Merge tag 'amd-drm-fixes-6.5-2023-08-16' of...

Merge tag 'amd-drm-fixes-6.5-2023-08-16' of https://gitlab.freedesktop.org/agd5f/linux into drm-fixes

amd-drm-fixes-6.5-2023-08-16:

amdgpu:
- SMU 13.x fixes
- Fix mcbp parameter for gfx9
- SMU 11.x fixes
- Temporary fix for large numbers of XCP partitions
- S0ix fixes
- DCN 2.0 fix
Signed-off-by: Dave Airlie <airlied@redhat.com>

From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230816200226.10771-1-alexander.deucher@amd.com

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

@@ -3722,10 +3722,11 @@ static void amdgpu_device_set_mcbp(struct amdgpu_device *adev)
 {
 	if (amdgpu_mcbp == 1)
 		adev->gfx.mcbp = true;
-
-	if ((adev->ip_versions[GC_HWIP][0] >= IP_VERSION(9, 0, 0)) &&
-	    (adev->ip_versions[GC_HWIP][0] < IP_VERSION(10, 0, 0)) &&
-	    adev->gfx.num_gfx_rings)
+	else if (amdgpu_mcbp == 0)
+		adev->gfx.mcbp = false;
+	else if ((adev->ip_versions[GC_HWIP][0] >= IP_VERSION(9, 0, 0)) &&
+		 (adev->ip_versions[GC_HWIP][0] < IP_VERSION(10, 0, 0)) &&
+		 adev->gfx.num_gfx_rings)
 		adev->gfx.mcbp = true;
 
 	if (amdgpu_sriov_vf(adev))
@@ -4393,6 +4394,7 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon)
 		drm_fb_helper_set_suspend_unlocked(adev_to_drm(adev)->fb_helper, true);
 
 	cancel_delayed_work_sync(&adev->delayed_init_work);
+	flush_delayed_work(&adev->gfx.gfx_off_delay_work);
 
 	amdgpu_ras_suspend(adev);
 

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

@@ -551,6 +551,41 @@ int amdgpu_fence_driver_sw_init(struct amdgpu_device *adev)
 	return 0;
 }
 
+/**
+ * amdgpu_fence_need_ring_interrupt_restore - helper function to check whether
+ * fence driver interrupts need to be restored.
+ *
+ * @ring: ring that to be checked
+ *
+ * Interrupts for rings that belong to GFX IP don't need to be restored
+ * when the target power state is s0ix.
+ *
+ * Return true if need to restore interrupts, false otherwise.
+ */
+static bool amdgpu_fence_need_ring_interrupt_restore(struct amdgpu_ring *ring)
+{
+	struct amdgpu_device *adev = ring->adev;
+	bool is_gfx_power_domain = false;
+
+	switch (ring->funcs->type) {
+	case AMDGPU_RING_TYPE_SDMA:
+	/* SDMA 5.x+ is part of GFX power domain so it's covered by GFXOFF */
+		if (adev->ip_versions[SDMA0_HWIP][0] >= IP_VERSION(5, 0, 0))
+			is_gfx_power_domain = true;
+		break;
+	case AMDGPU_RING_TYPE_GFX:
+	case AMDGPU_RING_TYPE_COMPUTE:
+	case AMDGPU_RING_TYPE_KIQ:
+	case AMDGPU_RING_TYPE_MES:
+		is_gfx_power_domain = true;
+		break;
+	default:
+		break;
+	}
+
+	return !(adev->in_s0ix && is_gfx_power_domain);
+}
+
 /**
  * amdgpu_fence_driver_hw_fini - tear down the fence driver
  * for all possible rings.
@@ -579,7 +614,8 @@ void amdgpu_fence_driver_hw_fini(struct amdgpu_device *adev)
 			amdgpu_fence_driver_force_completion(ring);
 
 		if (!drm_dev_is_unplugged(adev_to_drm(adev)) &&
-		    ring->fence_drv.irq_src)
+		    ring->fence_drv.irq_src &&
+		    amdgpu_fence_need_ring_interrupt_restore(ring))
 			amdgpu_irq_put(adev, ring->fence_drv.irq_src,
 				       ring->fence_drv.irq_type);
 
@@ -655,7 +691,8 @@ void amdgpu_fence_driver_hw_init(struct amdgpu_device *adev)
 			continue;
 
 		/* enable the interrupt */
-		if (ring->fence_drv.irq_src)
+		if (ring->fence_drv.irq_src &&
+		    amdgpu_fence_need_ring_interrupt_restore(ring))
 			amdgpu_irq_get(adev, ring->fence_drv.irq_src,
 				       ring->fence_drv.irq_type);
 	}

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

@@ -692,15 +692,8 @@ void amdgpu_gfx_off_ctrl(struct amdgpu_device *adev, bool enable)
 
 		if (adev->gfx.gfx_off_req_count == 0 &&
 		    !adev->gfx.gfx_off_state) {
-			/* If going to s2idle, no need to wait */
-			if (adev->in_s0ix) {
-				if (!amdgpu_dpm_set_powergating_by_smu(adev,
-						AMD_IP_BLOCK_TYPE_GFX, true))
-					adev->gfx.gfx_off_state = true;
-			} else {
-				schedule_delayed_work(&adev->gfx.gfx_off_delay_work,
+			schedule_delayed_work(&adev->gfx.gfx_off_delay_work,
 					      delay);
-			}
 		}
 	} else {
 		if (adev->gfx.gfx_off_req_count == 0) {

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

@@ -397,7 +397,7 @@ void amdgpu_sw_ring_ib_begin(struct amdgpu_ring *ring)
 	struct amdgpu_ring_mux *mux = &adev->gfx.muxer;
 
 	WARN_ON(!ring->is_sw_ring);
-	if (ring->hw_prio > AMDGPU_RING_PRIO_DEFAULT) {
+	if (adev->gfx.mcbp && ring->hw_prio > AMDGPU_RING_PRIO_DEFAULT) {
 		if (amdgpu_mcbp_scan(mux) > 0)
 			amdgpu_mcbp_trigger_preempt(mux);
 		return;

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

@@ -239,8 +239,13 @@ static int amdgpu_xcp_dev_alloc(struct amdgpu_device *adev)
 
 	for (i = 1; i < MAX_XCP; i++) {
 		ret = amdgpu_xcp_drm_dev_alloc(&p_ddev);
-		if (ret)
+		if (ret == -ENOSPC) {
+			dev_warn(adev->dev,
+			"Skip xcp node #%d when out of drm node resource.", i);
+			return 0;
+		} else if (ret) {
 			return ret;
+		}
 
 		/* Redirect all IOCTLs to the primary device */
 		adev->xcp_mgr->xcp[i].rdev = p_ddev->render->dev;
@@ -328,6 +333,9 @@ int amdgpu_xcp_dev_register(struct amdgpu_device *adev,
 		return 0;
 
 	for (i = 1; i < MAX_XCP; i++) {
+		if (!adev->xcp_mgr->xcp[i].ddev)
+			break;
+
 		ret = drm_dev_register(adev->xcp_mgr->xcp[i].ddev, ent->driver_data);
 		if (ret)
 			return ret;
@@ -345,6 +353,9 @@ void amdgpu_xcp_dev_unplug(struct amdgpu_device *adev)
 		return;
 
 	for (i = 1; i < MAX_XCP; i++) {
+		if (!adev->xcp_mgr->xcp[i].ddev)
+			break;
+
 		p_ddev = adev->xcp_mgr->xcp[i].ddev;
 		drm_dev_unplug(p_ddev);
 		p_ddev->render->dev = adev->xcp_mgr->xcp[i].rdev;

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

@@ -1965,7 +1965,14 @@ int kfd_topology_add_device(struct kfd_node *gpu)
 	const char *asic_name = amdgpu_asic_name[gpu->adev->asic_type];
 
 	gpu_id = kfd_generate_gpu_id(gpu);
-	pr_debug("Adding new GPU (ID: 0x%x) to topology\n", gpu_id);
+	if (gpu->xcp && !gpu->xcp->ddev) {
+		dev_warn(gpu->adev->dev,
+		"Won't add GPU (ID: 0x%x) to topology since it has no drm node assigned.",
+		gpu_id);
+		return 0;
+	} else {
+		pr_debug("Adding new GPU (ID: 0x%x) to topology\n", gpu_id);
+	}
 
 	/* Check to see if this gpu device exists in the topology_device_list.
 	 * If so, assign the gpu to that device,

drivers/gpu/drm/amd/display/dc/dcn20/dcn20_resource.c

@@ -712,7 +712,7 @@ static const struct dc_debug_options debug_defaults_drv = {
 		.timing_trace = false,
 		.clock_trace = true,
 		.disable_pplib_clock_request = true,
-		.pipe_split_policy = MPC_SPLIT_DYNAMIC,
+		.pipe_split_policy = MPC_SPLIT_AVOID_MULT_DISP,
 		.force_single_disp_pipe_split = false,
 		.disable_dcc = DCC_ENABLE,
 		.vsr_support = true,

drivers/gpu/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c

@@ -588,7 +588,9 @@ static int sienna_cichlid_tables_init(struct smu_context *smu)
 	return -ENOMEM;
 }
 
-static uint32_t sienna_cichlid_get_throttler_status_locked(struct smu_context *smu)
+static uint32_t sienna_cichlid_get_throttler_status_locked(struct smu_context *smu,
+							   bool use_metrics_v3,
+							   bool use_metrics_v2)
 {
 	struct smu_table_context *smu_table= &smu->smu_table;
 	SmuMetricsExternal_t *metrics_ext =
@@ -596,13 +598,11 @@ static uint32_t sienna_cichlid_get_throttler_status_locked(struct smu_context *s
 	uint32_t throttler_status = 0;
 	int i;
 
-	if ((smu->adev->ip_versions[MP1_HWIP][0] == IP_VERSION(11, 0, 7)) &&
-	     (smu->smc_fw_version >= 0x3A4900)) {
+	if (use_metrics_v3) {
 		for (i = 0; i < THROTTLER_COUNT; i++)
 			throttler_status |=
 				(metrics_ext->SmuMetrics_V3.ThrottlingPercentage[i] ? 1U << i : 0);
-	} else if ((smu->adev->ip_versions[MP1_HWIP][0] == IP_VERSION(11, 0, 7)) &&
-	     (smu->smc_fw_version >= 0x3A4300)) {
+	} else if (use_metrics_v2) {
 		for (i = 0; i < THROTTLER_COUNT; i++)
 			throttler_status |=
 				(metrics_ext->SmuMetrics_V2.ThrottlingPercentage[i] ? 1U << i : 0);
@@ -864,7 +864,7 @@ static int sienna_cichlid_get_smu_metrics_data(struct smu_context *smu,
 			metrics->TemperatureVrSoc) * SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
 		break;
 	case METRICS_THROTTLER_STATUS:
-		*value = sienna_cichlid_get_throttler_status_locked(smu);
+		*value = sienna_cichlid_get_throttler_status_locked(smu, use_metrics_v3, use_metrics_v2);
 		break;
 	case METRICS_CURR_FANSPEED:
 		*value = use_metrics_v3 ? metrics_v3->CurrFanSpeed :
@@ -4017,7 +4017,7 @@ static ssize_t sienna_cichlid_get_gpu_metrics(struct smu_context *smu,
 	gpu_metrics->current_dclk1 = use_metrics_v3 ? metrics_v3->CurrClock[PPCLK_DCLK_1] :
 		use_metrics_v2 ? metrics_v2->CurrClock[PPCLK_DCLK_1] : metrics->CurrClock[PPCLK_DCLK_1];
 
-	gpu_metrics->throttle_status = sienna_cichlid_get_throttler_status_locked(smu);
+	gpu_metrics->throttle_status = sienna_cichlid_get_throttler_status_locked(smu, use_metrics_v3, use_metrics_v2);
 	gpu_metrics->indep_throttle_status =
 			smu_cmn_get_indep_throttler_status(gpu_metrics->throttle_status,
 							   sienna_cichlid_throttler_map);

drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_0_ppt.c

@@ -331,6 +331,7 @@ static int smu_v13_0_0_check_powerplay_table(struct smu_context *smu)
 	struct smu_13_0_0_powerplay_table *powerplay_table =
 		table_context->power_play_table;
 	struct smu_baco_context *smu_baco = &smu->smu_baco;
+	PPTable_t *pptable = smu->smu_table.driver_pptable;
 #if 0
 	PPTable_t *pptable = smu->smu_table.driver_pptable;
 	const OverDriveLimits_t * const overdrive_upperlimits =
@@ -371,6 +372,9 @@ static int smu_v13_0_0_check_powerplay_table(struct smu_context *smu)
 	table_context->thermal_controller_type =
 		powerplay_table->thermal_controller_type;
 
+	smu->adev->pm.no_fan =
+		!(pptable->SkuTable.FeaturesToRun[0] & (1 << FEATURE_FAN_CONTROL_BIT));
+
 	return 0;
 }
 

drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c

@@ -81,9 +81,10 @@
 #define EPSILON 1
 
 #define smnPCIE_ESM_CTRL 0x193D0
-#define smnPCIE_LC_LINK_WIDTH_CNTL 0x1ab40288
+#define smnPCIE_LC_LINK_WIDTH_CNTL 0x1a340288
 #define PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_RD_MASK 0x00000070L
 #define PCIE_LC_LINK_WIDTH_CNTL__LC_LINK_WIDTH_RD__SHIFT 0x4
+#define MAX_LINK_WIDTH 6
 
 static const struct cmn2asic_msg_mapping smu_v13_0_6_message_map[SMU_MSG_MAX_COUNT] = {
 	MSG_MAP(TestMessage,			     PPSMC_MSG_TestMessage,			0),
@@ -708,16 +709,19 @@ static int smu_v13_0_6_get_smu_metrics_data(struct smu_context *smu,
 		*value = SMUQ10_TO_UINT(metrics->SocketPower) << 8;
 		break;
 	case METRICS_TEMPERATURE_HOTSPOT:
-		*value = SMUQ10_TO_UINT(metrics->MaxSocketTemperature);
+		*value = SMUQ10_TO_UINT(metrics->MaxSocketTemperature) *
+			 SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
 		break;
 	case METRICS_TEMPERATURE_MEM:
-		*value = SMUQ10_TO_UINT(metrics->MaxHbmTemperature);
+		*value = SMUQ10_TO_UINT(metrics->MaxHbmTemperature) *
+			 SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
 		break;
 	/* This is the max of all VRs and not just SOC VR.
 	 * No need to define another data type for the same.
 	 */
 	case METRICS_TEMPERATURE_VRSOC:
-		*value = SMUQ10_TO_UINT(metrics->MaxVrTemperature);
+		*value = SMUQ10_TO_UINT(metrics->MaxVrTemperature) *
+			 SMU_TEMPERATURE_UNITS_PER_CENTIGRADES;
 		break;
 	default:
 		*value = UINT_MAX;
@@ -1966,6 +1970,7 @@ static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table
 	struct amdgpu_device *adev = smu->adev;
 	int ret = 0, inst0, xcc0;
 	MetricsTable_t *metrics;
+	u16 link_width_level;
 
 	inst0 = adev->sdma.instance[0].aid_id;
 	xcc0 = GET_INST(GC, 0);
@@ -2016,8 +2021,12 @@ static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table
 	gpu_metrics->throttle_status = 0;
 
 	if (!(adev->flags & AMD_IS_APU)) {
+		link_width_level = smu_v13_0_6_get_current_pcie_link_width_level(smu);
+		if (link_width_level > MAX_LINK_WIDTH)
+			link_width_level = 0;
+
 		gpu_metrics->pcie_link_width =
-			smu_v13_0_6_get_current_pcie_link_width_level(smu);
+			DECODE_LANE_WIDTH(link_width_level);
 		gpu_metrics->pcie_link_speed =
 			smu_v13_0_6_get_current_pcie_link_speed(smu);
 	}