Merge tag 'amd-drm-fixes-5.18-2022-04-27' of https://gitlab.freedesktop.org/agd5f/linux into drm-fixes

amd-drm-fixes-5.18-2022-04-27:

amdgpu:
- Runtime pm fix
- DCN memory leak fix in error path
- SI DPM deadlock fix
- S0ix fix

amdkfd:
- GWS fix
- GWS support for CRIU

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220428023232.5794-1-alexander.deucher@amd.com

drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c

@@ -2395,6 +2395,71 @@ static int amdgpu_pmops_restore(struct device *dev)
 	return amdgpu_device_resume(drm_dev, true);
 }
 
+static int amdgpu_runtime_idle_check_display(struct device *dev)
+{
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct drm_device *drm_dev = pci_get_drvdata(pdev);
+	struct amdgpu_device *adev = drm_to_adev(drm_dev);
+
+	if (adev->mode_info.num_crtc) {
+		struct drm_connector *list_connector;
+		struct drm_connector_list_iter iter;
+		int ret = 0;
+
+		/* XXX: Return busy if any displays are connected to avoid
+		 * possible display wakeups after runtime resume due to
+		 * hotplug events in case any displays were connected while
+		 * the GPU was in suspend.  Remove this once that is fixed.
+		 */
+		mutex_lock(&drm_dev->mode_config.mutex);
+		drm_connector_list_iter_begin(drm_dev, &iter);
+		drm_for_each_connector_iter(list_connector, &iter) {
+			if (list_connector->status == connector_status_connected) {
+				ret = -EBUSY;
+				break;
+			}
+		}
+		drm_connector_list_iter_end(&iter);
+		mutex_unlock(&drm_dev->mode_config.mutex);
+
+		if (ret)
+			return ret;
+
+		if (amdgpu_device_has_dc_support(adev)) {
+			struct drm_crtc *crtc;
+
+			drm_for_each_crtc(crtc, drm_dev) {
+				drm_modeset_lock(&crtc->mutex, NULL);
+				if (crtc->state->active)
+					ret = -EBUSY;
+				drm_modeset_unlock(&crtc->mutex);
+				if (ret < 0)
+					break;
+			}
+		} else {
+			mutex_lock(&drm_dev->mode_config.mutex);
+			drm_modeset_lock(&drm_dev->mode_config.connection_mutex, NULL);
+
+			drm_connector_list_iter_begin(drm_dev, &iter);
+			drm_for_each_connector_iter(list_connector, &iter) {
+				if (list_connector->dpms ==  DRM_MODE_DPMS_ON) {
+					ret = -EBUSY;
+					break;
+				}
+			}
+
+			drm_connector_list_iter_end(&iter);
+
+			drm_modeset_unlock(&drm_dev->mode_config.connection_mutex);
+			mutex_unlock(&drm_dev->mode_config.mutex);
+		}
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
 static int amdgpu_pmops_runtime_suspend(struct device *dev)
 {
 	struct pci_dev *pdev = to_pci_dev(dev);
@@ -2407,6 +2472,10 @@ static int amdgpu_pmops_runtime_suspend(struct device *dev)
 		return -EBUSY;
 	}
 
+	ret = amdgpu_runtime_idle_check_display(dev);
+	if (ret)
+		return ret;
+
 	/* wait for all rings to drain before suspending */
 	for (i = 0; i < AMDGPU_MAX_RINGS; i++) {
 		struct amdgpu_ring *ring = adev->rings[i];
@@ -2516,41 +2585,7 @@ static int amdgpu_pmops_runtime_idle(struct device *dev)
 		return -EBUSY;
 	}
 
-	if (amdgpu_device_has_dc_support(adev)) {
-		struct drm_crtc *crtc;
-
-		drm_for_each_crtc(crtc, drm_dev) {
-			drm_modeset_lock(&crtc->mutex, NULL);
-			if (crtc->state->active)
-				ret = -EBUSY;
-			drm_modeset_unlock(&crtc->mutex);
-			if (ret < 0)
-				break;
-		}
-
-	} else {
-		struct drm_connector *list_connector;
-		struct drm_connector_list_iter iter;
-
-		mutex_lock(&drm_dev->mode_config.mutex);
-		drm_modeset_lock(&drm_dev->mode_config.connection_mutex, NULL);
-
-		drm_connector_list_iter_begin(drm_dev, &iter);
-		drm_for_each_connector_iter(list_connector, &iter) {
-			if (list_connector->dpms ==  DRM_MODE_DPMS_ON) {
-				ret = -EBUSY;
-				break;
-			}
-		}
-
-		drm_connector_list_iter_end(&iter);
-
-		drm_modeset_unlock(&drm_dev->mode_config.connection_mutex);
-		mutex_unlock(&drm_dev->mode_config.mutex);
-	}
-
-	if (ret == -EBUSY)
-		DRM_DEBUG_DRIVER("failing to power off - crtc active\n");
+	ret = amdgpu_runtime_idle_check_display(dev);
 
 	pm_runtime_mark_last_busy(dev);
 	pm_runtime_autosuspend(dev);

drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c

@@ -1151,6 +1151,16 @@ static int gmc_v10_0_set_clockgating_state(void *handle,
 	int r;
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	/*
+	 * The issue mmhub can't disconnect from DF with MMHUB clock gating being disabled
+	 * is a new problem observed at DF 3.0.3, however with the same suspend sequence not
+	 * seen any issue on the DF 3.0.2 series platform.
+	 */
+	if (adev->in_s0ix && adev->ip_versions[DF_HWIP][0] > IP_VERSION(3, 0, 2)) {
+		dev_dbg(adev->dev, "keep mmhub clock gating being enabled for s0ix\n");
+		return 0;
+	}
+
 	r = adev->mmhub.funcs->set_clockgating(adev, state);
 	if (r)
 		return r;

drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c

@@ -130,19 +130,33 @@ void program_sh_mem_settings(struct device_queue_manager *dqm,
 }
 
 static void increment_queue_count(struct device_queue_manager *dqm,
-			enum kfd_queue_type type)
+				  struct qcm_process_device *qpd,
+				  struct queue *q)
 {
 	dqm->active_queue_count++;
-	if (type == KFD_QUEUE_TYPE_COMPUTE || type == KFD_QUEUE_TYPE_DIQ)
+	if (q->properties.type == KFD_QUEUE_TYPE_COMPUTE ||
+	    q->properties.type == KFD_QUEUE_TYPE_DIQ)
 		dqm->active_cp_queue_count++;
+
+	if (q->properties.is_gws) {
+		dqm->gws_queue_count++;
+		qpd->mapped_gws_queue = true;
+	}
 }
 
 static void decrement_queue_count(struct device_queue_manager *dqm,
-			enum kfd_queue_type type)
+				  struct qcm_process_device *qpd,
+				  struct queue *q)
 {
 	dqm->active_queue_count--;
-	if (type == KFD_QUEUE_TYPE_COMPUTE || type == KFD_QUEUE_TYPE_DIQ)
+	if (q->properties.type == KFD_QUEUE_TYPE_COMPUTE ||
+	    q->properties.type == KFD_QUEUE_TYPE_DIQ)
 		dqm->active_cp_queue_count--;
+
+	if (q->properties.is_gws) {
+		dqm->gws_queue_count--;
+		qpd->mapped_gws_queue = false;
+	}
 }
 
 /*
@@ -412,7 +426,7 @@ add_queue_to_list:
 	list_add(&q->list, &qpd->queues_list);
 	qpd->queue_count++;
 	if (q->properties.is_active)
-		increment_queue_count(dqm, q->properties.type);
+		increment_queue_count(dqm, qpd, q);
 
 	/*
 	 * Unconditionally increment this counter, regardless of the queue's
@@ -601,13 +615,8 @@ static int destroy_queue_nocpsch_locked(struct device_queue_manager *dqm,
 		deallocate_vmid(dqm, qpd, q);
 	}
 	qpd->queue_count--;
-	if (q->properties.is_active) {
-		decrement_queue_count(dqm, q->properties.type);
-		if (q->properties.is_gws) {
-			dqm->gws_queue_count--;
-			qpd->mapped_gws_queue = false;
-		}
-	}
+	if (q->properties.is_active)
+		decrement_queue_count(dqm, qpd, q);
 
 	return retval;
 }
@@ -700,12 +709,11 @@ static int update_queue(struct device_queue_manager *dqm, struct queue *q,
 	 * dqm->active_queue_count to determine whether a new runlist must be
 	 * uploaded.
 	 */
-	if (q->properties.is_active && !prev_active)
-		increment_queue_count(dqm, q->properties.type);
-	else if (!q->properties.is_active && prev_active)
-		decrement_queue_count(dqm, q->properties.type);
-
-	if (q->gws && !q->properties.is_gws) {
+	if (q->properties.is_active && !prev_active) {
+		increment_queue_count(dqm, &pdd->qpd, q);
+	} else if (!q->properties.is_active && prev_active) {
+		decrement_queue_count(dqm, &pdd->qpd, q);
+	} else if (q->gws && !q->properties.is_gws) {
 		if (q->properties.is_active) {
 			dqm->gws_queue_count++;
 			pdd->qpd.mapped_gws_queue = true;
@@ -767,11 +775,7 @@ static int evict_process_queues_nocpsch(struct device_queue_manager *dqm,
 		mqd_mgr = dqm->mqd_mgrs[get_mqd_type_from_queue_type(
 				q->properties.type)];
 		q->properties.is_active = false;
-		decrement_queue_count(dqm, q->properties.type);
-		if (q->properties.is_gws) {
-			dqm->gws_queue_count--;
-			qpd->mapped_gws_queue = false;
-		}
+		decrement_queue_count(dqm, qpd, q);
 
 		if (WARN_ONCE(!dqm->sched_running, "Evict when stopped\n"))
 			continue;
@@ -817,7 +821,7 @@ static int evict_process_queues_cpsch(struct device_queue_manager *dqm,
 			continue;
 
 		q->properties.is_active = false;
-		decrement_queue_count(dqm, q->properties.type);
+		decrement_queue_count(dqm, qpd, q);
 	}
 	pdd->last_evict_timestamp = get_jiffies_64();
 	retval = execute_queues_cpsch(dqm,
@@ -888,11 +892,7 @@ static int restore_process_queues_nocpsch(struct device_queue_manager *dqm,
 		mqd_mgr = dqm->mqd_mgrs[get_mqd_type_from_queue_type(
 				q->properties.type)];
 		q->properties.is_active = true;
-		increment_queue_count(dqm, q->properties.type);
-		if (q->properties.is_gws) {
-			dqm->gws_queue_count++;
-			qpd->mapped_gws_queue = true;
-		}
+		increment_queue_count(dqm, qpd, q);
 
 		if (WARN_ONCE(!dqm->sched_running, "Restore when stopped\n"))
 			continue;
@@ -950,7 +950,7 @@ static int restore_process_queues_cpsch(struct device_queue_manager *dqm,
 			continue;
 
 		q->properties.is_active = true;
-		increment_queue_count(dqm, q->properties.type);
+		increment_queue_count(dqm, &pdd->qpd, q);
 	}
 	retval = execute_queues_cpsch(dqm,
 				KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
@@ -1378,7 +1378,7 @@ static int create_kernel_queue_cpsch(struct device_queue_manager *dqm,
 			dqm->total_queue_count);
 
 	list_add(&kq->list, &qpd->priv_queue_list);
-	increment_queue_count(dqm, kq->queue->properties.type);
+	increment_queue_count(dqm, qpd, kq->queue);
 	qpd->is_debug = true;
 	execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
 	dqm_unlock(dqm);
@@ -1392,7 +1392,7 @@ static void destroy_kernel_queue_cpsch(struct device_queue_manager *dqm,
 {
 	dqm_lock(dqm);
 	list_del(&kq->list);
-	decrement_queue_count(dqm, kq->queue->properties.type);
+	decrement_queue_count(dqm, qpd, kq->queue);
 	qpd->is_debug = false;
 	execute_queues_cpsch(dqm, KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES, 0);
 	/*
@@ -1467,7 +1467,7 @@ static int create_queue_cpsch(struct device_queue_manager *dqm, struct queue *q,
 	qpd->queue_count++;
 
 	if (q->properties.is_active) {
-		increment_queue_count(dqm, q->properties.type);
+		increment_queue_count(dqm, qpd, q);
 
 		execute_queues_cpsch(dqm,
 				KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
@@ -1683,15 +1683,11 @@ static int destroy_queue_cpsch(struct device_queue_manager *dqm,
 	list_del(&q->list);
 	qpd->queue_count--;
 	if (q->properties.is_active) {
-		decrement_queue_count(dqm, q->properties.type);
+		decrement_queue_count(dqm, qpd, q);
 		retval = execute_queues_cpsch(dqm,
 				KFD_UNMAP_QUEUES_FILTER_DYNAMIC_QUEUES, 0);
 		if (retval == -ETIME)
 			qpd->reset_wavefronts = true;
-		if (q->properties.is_gws) {
-			dqm->gws_queue_count--;
-			qpd->mapped_gws_queue = false;
-		}
 	}
 
 	/*
@@ -1932,7 +1928,7 @@ static int process_termination_cpsch(struct device_queue_manager *dqm,
 	/* Clean all kernel queues */
 	list_for_each_entry_safe(kq, kq_next, &qpd->priv_queue_list, list) {
 		list_del(&kq->list);
-		decrement_queue_count(dqm, kq->queue->properties.type);
+		decrement_queue_count(dqm, qpd, kq->queue);
 		qpd->is_debug = false;
 		dqm->total_queue_count--;
 		filter = KFD_UNMAP_QUEUES_FILTER_ALL_QUEUES;
@@ -1945,13 +1941,8 @@ static int process_termination_cpsch(struct device_queue_manager *dqm,
 		else if (q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI)
 			deallocate_sdma_queue(dqm, q);
 
-		if (q->properties.is_active) {
-			decrement_queue_count(dqm, q->properties.type);
-			if (q->properties.is_gws) {
-				dqm->gws_queue_count--;
-				qpd->mapped_gws_queue = false;
-			}
-		}
+		if (q->properties.is_active)
+			decrement_queue_count(dqm, qpd, q);
 
 		dqm->total_queue_count--;
 	}

drivers/gpu/drm/amd/amdkfd/kfd_priv.h

@@ -1103,7 +1103,7 @@ struct kfd_criu_queue_priv_data {
 	uint32_t priority;
 	uint32_t q_percent;
 	uint32_t doorbell_id;
-	uint32_t is_gws;
+	uint32_t gws;
 	uint32_t sdma_id;
 	uint32_t eop_ring_buffer_size;
 	uint32_t ctx_save_restore_area_size;

drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c

@@ -636,6 +636,8 @@ static int criu_checkpoint_queue(struct kfd_process_device *pdd,
 	q_data->ctx_save_restore_area_size =
 		q->properties.ctx_save_restore_area_size;
 
+	q_data->gws = !!q->gws;
+
 	ret = pqm_checkpoint_mqd(&pdd->process->pqm, q->properties.queue_id, mqd, ctl_stack);
 	if (ret) {
 		pr_err("Failed checkpoint queue_mqd (%d)\n", ret);
@@ -743,7 +745,6 @@ static void set_queue_properties_from_criu(struct queue_properties *qp,
 					  struct kfd_criu_queue_priv_data *q_data)
 {
 	qp->is_interop = false;
-	qp->is_gws = q_data->is_gws;
 	qp->queue_percent = q_data->q_percent;
 	qp->priority = q_data->priority;
 	qp->queue_address = q_data->q_address;
@@ -826,12 +827,15 @@ int kfd_criu_restore_queue(struct kfd_process *p,
 				NULL);
 	if (ret) {
 		pr_err("Failed to create new queue err:%d\n", ret);
-		ret = -EINVAL;
+		goto exit;
 	}
 
+	if (q_data->gws)
+		ret = pqm_set_gws(&p->pqm, q_data->q_id, pdd->dev->gws);
+
 exit:
 	if (ret)
-		pr_err("Failed to create queue (%d)\n", ret);
+		pr_err("Failed to restore queue (%d)\n", ret);
 	else
 		pr_debug("Queue id %d was restored successfully\n", queue_id);
 

drivers/gpu/drm/amd/display/dc/dcn21/dcn21_resource.c

@@ -997,6 +997,7 @@ static struct clock_source *dcn21_clock_source_create(
 		return &clk_src->base;
 	}
 
+	kfree(clk_src);
 	BREAK_TO_DEBUGGER();
 	return NULL;
 }

drivers/gpu/drm/amd/pm/amdgpu_dpm.c

@@ -427,6 +427,7 @@ int amdgpu_dpm_read_sensor(struct amdgpu_device *adev, enum amd_pp_sensors senso
 void amdgpu_dpm_compute_clocks(struct amdgpu_device *adev)
 {
 	const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
+	int i;
 
 	if (!adev->pm.dpm_enabled)
 		return;
@@ -434,6 +435,15 @@ void amdgpu_dpm_compute_clocks(struct amdgpu_device *adev)
 	if (!pp_funcs->pm_compute_clocks)
 		return;
 
+	if (adev->mode_info.num_crtc)
+		amdgpu_display_bandwidth_update(adev);
+
+	for (i = 0; i < AMDGPU_MAX_RINGS; i++) {
+		struct amdgpu_ring *ring = adev->rings[i];
+		if (ring && ring->sched.ready)
+			amdgpu_fence_wait_empty(ring);
+	}
+
 	mutex_lock(&adev->pm.mutex);
 	pp_funcs->pm_compute_clocks(adev->powerplay.pp_handle);
 	mutex_unlock(&adev->pm.mutex);
@@ -443,6 +453,20 @@ void amdgpu_dpm_enable_uvd(struct amdgpu_device *adev, bool enable)
 {
 	int ret = 0;
 
+	if (adev->family == AMDGPU_FAMILY_SI) {
+		mutex_lock(&adev->pm.mutex);
+		if (enable) {
+			adev->pm.dpm.uvd_active = true;
+			adev->pm.dpm.state = POWER_STATE_TYPE_INTERNAL_UVD;
+		} else {
+			adev->pm.dpm.uvd_active = false;
+		}
+		mutex_unlock(&adev->pm.mutex);
+
+		amdgpu_dpm_compute_clocks(adev);
+		return;
+	}
+
 	ret = amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_UVD, !enable);
 	if (ret)
 		DRM_ERROR("Dpm %s uvd failed, ret = %d. \n",
@@ -453,6 +477,21 @@ void amdgpu_dpm_enable_vce(struct amdgpu_device *adev, bool enable)
 {
 	int ret = 0;
 
+	if (adev->family == AMDGPU_FAMILY_SI) {
+		mutex_lock(&adev->pm.mutex);
+		if (enable) {
+			adev->pm.dpm.vce_active = true;
+			/* XXX select vce level based on ring/task */
+			adev->pm.dpm.vce_level = AMD_VCE_LEVEL_AC_ALL;
+		} else {
+			adev->pm.dpm.vce_active = false;
+		}
+		mutex_unlock(&adev->pm.mutex);
+
+		amdgpu_dpm_compute_clocks(adev);
+		return;
+	}
+
 	ret = amdgpu_dpm_set_powergating_by_smu(adev, AMD_IP_BLOCK_TYPE_VCE, !enable);
 	if (ret)
 		DRM_ERROR("Dpm %s vce failed, ret = %d. \n",

drivers/gpu/drm/amd/pm/legacy-dpm/legacy_dpm.c

@@ -1028,16 +1028,6 @@ static int amdgpu_dpm_change_power_state_locked(struct amdgpu_device *adev)
 void amdgpu_legacy_dpm_compute_clocks(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-	int i = 0;
-
-	if (adev->mode_info.num_crtc)
-		amdgpu_display_bandwidth_update(adev);
-
-	for (i = 0; i < AMDGPU_MAX_RINGS; i++) {
-		struct amdgpu_ring *ring = adev->rings[i];
-		if (ring && ring->sched.ready)
-			amdgpu_fence_wait_empty(ring);
-	}
 
 	amdgpu_dpm_get_active_displays(adev);
 

drivers/gpu/drm/amd/pm/legacy-dpm/si_dpm.c

@@ -3892,40 +3892,6 @@ static int si_set_boot_state(struct amdgpu_device *adev)
 }
 #endif
 
-static int si_set_powergating_by_smu(void *handle,
-				     uint32_t block_type,
-				     bool gate)
-{
-	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
-
-	switch (block_type) {
-	case AMD_IP_BLOCK_TYPE_UVD:
-		if (!gate) {
-			adev->pm.dpm.uvd_active = true;
-			adev->pm.dpm.state = POWER_STATE_TYPE_INTERNAL_UVD;
-		} else {
-			adev->pm.dpm.uvd_active = false;
-		}
-
-		amdgpu_legacy_dpm_compute_clocks(handle);
-		break;
-	case AMD_IP_BLOCK_TYPE_VCE:
-		if (!gate) {
-			adev->pm.dpm.vce_active = true;
-			/* XXX select vce level based on ring/task */
-			adev->pm.dpm.vce_level = AMD_VCE_LEVEL_AC_ALL;
-		} else {
-			adev->pm.dpm.vce_active = false;
-		}
-
-		amdgpu_legacy_dpm_compute_clocks(handle);
-		break;
-	default:
-		break;
-	}
-	return 0;
-}
-
 static int si_set_sw_state(struct amdgpu_device *adev)
 {
 	return (amdgpu_si_send_msg_to_smc(adev, PPSMC_MSG_SwitchToSwState) == PPSMC_Result_OK) ?
@@ -8125,7 +8091,6 @@ static const struct amd_pm_funcs si_dpm_funcs = {
 	.print_power_state = &si_dpm_print_power_state,
 	.debugfs_print_current_performance_level = &si_dpm_debugfs_print_current_performance_level,
 	.force_performance_level = &si_dpm_force_performance_level,
-	.set_powergating_by_smu = &si_set_powergating_by_smu,
 	.vblank_too_short = &si_dpm_vblank_too_short,
 	.set_fan_control_mode = &si_dpm_set_fan_control_mode,
 	.get_fan_control_mode = &si_dpm_get_fan_control_mode,

drivers/gpu/drm/amd/pm/powerplay/amd_powerplay.c

@@ -1487,16 +1487,6 @@ static void pp_pm_compute_clocks(void *handle)
 {
 	struct pp_hwmgr *hwmgr = handle;
 	struct amdgpu_device *adev = hwmgr->adev;
-	int i = 0;
-
-	if (adev->mode_info.num_crtc)
-		amdgpu_display_bandwidth_update(adev);
-
-	for (i = 0; i < AMDGPU_MAX_RINGS; i++) {
-		struct amdgpu_ring *ring = adev->rings[i];
-		if (ring && ring->sched.ready)
-			amdgpu_fence_wait_empty(ring);
-	}
 
 	if (!amdgpu_device_has_dc_support(adev)) {
 		amdgpu_dpm_get_active_displays(adev);