Merge branch 'drm-next-4.3' of git://people.freedesktop.org/~agd5f/linux into drm-next

- DP fixes for radeon and amdgpu
- IH ring fix for tonga and fiji
- Lots of GPU scheduler fixes
- Misc additional fixes

* 'drm-next-4.3' of git://people.freedesktop.org/~agd5f/linux: (42 commits)
  drm/amdgpu: fix wait queue handling in the scheduler
  drm/amdgpu: remove extra parameters from scheduler callbacks
  drm/amdgpu: wake up scheduler only when neccessary
  drm/amdgpu: remove entity idle timeout v2
  drm/amdgpu: fix postclose order
  drm/amdgpu: use IB for copy buffer of eviction
  drm/amdgpu: adjust the judgement of removing fence callback
  drm/amdgpu: fix no sync_wait in copy_buffer
  drm/amdgpu: fix last_vm_update fence is not effetive for sched fence
  drm/amdgpu: add priv data to sched
  drm/amdgpu: add owner for sched fence
  drm/amdgpu: remove entity reference from sched fence
  drm/amdgpu: fix and cleanup amd_sched_entity_push_job
  drm/amdgpu: remove amdgpu_bo_list_clone
  drm/amdgpu: remove the context from amdgpu_job
  drm/amdgpu: remove unused parameters to amd_sched_create
  drm/amdgpu: remove sched_lock
  drm/amdgpu: remove prepare_job callback
  drm/amdgpu: cleanup a scheduler function name
  drm/amdgpu: reorder scheduler functions
  ...

30 files changed, 671 insertions, 727 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 2fc58e658986..aa2dcf578dd6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -183,6 +183,7 @@ struct amdgpu_vm;
 struct amdgpu_ring;
 struct amdgpu_semaphore;
 struct amdgpu_cs_parser;
+struct amdgpu_job;
 struct amdgpu_irq_src;
 struct amdgpu_fpriv;
 
@@ -246,7 +247,7 @@ struct amdgpu_buffer_funcs {
 	unsigned	copy_num_dw;
 
 	/* used for buffer migration */
-	void (*emit_copy_buffer)(struct amdgpu_ring *ring,
+	void (*emit_copy_buffer)(struct amdgpu_ib *ib,
 				 /* src addr in bytes */
 				 uint64_t src_offset,
 				 /* dst addr in bytes */
@@ -439,9 +440,12 @@ int amdgpu_fence_wait_next(struct amdgpu_ring *ring);
 int amdgpu_fence_wait_empty(struct amdgpu_ring *ring);
 unsigned amdgpu_fence_count_emitted(struct amdgpu_ring *ring);
 
-signed long amdgpu_fence_wait_any(struct amdgpu_device *adev,
-			  struct amdgpu_fence **fences,
-			  bool intr, long t);
+signed long amdgpu_fence_wait_multiple(struct amdgpu_device *adev,
+				       struct fence **array,
+				       uint32_t count,
+				       bool wait_all,
+				       bool intr,
+				       signed long t);
 struct amdgpu_fence *amdgpu_fence_ref(struct amdgpu_fence *fence);
 void amdgpu_fence_unref(struct amdgpu_fence **fence);
 
@@ -514,7 +518,7 @@ int amdgpu_copy_buffer(struct amdgpu_ring *ring,
 		       uint64_t dst_offset,
 		       uint32_t byte_count,
 		       struct reservation_object *resv,
-		       struct amdgpu_fence **fence);
+		       struct fence **fence);
 int amdgpu_mmap(struct file *filp, struct vm_area_struct *vma);
 
 struct amdgpu_bo_list_entry {
@@ -650,7 +654,7 @@ struct amdgpu_sa_bo {
 	struct amdgpu_sa_manager	*manager;
 	unsigned			soffset;
 	unsigned			eoffset;
-	struct amdgpu_fence		*fence;
+	struct fence		        *fence;
 };
 
 /*
@@ -692,7 +696,7 @@ bool amdgpu_semaphore_emit_wait(struct amdgpu_ring *ring,
 				struct amdgpu_semaphore *semaphore);
 void amdgpu_semaphore_free(struct amdgpu_device *adev,
 			   struct amdgpu_semaphore **semaphore,
-			   struct amdgpu_fence *fence);
+			   struct fence *fence);
 
 /*
  * Synchronization
@@ -700,7 +704,8 @@ void amdgpu_semaphore_free(struct amdgpu_device *adev,
 struct amdgpu_sync {
 	struct amdgpu_semaphore *semaphores[AMDGPU_NUM_SYNCS];
 	struct amdgpu_fence	*sync_to[AMDGPU_MAX_RINGS];
-	struct amdgpu_fence	*last_vm_update;
+	DECLARE_HASHTABLE(fences, 4);
+	struct fence	        *last_vm_update;
 };
 
 void amdgpu_sync_create(struct amdgpu_sync *sync);
@@ -712,8 +717,9 @@ int amdgpu_sync_resv(struct amdgpu_device *adev,
 		     void *owner);
 int amdgpu_sync_rings(struct amdgpu_sync *sync,
 		      struct amdgpu_ring *ring);
+int amdgpu_sync_wait(struct amdgpu_sync *sync);
 void amdgpu_sync_free(struct amdgpu_device *adev, struct amdgpu_sync *sync,
-		      struct amdgpu_fence *fence);
+		      struct fence *fence);
 
 /*
  * GART structures, functions & helpers
@@ -871,7 +877,7 @@ int amdgpu_sched_ib_submit_kernel_helper(struct amdgpu_device *adev,
 					 struct amdgpu_ring *ring,
 					 struct amdgpu_ib *ibs,
 					 unsigned num_ibs,
-					 int (*free_job)(struct amdgpu_cs_parser *),
+					 int (*free_job)(struct amdgpu_job *),
 					 void *owner,
 					 struct fence **fence);
 
@@ -957,7 +963,7 @@ struct amdgpu_vm_id {
 	unsigned		id;
 	uint64_t		pd_gpu_addr;
 	/* last flushed PD/PT update */
-	struct amdgpu_fence	*flushed_updates;
+	struct fence	        *flushed_updates;
 	/* last use of vmid */
 	struct amdgpu_fence	*last_id_use;
 };
@@ -1042,7 +1048,7 @@ struct amdgpu_ctx *amdgpu_ctx_get(struct amdgpu_fpriv *fpriv, uint32_t id);
 int amdgpu_ctx_put(struct amdgpu_ctx *ctx);
 
 uint64_t amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx, struct amdgpu_ring *ring,
-			      struct fence *fence, uint64_t queued_seq);
+			      struct fence *fence);
 struct fence *amdgpu_ctx_get_fence(struct amdgpu_ctx *ctx,
 				   struct amdgpu_ring *ring, uint64_t seq);
 
@@ -1078,8 +1084,6 @@ struct amdgpu_bo_list {
 };
 
 struct amdgpu_bo_list *
-amdgpu_bo_list_clone(struct amdgpu_bo_list *list);
-struct amdgpu_bo_list *
 amdgpu_bo_list_get(struct amdgpu_fpriv *fpriv, int id);
 void amdgpu_bo_list_put(struct amdgpu_bo_list *list);
 void amdgpu_bo_list_free(struct amdgpu_bo_list *list);
@@ -1255,14 +1259,16 @@ struct amdgpu_cs_parser {
 
 	/* user fence */
 	struct amdgpu_user_fence uf;
+};
 
-	struct amdgpu_ring *ring;
-	struct mutex job_lock;
-	struct work_struct job_work;
-	int (*prepare_job)(struct amdgpu_cs_parser *sched_job);
-	int (*run_job)(struct amdgpu_cs_parser *sched_job);
-	int (*free_job)(struct amdgpu_cs_parser *sched_job);
-	struct amd_sched_fence *s_fence;
+struct amdgpu_job {
+	struct amd_sched_job    base;
+	struct amdgpu_device	*adev;
+	struct amdgpu_ib	*ibs;
+	uint32_t		num_ibs;
+	struct mutex            job_lock;
+	struct amdgpu_user_fence uf;
+	int (*free_job)(struct amdgpu_job *sched_job);
 };
 
 static inline u32 amdgpu_get_ib_value(struct amdgpu_cs_parser *p, uint32_t ib_idx, int idx)
@@ -2241,7 +2247,7 @@ static inline void amdgpu_ring_write(struct amdgpu_ring *ring, uint32_t v)
 #define amdgpu_display_add_connector(adev, ci, sd, ct, ib, coi, h, r) (adev)->mode_info.funcs->add_connector((adev), (ci), (sd), (ct), (ib), (coi), (h), (r))
 #define amdgpu_display_stop_mc_access(adev, s) (adev)->mode_info.funcs->stop_mc_access((adev), (s))
 #define amdgpu_display_resume_mc_access(adev, s) (adev)->mode_info.funcs->resume_mc_access((adev), (s))
-#define amdgpu_emit_copy_buffer(adev, r, s, d, b) (adev)->mman.buffer_funcs->emit_copy_buffer((r), (s), (d), (b))
+#define amdgpu_emit_copy_buffer(adev, ib, s, d, b) (adev)->mman.buffer_funcs->emit_copy_buffer((ib),  (s), (d), (b))
 #define amdgpu_emit_fill_buffer(adev, r, s, d, b) (adev)->mman.buffer_funcs->emit_fill_buffer((r), (s), (d), (b))
 #define amdgpu_dpm_get_temperature(adev) (adev)->pm.funcs->get_temperature((adev))
 #define amdgpu_dpm_pre_set_power_state(adev) (adev)->pm.funcs->pre_set_power_state((adev))
@@ -2343,7 +2349,7 @@ int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring,
 		      struct amdgpu_sync *sync);
 void amdgpu_vm_flush(struct amdgpu_ring *ring,
 		     struct amdgpu_vm *vm,
-		     struct amdgpu_fence *updates);
+		     struct fence *updates);
 void amdgpu_vm_fence(struct amdgpu_device *adev,
 		     struct amdgpu_vm *vm,
 		     struct amdgpu_fence *fence);
@@ -2373,7 +2379,7 @@ int amdgpu_vm_bo_unmap(struct amdgpu_device *adev,
 		       uint64_t addr);
 void amdgpu_vm_bo_rmv(struct amdgpu_device *adev,
 		      struct amdgpu_bo_va *bo_va);
-
+int amdgpu_vm_free_job(struct amdgpu_job *job);
 /*
  * functions used by amdgpu_encoder.c
  */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_benchmark.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_benchmark.c
index 759482e4300d..98d59ee640ce 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_benchmark.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_benchmark.c
@@ -33,7 +33,7 @@ static int amdgpu_benchmark_do_move(struct amdgpu_device *adev, unsigned size,
 {
 	unsigned long start_jiffies;
 	unsigned long end_jiffies;
-	struct amdgpu_fence *fence = NULL;
+	struct fence *fence = NULL;
 	int i, r;
 
 	start_jiffies = jiffies;
@@ -42,17 +42,17 @@ static int amdgpu_benchmark_do_move(struct amdgpu_device *adev, unsigned size,
 		r = amdgpu_copy_buffer(ring, saddr, daddr, size, NULL, &fence);
 		if (r)
 			goto exit_do_move;
-		r = fence_wait(&fence->base, false);
+		r = fence_wait(fence, false);
 		if (r)
 			goto exit_do_move;
-		amdgpu_fence_unref(&fence);
+		fence_put(fence);
 	}
 	end_jiffies = jiffies;
 	r = jiffies_to_msecs(end_jiffies - start_jiffies);
 
 exit_do_move:
 	if (fence)
-		amdgpu_fence_unref(&fence);
+		fence_put(fence);
 	return r;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c
index 7eed523bf28f..f82a2dd83874 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c
@@ -62,39 +62,6 @@ static int amdgpu_bo_list_create(struct amdgpu_fpriv *fpriv,
 	return 0;
 }
 
-struct amdgpu_bo_list *
-amdgpu_bo_list_clone(struct amdgpu_bo_list *list)
-{
-	struct amdgpu_bo_list *result;
-	unsigned i;
-
-	result = kmalloc(sizeof(struct amdgpu_bo_list), GFP_KERNEL);
-	if (!result)
-		return NULL;
-
-	result->array = drm_calloc_large(list->num_entries,
-		sizeof(struct amdgpu_bo_list_entry));
-	if (!result->array) {
-		kfree(result);
-		return NULL;
-	}
-
-	mutex_init(&result->lock);
-	result->gds_obj = list->gds_obj;
-	result->gws_obj = list->gws_obj;
-	result->oa_obj = list->oa_obj;
-	result->has_userptr = list->has_userptr;
-	result->num_entries = list->num_entries;
-
-	memcpy(result->array, list->array, list->num_entries *
-	       sizeof(struct amdgpu_bo_list_entry));
-
-	for (i = 0; i < result->num_entries; ++i)
-		amdgpu_bo_ref(result->array[i].robj);
-
-	return result;
-}
-
 static void amdgpu_bo_list_destroy(struct amdgpu_fpriv *fpriv, int id)
 {
 	struct amdgpu_bo_list *list;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c
index 27df17a0e620..89c3dd62ba21 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c
@@ -75,6 +75,11 @@ void amdgpu_connector_hotplug(struct drm_connector *connector)
 			if (!amdgpu_display_hpd_sense(adev, amdgpu_connector->hpd.hpd)) {
 				drm_helper_connector_dpms(connector, DRM_MODE_DPMS_OFF);
 			} else if (amdgpu_atombios_dp_needs_link_train(amdgpu_connector)) {
+				/* Don't try to start link training before we
+				 * have the dpcd */
+				if (!amdgpu_atombios_dp_get_dpcd(amdgpu_connector))
+					return;
+
 				/* set it to OFF so that drm_helper_connector_dpms()
 				 * won't return immediately since the current state
 				 * is ON at this point.
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index e4424b4db5d3..6a206f15635f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -126,19 +126,6 @@ int amdgpu_cs_get_ring(struct amdgpu_device *adev, u32 ip_type,
 	return 0;
 }
 
-static void amdgpu_job_work_func(struct work_struct *work)
-{
-	struct amdgpu_cs_parser *sched_job =
-		container_of(work, struct amdgpu_cs_parser,
-			     job_work);
-	mutex_lock(&sched_job->job_lock);
-	if (sched_job->free_job)
-		sched_job->free_job(sched_job);
-	mutex_unlock(&sched_job->job_lock);
-	/* after processing job, free memory */
-	fence_put(&sched_job->s_fence->base);
-	kfree(sched_job);
-}
 struct amdgpu_cs_parser *amdgpu_cs_parser_create(struct amdgpu_device *adev,
                                                struct drm_file *filp,
                                                struct amdgpu_ctx *ctx,
@@ -157,10 +144,6 @@ struct amdgpu_cs_parser *amdgpu_cs_parser_create(struct amdgpu_device *adev,
 	parser->ctx = ctx;
 	parser->ibs = ibs;
 	parser->num_ibs = num_ibs;
-	if (amdgpu_enable_scheduler) {
-		mutex_init(&parser->job_lock);
-		INIT_WORK(&parser->job_work, amdgpu_job_work_func);
-	}
 	for (i = 0; i < num_ibs; i++)
 		ibs[i].ctx = ctx;
 
@@ -173,7 +156,6 @@ int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data)
 	uint64_t *chunk_array_user;
 	uint64_t *chunk_array = NULL;
 	struct amdgpu_fpriv *fpriv = p->filp->driver_priv;
-	struct amdgpu_bo_list *bo_list = NULL;
 	unsigned size, i;
 	int r = 0;
 
@@ -185,20 +167,7 @@ int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data)
 		r = -EINVAL;
 		goto out;
 	}
-	bo_list = amdgpu_bo_list_get(fpriv, cs->in.bo_list_handle);
-	if (!amdgpu_enable_scheduler)
-		p->bo_list = bo_list;
-	else {
-		if (bo_list && !bo_list->has_userptr) {
-			p->bo_list = amdgpu_bo_list_clone(bo_list);
-			amdgpu_bo_list_put(bo_list);
-			if (!p->bo_list)
-				return -ENOMEM;
-		} else if (bo_list && bo_list->has_userptr)
-			p->bo_list = bo_list;
-		else
-			p->bo_list = NULL;
-	}
+	p->bo_list = amdgpu_bo_list_get(fpriv, cs->in.bo_list_handle);
 
 	/* get chunks */
 	INIT_LIST_HEAD(&p->validated);
@@ -291,7 +260,7 @@ int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data)
 	}
 
 
-	p->ibs = kmalloc_array(p->num_ibs, sizeof(struct amdgpu_ib), GFP_KERNEL);
+	p->ibs = kcalloc(p->num_ibs, sizeof(struct amdgpu_ib), GFP_KERNEL);
 	if (!p->ibs)
 		r = -ENOMEM;
 
@@ -498,25 +467,24 @@ static void amdgpu_cs_parser_fini_late(struct amdgpu_cs_parser *parser)
 	unsigned i;
 	if (parser->ctx)
 		amdgpu_ctx_put(parser->ctx);
-	if (parser->bo_list) {
-		if (amdgpu_enable_scheduler && !parser->bo_list->has_userptr)
-			amdgpu_bo_list_free(parser->bo_list);
-		else
-			amdgpu_bo_list_put(parser->bo_list);
-	}
+	if (parser->bo_list)
+		amdgpu_bo_list_put(parser->bo_list);
+
 	drm_free_large(parser->vm_bos);
 	for (i = 0; i < parser->nchunks; i++)
 		drm_free_large(parser->chunks[i].kdata);
 	kfree(parser->chunks);
-	if (parser->ibs)
-		for (i = 0; i < parser->num_ibs; i++)
-			amdgpu_ib_free(parser->adev, &parser->ibs[i]);
-	kfree(parser->ibs);
-	if (parser->uf.bo)
-		drm_gem_object_unreference_unlocked(&parser->uf.bo->gem_base);
-
 	if (!amdgpu_enable_scheduler)
-		kfree(parser);
+	{
+		if (parser->ibs)
+			for (i = 0; i < parser->num_ibs; i++)
+				amdgpu_ib_free(parser->adev, &parser->ibs[i]);
+		kfree(parser->ibs);
+		if (parser->uf.bo)
+			drm_gem_object_unreference_unlocked(&parser->uf.bo->gem_base);
+	}
+
+	kfree(parser);
 }
 
 /**
@@ -533,12 +501,6 @@ static void amdgpu_cs_parser_fini(struct amdgpu_cs_parser *parser, int error, bo
        amdgpu_cs_parser_fini_late(parser);
 }
 
-static int amdgpu_cs_parser_free_job(struct amdgpu_cs_parser *sched_job)
-{
-       amdgpu_cs_parser_fini_late(sched_job);
-       return 0;
-}
-
 static int amdgpu_bo_vm_update_pte(struct amdgpu_cs_parser *p,
 				   struct amdgpu_vm *vm)
 {
@@ -810,68 +772,16 @@ static int amdgpu_cs_dependencies(struct amdgpu_device *adev,
 	return 0;
 }
 
-static int amdgpu_cs_parser_prepare_job(struct amdgpu_cs_parser *sched_job)
+static int amdgpu_cs_free_job(struct amdgpu_job *sched_job)
 {
-	int r, i;
-	struct amdgpu_cs_parser *parser = sched_job;
-	struct amdgpu_device *adev = sched_job->adev;
-	bool reserved_buffers = false;
-
-	r = amdgpu_cs_parser_relocs(parser);
-	if (r) {
-		if (r != -ERESTARTSYS) {
-			if (r == -ENOMEM)
-				DRM_ERROR("Not enough memory for command submission!\n");
-			else
-				DRM_ERROR("Failed to process the buffer list %d!\n", r);
-		}
-	}
-
-	if (!r) {
-		reserved_buffers = true;
-		r = amdgpu_cs_ib_fill(adev, parser);
-	}
-	if (!r) {
-		r = amdgpu_cs_dependencies(adev, parser);
-		if (r)
-			DRM_ERROR("Failed in the dependencies handling %d!\n", r);
-	}
-	if (r) {
-		amdgpu_cs_parser_fini(parser, r, reserved_buffers);
-		return r;
-	}
-
-	for (i = 0; i < parser->num_ibs; i++)
-		trace_amdgpu_cs(parser, i);
-
-	r = amdgpu_cs_ib_vm_chunk(adev, parser);
-	return r;
-}
-
-static struct amdgpu_ring *amdgpu_cs_parser_get_ring(
-	struct amdgpu_device *adev,
-	struct amdgpu_cs_parser *parser)
-{
-	int i, r;
-
-	struct amdgpu_cs_chunk *chunk;
-	struct drm_amdgpu_cs_chunk_ib *chunk_ib;
-	struct amdgpu_ring *ring;
-	for (i = 0; i < parser->nchunks; i++) {
-		chunk = &parser->chunks[i];
-		chunk_ib = (struct drm_amdgpu_cs_chunk_ib *)chunk->kdata;
-
-		if (chunk->chunk_id != AMDGPU_CHUNK_ID_IB)
-			continue;
-
-		r = amdgpu_cs_get_ring(adev, chunk_ib->ip_type,
-				       chunk_ib->ip_instance, chunk_ib->ring,
-				       &ring);
-		if (r)
-			return NULL;
-		break;
-	}
-	return ring;
+	int i;
+	if (sched_job->ibs)
+		for (i = 0; i < sched_job->num_ibs; i++)
+			amdgpu_ib_free(sched_job->adev, &sched_job->ibs[i]);
+	kfree(sched_job->ibs);
+	if (sched_job->uf.bo)
+		drm_gem_object_unreference_unlocked(&sched_job->uf.bo->gem_base);
+	return 0;
 }
 
 int amdgpu_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
@@ -879,7 +789,8 @@ int amdgpu_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	struct amdgpu_device *adev = dev->dev_private;
 	union drm_amdgpu_cs *cs = data;
 	struct amdgpu_cs_parser *parser;
-	int r;
+	bool reserved_buffers = false;
+	int i, r;
 
 	down_read(&adev->exclusive_lock);
 	if (!adev->accel_working) {
@@ -899,44 +810,79 @@ int amdgpu_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		return r;
 	}
 
-	if (amdgpu_enable_scheduler && parser->num_ibs) {
-		struct amdgpu_ring * ring =
-			amdgpu_cs_parser_get_ring(adev, parser);
-		r = amdgpu_cs_parser_prepare_job(parser);
+	r = amdgpu_cs_parser_relocs(parser);
+	if (r == -ENOMEM)
+		DRM_ERROR("Not enough memory for command submission!\n");
+	else if (r && r != -ERESTARTSYS)
+		DRM_ERROR("Failed to process the buffer list %d!\n", r);
+	else if (!r) {
+		reserved_buffers = true;
+		r = amdgpu_cs_ib_fill(adev, parser);
+	}
+
+	if (!r) {
+		r = amdgpu_cs_dependencies(adev, parser);
 		if (r)
-			goto out;
-		parser->ring = ring;
-		parser->free_job = amdgpu_cs_parser_free_job;
-		mutex_lock(&parser->job_lock);
-		r = amd_sched_push_job(ring->scheduler,
-				       &parser->ctx->rings[ring->idx].entity,
-				       parser,
-				       &parser->s_fence);
+			DRM_ERROR("Failed in the dependencies handling %d!\n", r);
+	}
+
+	if (r)
+		goto out;
+
+	for (i = 0; i < parser->num_ibs; i++)
+		trace_amdgpu_cs(parser, i);
+
+	r = amdgpu_cs_ib_vm_chunk(adev, parser);
+	if (r)
+		goto out;
+
+	if (amdgpu_enable_scheduler && parser->num_ibs) {
+		struct amdgpu_job *job;
+		struct amdgpu_ring * ring =  parser->ibs->ring;
+		job = kzalloc(sizeof(struct amdgpu_job), GFP_KERNEL);
+		if (!job)
+			return -ENOMEM;
+		job->base.sched = ring->scheduler;
+		job->base.s_entity = &parser->ctx->rings[ring->idx].entity;
+		job->adev = parser->adev;
+		job->ibs = parser->ibs;
+		job->num_ibs = parser->num_ibs;
+		job->base.owner = parser->filp;
+		mutex_init(&job->job_lock);
+		if (job->ibs[job->num_ibs - 1].user) {
+			memcpy(&job->uf,  &parser->uf,
+			       sizeof(struct amdgpu_user_fence));
+			job->ibs[job->num_ibs - 1].user = &job->uf;
+		}
+
+		job->free_job = amdgpu_cs_free_job;
+		mutex_lock(&job->job_lock);
+		r = amd_sched_entity_push_job((struct amd_sched_job *)job);
 		if (r) {
-			mutex_unlock(&parser->job_lock);
+			mutex_unlock(&job->job_lock);
+			amdgpu_cs_free_job(job);
+			kfree(job);
 			goto out;
 		}
-		parser->ibs[parser->num_ibs - 1].sequence =
+		cs->out.handle =
 			amdgpu_ctx_add_fence(parser->ctx, ring,
-					     &parser->s_fence->base,
-					     parser->s_fence->v_seq);
-		cs->out.handle = parser->s_fence->v_seq;
+					     &job->base.s_fence->base);
+		parser->ibs[parser->num_ibs - 1].sequence = cs->out.handle;
+
 		list_sort(NULL, &parser->validated, cmp_size_smaller_first);
 		ttm_eu_fence_buffer_objects(&parser->ticket,
 				&parser->validated,
-				&parser->s_fence->base);
+				&job->base.s_fence->base);
 
-		mutex_unlock(&parser->job_lock);
+		mutex_unlock(&job->job_lock);
+		amdgpu_cs_parser_fini_late(parser);
 		up_read(&adev->exclusive_lock);
 		return 0;
 	}
-	r = amdgpu_cs_parser_prepare_job(parser);
-	if (r)
-		goto out;
 
 	cs->out.handle = parser->ibs[parser->num_ibs - 1].sequence;
 out:
-	amdgpu_cs_parser_fini(parser, r, true);
+	amdgpu_cs_parser_fini(parser, r, reserved_buffers);
 	up_read(&adev->exclusive_lock);
 	r = amdgpu_cs_handle_lockup(adev, r);
 	return r;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
index 08bc7722ddb8..20cbc4eb5a6f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
@@ -229,17 +229,13 @@ int amdgpu_ctx_put(struct amdgpu_ctx *ctx)
 }
 
 uint64_t amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx, struct amdgpu_ring *ring,
-			      struct fence *fence, uint64_t queued_seq)
+			      struct fence *fence)
 {
 	struct amdgpu_ctx_ring *cring = & ctx->rings[ring->idx];
-	uint64_t seq = 0;
+	uint64_t seq = cring->sequence;
 	unsigned idx = 0;
 	struct fence *other = NULL;
 
-	if (amdgpu_enable_scheduler)
-		seq = queued_seq;
-	else
-		seq = cring->sequence;
 	idx = seq % AMDGPU_CTX_MAX_CS_PENDING;
 	other = cring->fences[idx];
 	if (other) {
@@ -253,8 +249,7 @@ uint64_t amdgpu_ctx_add_fence(struct amdgpu_ctx *ctx, struct amdgpu_ring *ring,
 
 	spin_lock(&ctx->ring_lock);
 	cring->fences[idx] = fence;
-	if (!amdgpu_enable_scheduler)
-		cring->sequence++;
+	cring->sequence++;
 	spin_unlock(&ctx->ring_lock);
 
 	fence_put(other);
@@ -267,21 +262,16 @@ struct fence *amdgpu_ctx_get_fence(struct amdgpu_ctx *ctx,
 {
 	struct amdgpu_ctx_ring *cring = & ctx->rings[ring->idx];
 	struct fence *fence;
-	uint64_t queued_seq;
 
 	spin_lock(&ctx->ring_lock);
-	if (amdgpu_enable_scheduler)
-		queued_seq = amd_sched_next_queued_seq(&cring->entity);
-	else
-		queued_seq = cring->sequence;
 
-	if (seq >= queued_seq) {
+	if (seq >= cring->sequence) {
 		spin_unlock(&ctx->ring_lock);
 		return ERR_PTR(-EINVAL);
 	}
 
 
-	if (seq + AMDGPU_CTX_MAX_CS_PENDING < queued_seq) {
+	if (seq + AMDGPU_CTX_MAX_CS_PENDING < cring->sequence) {
 		spin_unlock(&ctx->ring_lock);
 		return NULL;
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index e6fa27805207..0fcc0bd1622c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -49,9 +49,10 @@
 /*
  * KMS wrapper.
  * - 3.0.0 - initial driver
+ * - 3.1.0 - allow reading more status registers (GRBM, SRBM, SDMA, CP)
  */
 #define KMS_DRIVER_MAJOR	3
-#define KMS_DRIVER_MINOR	0
+#define KMS_DRIVER_MINOR	1
 #define KMS_DRIVER_PATCHLEVEL	0
 
 int amdgpu_vram_limit = 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
index 98500f1756f7..f446bf2fedc9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
@@ -626,10 +626,10 @@ void amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring)
 	ring->fence_drv.ring = ring;
 
 	if (amdgpu_enable_scheduler) {
-		ring->scheduler = amd_sched_create((void *)ring->adev,
-						   &amdgpu_sched_ops,
-						   ring->idx, 5, 0,
-						   amdgpu_sched_hw_submission);
+		ring->scheduler = amd_sched_create(&amdgpu_sched_ops,
+						   ring->idx,
+						   amdgpu_sched_hw_submission,
+						   (void *)ring->adev);
 		if (!ring->scheduler)
 			DRM_ERROR("Failed to create scheduler on ring %d.\n",
 				  ring->idx);
@@ -836,22 +836,37 @@ static inline bool amdgpu_test_signaled(struct amdgpu_fence *fence)
 	return test_bit(FENCE_FLAG_SIGNALED_BIT, &fence->base.flags);
 }
 
-static inline bool amdgpu_test_signaled_any(struct amdgpu_fence **fences)
+static bool amdgpu_test_signaled_any(struct fence **fences, uint32_t count)
 {
 	int idx;
-	struct amdgpu_fence *fence;
+	struct fence *fence;
 
-	idx = 0;
-	for (idx = 0; idx < AMDGPU_MAX_RINGS; ++idx) {
+	for (idx = 0; idx < count; ++idx) {
 		fence = fences[idx];
 		if (fence) {
-			if (test_bit(FENCE_FLAG_SIGNALED_BIT, &fence->base.flags))
+			if (test_bit(FENCE_FLAG_SIGNALED_BIT, &fence->flags))
 				return true;
 		}
 	}
 	return false;
 }
 
+static bool amdgpu_test_signaled_all(struct fence **fences, uint32_t count)
+{
+	int idx;
+	struct fence *fence;
+
+	for (idx = 0; idx < count; ++idx) {
+		fence = fences[idx];
+		if (fence) {
+			if (!test_bit(FENCE_FLAG_SIGNALED_BIT, &fence->flags))
+				return false;
+		}
+	}
+
+	return true;
+}
+
 struct amdgpu_wait_cb {
 	struct fence_cb base;
 	struct task_struct *task;
@@ -867,33 +882,56 @@ static void amdgpu_fence_wait_cb(struct fence *fence, struct fence_cb *cb)
 static signed long amdgpu_fence_default_wait(struct fence *f, bool intr,
 					     signed long t)
 {
-	struct amdgpu_fence *array[AMDGPU_MAX_RINGS];
 	struct amdgpu_fence *fence = to_amdgpu_fence(f);
 	struct amdgpu_device *adev = fence->ring->adev;
 
-	memset(&array[0], 0, sizeof(array));
-	array[0] = fence;
-
-	return amdgpu_fence_wait_any(adev, array, intr, t);
+	return amdgpu_fence_wait_multiple(adev, &f, 1, false, intr, t);
 }
 
-/* wait until any fence in array signaled */
-signed long amdgpu_fence_wait_any(struct amdgpu_device *adev,
-				struct amdgpu_fence **array, bool intr, signed long t)
+/**
+ * Wait the fence array with timeout
+ *
+ * @adev:     amdgpu device
+ * @array:    the fence array with amdgpu fence pointer
+ * @count:    the number of the fence array
+ * @wait_all: the flag of wait all(true) or wait any(fal