_get_tiling(obj);
+		unsigned int stride = i915_gem_object_get_stride(obj);
 		uint32_t pitch_val;
 
 		WARN((i915_gem_obj_ggtt_offset(obj) & ~I830_FENCE_START_MASK) ||
@@ -169,11 +175,11 @@ static void i830_write_fence_reg(struct drm_device *dev, int reg,
 		     "object 0x%08llx not 512K or pot-size 0x%08x aligned\n",
 		     i915_gem_obj_ggtt_offset(obj), size);
 
-		pitch_val = obj->stride / 128;
+		pitch_val = stride / 128;
 		pitch_val = ffs(pitch_val) - 1;
 
 		val = i915_gem_obj_ggtt_offset(obj);
-		if (obj->tiling_mode == I915_TILING_Y)
+		if (tiling == I915_TILING_Y)
 			val |= 1 << I830_FENCE_TILING_Y_SHIFT;
 		val |= I830_FENCE_SIZE_BITS(size);
 		val |= pitch_val << I830_FENCE_PITCH_SHIFT;
@@ -201,9 +207,12 @@ static void i915_gem_write_fence(struct drm_device *dev, int reg,
 	if (i915_gem_object_needs_mb(dev_priv->fence_regs[reg].obj))
 		mb();
 
-	WARN(obj && (!obj->stride || !obj->tiling_mode),
+	WARN(obj &&
+	     (!i915_gem_object_get_stride(obj) ||
+	      !i915_gem_object_get_tiling(obj)),
 	     "bogus fence setup with stride: 0x%x, tiling mode: %i\n",
-	     obj->stride, obj->tiling_mode);
+	     i915_gem_object_get_stride(obj),
+	     i915_gem_object_get_tiling(obj));
 
 	if (IS_GEN2(dev))
 		i830_write_fence_reg(dev, reg, obj);
@@ -248,7 +257,7 @@ static void i915_gem_object_update_fence(struct drm_i915_gem_object *obj,
 
 static inline void i915_gem_object_fence_lost(struct drm_i915_gem_object *obj)
 {
-	if (obj->tiling_mode)
+	if (i915_gem_object_is_tiled(obj))
 		i915_gem_release_mmap(obj);
 
 	/* As we do not have an associated fence register, we will force
@@ -261,15 +270,8 @@ static inline void i915_gem_object_fence_lost(struct drm_i915_gem_object *obj)
 static int
 i915_gem_object_wait_fence(struct drm_i915_gem_object *obj)
 {
-	if (obj->last_fenced_req) {
-		int ret = i915_wait_request(obj->last_fenced_req);
-		if (ret)
-			return ret;
-
-		i915_gem_request_assign(&obj->last_fenced_req, NULL);
-	}
-
-	return 0;
+	return i915_gem_active_retire(&obj->last_fence,
+				      &obj->base.dev->struct_mutex);
 }
 
 /**
@@ -368,7 +370,7 @@ i915_gem_object_get_fence(struct drm_i915_gem_object *obj)
 {
 	struct drm_device *dev = obj->base.dev;
 	struct drm_i915_private *dev_priv = to_i915(dev);
-	bool enable = obj->tiling_mode != I915_TILING_NONE;
+	bool enable = i915_gem_object_is_tiled(obj);
 	struct drm_i915_fence_reg *reg;
 	int ret;
 
@@ -438,7 +440,7 @@ i915_gem_object_pin_fence(struct drm_i915_gem_object *obj)
 
 		WARN_ON(!ggtt_vma ||
 			dev_priv->fence_regs[obj->fence_reg].pin_count >
-			ggtt_vma->pin_count);
+			i915_vma_pin_count(ggtt_vma));
 		dev_priv->fence_regs[obj->fence_reg].pin_count++;
 		return true;
 	} else
@@ -484,7 +486,7 @@ void i915_gem_restore_fences(struct drm_device *dev)
 		 */
 		if (reg->obj) {
 			i915_gem_object_update_fence(reg->obj, reg,
-						     reg->obj->tiling_mode);
+						     i915_gem_object_get_tiling(reg->obj));
 		} else {
 			i915_gem_write_fence(dev, i, NULL);
 		}

drivers/gpu/drm/i915/i915_gem_gtt.h

@@ -36,6 +36,8 @@
 
 #include <linux/io-mapping.h>
 
+#include "i915_gem_request.h"
+
 struct drm_i915_file_private;
 
 typedef uint32_t gen6_pte_t;
@@ -178,12 +180,32 @@ struct i915_vma {
 	struct drm_i915_gem_object *obj;
 	struct i915_address_space *vm;
 	void __iomem *iomap;
+	u64 size;
+
+	unsigned int flags;
+	/**
+	 * How many users have pinned this object in GTT space. The following
+	 * users can each hold at most one reference: pwrite/pread, execbuffer
+	 * (objects are not allowed multiple times for the same batchbuffer),
+	 * and the framebuffer code. When switching/pageflipping, the
+	 * framebuffer code has at most two buffers pinned per crtc.
+	 *
+	 * In the worst case this is 1 + 1 + 1 + 2*2 = 7. That would fit into 3
+	 * bits with absolutely no headroom. So use 4 bits.
+	 */
+#define I915_VMA_PIN_MASK 0xf
+#define I915_VMA_PIN_OVERFLOW	BIT(5)
 
 	/** Flags and address space this VMA is bound to */
-#define GLOBAL_BIND	(1<<0)
-#define LOCAL_BIND	(1<<1)
-	unsigned int bound : 4;
-	bool is_ggtt : 1;
+#define I915_VMA_GLOBAL_BIND	BIT(6)
+#define I915_VMA_LOCAL_BIND	BIT(7)
+#define I915_VMA_BIND_MASK (I915_VMA_GLOBAL_BIND | I915_VMA_LOCAL_BIND | I915_VMA_PIN_OVERFLOW)
+
+#define I915_VMA_GGTT	BIT(8)
+#define I915_VMA_CLOSED BIT(9)
+
+	unsigned int active;
+	struct i915_gem_active last_read[I915_NUM_ENGINES];
 
 	/**
 	 * Support different GGTT views into the same object.
@@ -208,20 +230,46 @@ struct i915_vma {
 	struct hlist_node exec_node;
 	unsigned long exec_handle;
 	struct drm_i915_gem_exec_object2 *exec_entry;
-
-	/**
-	 * How many users have pinned this object in GTT space. The following
-	 * users can each hold at most one reference: pwrite/pread, execbuffer
-	 * (objects are not allowed multiple times for the same batchbuffer),
-	 * and the framebuffer code. When switching/pageflipping, the
-	 * framebuffer code has at most two buffers pinned per crtc.
-	 *
-	 * In the worst case this is 1 + 1 + 1 + 2*2 = 7. That would fit into 3
-	 * bits with absolutely no headroom. So use 4 bits. */
-	unsigned int pin_count:4;
-#define DRM_I915_GEM_OBJECT_MAX_PIN_COUNT 0xf
 };
 
+static inline bool i915_vma_is_ggtt(const struct i915_vma *vma)
+{
+	return vma->flags & I915_VMA_GGTT;
+}
+
+static inline bool i915_vma_is_closed(const struct i915_vma *vma)
+{
+	return vma->flags & I915_VMA_CLOSED;
+}
+
+static inline unsigned int i915_vma_get_active(const struct i915_vma *vma)
+{
+	return vma->active;
+}
+
+static inline bool i915_vma_is_active(const struct i915_vma *vma)
+{
+	return i915_vma_get_active(vma);
+}
+
+static inline void i915_vma_set_active(struct i915_vma *vma,
+				       unsigned int engine)
+{
+	vma->active |= BIT(engine);
+}
+
+static inline void i915_vma_clear_active(struct i915_vma *vma,
+					 unsigned int engine)
+{
+	vma->active &= ~BIT(engine);
+}
+
+static inline bool i915_vma_has_active_engine(const struct i915_vma *vma,
+					      unsigned int engine)
+{
+	return vma->active & BIT(engine);
+}
+
 struct i915_page_dma {
 	struct page *page;
 	union {
@@ -272,11 +320,20 @@ struct i915_pml4 {
 struct i915_address_space {
 	struct drm_mm mm;
 	struct drm_device *dev;
+	/* Every address space belongs to a struct file - except for the global
+	 * GTT that is owned by the driver (and so @file is set to NULL). In
+	 * principle, no information should leak from one context to another
+	 * (or between files/processes etc) unless explicitly shared by the
+	 * owner. Tracking the owner is important in order to free up per-file
+	 * objects along with the file, to aide resource tracking, and to
+	 * assign blame.
+	 */
+	struct drm_i915_file_private *file;
 	struct list_head global_link;
 	u64 start;		/* Start offset always 0 for dri2 */
 	u64 total;		/* size addr space maps (ex. 2GB for ggtt) */
 
-	bool is_ggtt;
+	bool closed;
 
 	struct i915_page_scratch *scratch_page;
 	struct i915_page_table *scratch_pt;
@@ -306,6 +363,13 @@ struct i915_address_space {
 	 */
 	struct list_head inactive_list;
 
+	/**
+	 * List of vma that have been unbound.
+	 *
+	 * A reference is not held on the buffer while on this list.
+	 */
+	struct list_head unbound_list;
+
 	/* FIXME: Need a more generic return type */
 	gen6_pte_t (*pte_encode)(dma_addr_t addr,
 				 enum i915_cache_level level,
@@ -338,7 +402,7 @@ struct i915_address_space {
 			u32 flags);
 };
 
-#define i915_is_ggtt(V) ((V)->is_ggtt)
+#define i915_is_ggtt(V) (!(V)->file)
 
 /* The Graphics Translation Table is the way in which GEN hardware translates a
  * Graphics Virtual Address into a Physical Address. In addition to the normal
@@ -354,7 +418,6 @@ struct i915_ggtt {
 	size_t stolen_usable_size;	/* Total size minus BIOS reserved */
 	size_t stolen_reserved_base;
 	size_t stolen_reserved_size;
-	size_t size;			/* Total size of Global GTT */
 	u64 mappable_end;		/* End offset that we can CPU map */
 	struct io_mapping *mappable;	/* Mapping to our CPU mappable region */
 	phys_addr_t mappable_base;	/* PA of our GMADR */
@@ -365,8 +428,6 @@ struct i915_ggtt {
 	bool do_idle_maps;
 
 	int mtrr;
-
-	int (*probe)(struct i915_ggtt *ggtt);
 };
 
 struct i915_hw_ppgtt {
@@ -380,8 +441,6 @@ struct i915_hw_ppgtt {
 		struct i915_page_directory pd;		/* GEN6-7 */
 	};
 
-	struct drm_i915_file_private *file_priv;
-
 	gen6_pte_t __iomem *pd_addr;
 
 	int (*enable)(struct i915_hw_ppgtt *ppgtt);
@@ -521,14 +580,15 @@ i915_page_dir_dma_addr(const struct i915_hw_ppgtt *ppgtt, const unsigned n)
 		px_dma(ppgtt->base.scratch_pd);
 }
 
-int i915_ggtt_init_hw(struct drm_device *dev);
-int i915_ggtt_enable_hw(struct drm_device *dev);
-void i915_gem_init_ggtt(struct drm_device *dev);
-void i915_ggtt_cleanup_hw(struct drm_device *dev);
+int i915_ggtt_probe_hw(struct drm_i915_private *dev_priv);
+int i915_ggtt_init_hw(struct drm_i915_private *dev_priv);
+int i915_ggtt_enable_hw(struct drm_i915_private *dev_priv);
+int i915_gem_init_ggtt(struct drm_i915_private *dev_priv);
+void i915_ggtt_cleanup_hw(struct drm_i915_private *dev_priv);
 
 int i915_ppgtt_init_hw(struct drm_device *dev);
 void i915_ppgtt_release(struct kref *kref);
-struct i915_hw_ppgtt *i915_ppgtt_create(struct drm_device *dev,
+struct i915_hw_ppgtt *i915_ppgtt_create(struct drm_i915_private *dev_priv,
 					struct drm_i915_file_private *fpriv);
 static inline void i915_ppgtt_get(struct i915_hw_ppgtt *ppgtt)
 {
@@ -562,9 +622,66 @@ i915_ggtt_view_equal(const struct i915_ggtt_view *a,
 	return true;
 }
 
-size_t
-i915_ggtt_view_size(struct drm_i915_gem_object *obj,
-		    const struct i915_ggtt_view *view);
+/* Flags used by pin/bind&friends. */
+#define PIN_NONBLOCK		BIT(0)
+#define PIN_MAPPABLE		BIT(1)
+#define PIN_ZONE_4G		BIT(2)
+
+#define PIN_MBZ			BIT(5) /* I915_VMA_PIN_OVERFLOW */
+#define PIN_GLOBAL		BIT(6) /* I915_VMA_GLOBAL_BIND */
+#define PIN_USER		BIT(7) /* I915_VMA_LOCAL_BIND */
+#define PIN_UPDATE		BIT(8)
+
+#define PIN_HIGH		BIT(9)
+#define PIN_OFFSET_BIAS		BIT(10)
+#define PIN_OFFSET_FIXED	BIT(11)
+#define PIN_OFFSET_MASK		(~4095)
+
+int __i915_vma_do_pin(struct i915_vma *vma,
+		      u64 size, u64 alignment, u64 flags);
+static inline int __must_check
+i915_vma_pin(struct i915_vma *vma, u64 size, u64 alignment, u64 flags)
+{
+	BUILD_BUG_ON(PIN_MBZ != I915_VMA_PIN_OVERFLOW);
+	BUILD_BUG_ON(PIN_GLOBAL != I915_VMA_GLOBAL_BIND);
+	BUILD_BUG_ON(PIN_USER != I915_VMA_LOCAL_BIND);
+
+	/* Pin early to prevent the shrinker/eviction logic from destroying
+	 * our vma as we insert and bind.
+	 */
+	if (likely(((++vma->flags ^ flags) & I915_VMA_BIND_MASK) == 0))
+		return 0;
+
+	return __i915_vma_do_pin(vma, size, alignment, flags);
+}
+
+static inline int i915_vma_pin_count(const struct i915_vma *vma)
+{
+	return vma->flags & I915_VMA_PIN_MASK;
+}
+
+static inline bool i915_vma_is_pinned(const struct i915_vma *vma)
+{
+	return i915_vma_pin_count(vma);
+}
+
+static inline void __i915_vma_pin(struct i915_vma *vma)
+{
+	vma->flags++;
+	GEM_BUG_ON(vma->flags & I915_VMA_PIN_OVERFLOW);
+}
+
+static inline void __i915_vma_unpin(struct i915_vma *vma)
+{
+	GEM_BUG_ON(!i915_vma_is_pinned(vma));
+	vma->flags--;
+}
+
+static inline void i915_vma_unpin(struct i915_vma *vma)
+{
+	GEM_BUG_ON(!drm_mm_node_allocated(&vma->node));
+	__i915_vma_unpin(vma);
+}
 
 /**
  * i915_vma_pin_iomap - calls ioremap_wc to map the GGTT VMA via the aperture
@@ -580,6 +697,7 @@ i915_ggtt_view_size(struct drm_i915_gem_object *obj,
  * Returns a valid iomapped pointer or ERR_PTR.
  */
 void __iomem *i915_vma_pin_iomap(struct i915_vma *vma);
+#define IO_ERR_PTR(x) ((void __iomem *)ERR_PTR(x))
 
 /**
  * i915_vma_unpin_iomap - unpins the mapping returned from i915_vma_iomap
@@ -593,9 +711,8 @@ void __iomem *i915_vma_pin_iomap(struct i915_vma *vma);
 static inline void i915_vma_unpin_iomap(struct i915_vma *vma)
 {
 	lockdep_assert_held(&vma->vm->dev->struct_mutex);
-	GEM_BUG_ON(vma->pin_count == 0);
 	GEM_BUG_ON(vma->iomap == NULL);
-	vma->pin_count--;
+	i915_vma_unpin(vma);
 }
 
 #endif

drivers/gpu/drm/i915/i915_gem_render_state.h

@@ -26,24 +26,6 @@
 
 #include <linux/types.h>
 
-struct intel_renderstate_rodata {
-	const u32 *reloc;
-	const u32 *batch;
-	const u32 batch_items;
-};
-
-struct render_state {
-	const struct intel_renderstate_rodata *rodata;
-	struct drm_i915_gem_object *obj;
-	u64 ggtt_offset;
-	int gen;
-	u32 aux_batch_size;
-	u32 aux_batch_offset;
-};
-
 int i915_gem_render_state_init(struct drm_i915_gem_request *req);
-void i915_gem_render_state_fini(struct render_state *so);
-int i915_gem_render_state_prepare(struct intel_engine_cs *engine,
-				  struct render_state *so);
 
 #endif /* _I915_GEM_RENDER_STATE_H_ */

drivers/gpu/drm/i915/i915_gem_stolen.c

@@ -698,24 +698,24 @@ i915_gem_object_create_stolen_for_preallocated(struct drm_device *dev,
 	 */
 	vma->node.start = gtt_offset;
 	vma->node.size = size;
-	if (drm_mm_initialized(&ggtt->base.mm)) {
-		ret = drm_mm_reserve_node(&ggtt->base.mm, &vma->node);
-		if (ret) {
-			DRM_DEBUG_KMS("failed to allocate stolen GTT space\n");
-			goto err;
-		}
 
-		vma->bound |= GLOBAL_BIND;
-		__i915_vma_set_map_and_fenceable(vma);
-		list_add_tail(&vma->vm_link, &ggtt->base.inactive_list);
+	ret = drm_mm_reserve_node(&ggtt->base.mm, &vma->node);
+	if (ret) {
+		DRM_DEBUG_KMS("failed to allocate stolen GTT space\n");
+		goto err;
 	}
 
+	vma->flags |= I915_VMA_GLOBAL_BIND;
+	__i915_vma_set_map_and_fenceable(vma);
+	list_move_tail(&vma->vm_link, &ggtt->base.inactive_list);
+	obj->bind_count++;
+
 	list_add_tail(&obj->global_list, &dev_priv->mm.bound_list);
 	i915_gem_object_pin_pages(obj);
 
 	return obj;
 
 err:
-	drm_gem_object_unreference(&obj->base);
+	i915_gem_object_put(obj);
 	return NULL;
 }

drivers/gpu/drm/i915/i915_pci.c

@@ -173,6 +173,7 @@ static const struct intel_device_info intel_pineview_info = {
 	.gen = 3, .is_g33 = 1, .is_pineview = 1, .is_mobile = 1, .num_pipes = 2,
 	.need_gfx_hws = 1, .has_hotplug = 1,
 	.has_overlay = 1,
+	.ring_mask = RENDER_RING,
 	GEN_DEFAULT_PIPEOFFSETS,
 	CURSOR_OFFSETS,
 };

drivers/gpu/drm/i915/i915_reg.h

@@ -186,13 +186,13 @@ static inline bool i915_mmio_reg_valid(i915_reg_t reg)
 #define  GEN9_GRDOM_GUC			(1 << 5)
 #define  GEN8_GRDOM_MEDIA2		(1 << 7)
 
-#define RING_PP_DIR_BASE(ring)		_MMIO((ring)->mmio_base+0x228)
-#define RING_PP_DIR_BASE_READ(ring)	_MMIO((ring)->mmio_base+0x518)
-#define RING_PP_DIR_DCLV(ring)		_MMIO((ring)->mmio_base+0x220)
+#define RING_PP_DIR_BASE(engine)	_MMIO((engine)->mmio_base+0x228)
+#define RING_PP_DIR_BASE_READ(engine)	_MMIO((engine)->mmio_base+0x518)
+#define RING_PP_DIR_DCLV(engine)	_MMIO((engine)->mmio_base+0x220)
 #define   PP_DIR_DCLV_2G		0xffffffff
 
-#define GEN8_RING_PDP_UDW(ring, n)	_MMIO((ring)->mmio_base+0x270 + (n) * 8 + 4)
-#define GEN8_RING_PDP_LDW(ring, n)	_MMIO((ring)->mmio_base+0x270 + (n) * 8)
+#define GEN8_RING_PDP_UDW(engine, n)	_MMIO((engine)->mmio_base+0x270 + (n) * 8 + 4)
+#define GEN8_RING_PDP_LDW(engine, n)	_MMIO((engine)->mmio_base+0x270 + (n) * 8)
 
 #define GEN8_R_PWR_CLK_STATE		_MMIO(0x20C8)
 #define   GEN8_RPCS_ENABLE		(1 << 31)
@@ -1536,6 +1536,7 @@ enum skl_disp_power_wells {
 #define BALANCE_LEG_MASK(port)		(7<<(8+3*(port)))
 /* Balance leg disable bits */
 #define BALANCE_LEG_DISABLE_SHIFT	23
+#define BALANCE_LEG_DISABLE(port)	(1 << (23 + (port)))
 
 /*
  * Fence registers
@@ -1647,7 +1648,7 @@ enum skl_disp_power_wells {
 #define   ARB_MODE_BWGTLB_DISABLE (1<<9)
 #define   ARB_MODE_SWIZZLE_BDW	(1<<1)
 #define RENDER_HWS_PGA_GEN7	_MMIO(0x04080)
-#define RING_FAULT_REG(ring)	_MMIO(0x4094 + 0x100*(ring)->id)
+#define RING_FAULT_REG(engine)	_MMIO(0x4094 + 0x100*(engine)->hw_id)
 #define   RING_FAULT_GTTSEL_MASK (1<<11)
 #define   RING_FAULT_SRCID(x)	(((x) >> 3) & 0xff)
 #define   RING_FAULT_FAULT_TYPE(x) (((x) >> 1) & 0x3)
@@ -1845,7 +1846,7 @@ enum skl_disp_power_wells {
 
 #define GFX_MODE	_MMIO(0x2520)
 #define GFX_MODE_GEN7	_MMIO(0x229c)
-#define RING_MODE_GEN7(ring)	_MMIO((ring)->mmio_base+0x29c)
+#define RING_MODE_GEN7(engine)	_MMIO((engine)->mmio_base+0x29c)
 #define   GFX_RUN_LIST_ENABLE		(1<<15)
 #define   GFX_INTERRUPT_STEERING	(1<<14)
 #define   GFX_TLB_INVALIDATE_EXPLICIT	(1<<13)
@@ -6132,6 +6133,7 @@ enum {
 # define GEN7_CSC1_RHWO_OPT_DISABLE_IN_RCC	((1<<10) | (1<<26))
 # define GEN9_RHWO_OPTIMIZATION_DISABLE		(1<<14)
 #define COMMON_SLICE_CHICKEN2			_MMIO(0x7014)
+# define GEN9_DISABLE_GATHER_AT_SET_SHADER_COMMON_SLICE (1<<12)
 # define GEN8_SBE_DISABLE_REPLAY_BUF_OPTIMIZATION (1<<8)
 # define GEN8_CSC2_SBE_VUE_CACHE_CONSERVATIVE	(1<<0)
 
@@ -6958,6 +6960,9 @@ enum {
 #define  ECOBUS					_MMIO(0xa180)
 #define    FORCEWAKE_MT_ENABLE			(1<<5)
 #define  VLV_SPAREG2H				_MMIO(0xA194)
+#define  GEN9_PWRGT_DOMAIN_STATUS		_MMIO(0xA2A0)
+#define   GEN9_PWRGT_MEDIA_STATUS_MASK		(1 << 0)
+#define   GEN9_PWRGT_RENDER_STATUS_MASK		(1 << 1)
 
 #define  GTFIFODBG				_MMIO(0x120000)
 #define    GT_FIFO_SBDEDICATE_FREE_ENTRY_CHV	(0x1f << 20)
@@ -7485,6 +7490,7 @@ enum {
 #define _DDI_BUF_TRANS_A		0x64E00
 #define _DDI_BUF_TRANS_B		0x64E60
 #define DDI_BUF_TRANS_LO(port, i)	_MMIO(_PORT(port, _DDI_BUF_TRANS_A, _DDI_BUF_TRANS_B) + (i) * 8)
+#define  DDI_BUF_BALANCE_LEG_ENABLE	(1 << 31)
 #define DDI_BUF_TRANS_HI(port, i)	_MMIO(_PORT(port, _DDI_BUF_TRANS_A, _DDI_BUF_TRANS_B) + (i) * 8 + 4)
 
 /* Sideband Interface (SBI) is programmed indirectly, via