Merge tag 'amd-drm-fixes-5.5-2020-01-01' of git://people.freedesktop.org/~agd5f/linux into drm-fixes

amd-drm-fixes-5.5-2020-01-01:

amdgpu:
- ATPX regression fix
- SMU metrics table locking fixes
- gfxoff fix for raven
- RLC firmware loading stability fix

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200101151307.5242-1-alexander.deucher@amd.com

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

@@ -613,7 +613,17 @@ static bool amdgpu_atpx_detect(void)
 	bool d3_supported = false;
 	struct pci_dev *parent_pdev;
 
-	while ((pdev = pci_get_class(PCI_BASE_CLASS_DISPLAY << 16, pdev)) != NULL) {
+	while ((pdev = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, pdev)) != NULL) {
+		vga_count++;
+
+		has_atpx |= (amdgpu_atpx_pci_probe_handle(pdev) == true);
+
+		parent_pdev = pci_upstream_bridge(pdev);
+		d3_supported |= parent_pdev && parent_pdev->bridge_d3;
+		amdgpu_atpx_get_quirks(pdev);
+	}
+
+	while ((pdev = pci_get_class(PCI_CLASS_DISPLAY_OTHER << 8, pdev)) != NULL) {
 		vga_count++;
 
 		has_atpx |= (amdgpu_atpx_pci_probe_handle(pdev) == true);

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

@@ -1488,7 +1488,7 @@ out:
 
 		/* Start rlc autoload after psp recieved all the gfx firmware */
 		if (psp->autoload_supported && ucode->ucode_id == (amdgpu_sriov_vf(adev) ?
-		    AMDGPU_UCODE_ID_CP_MEC2 : AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM)) {
+		    AMDGPU_UCODE_ID_CP_MEC2 : AMDGPU_UCODE_ID_RLC_G)) {
 			ret = psp_rlc_autoload(psp);
 			if (ret) {
 				DRM_ERROR("Failed to start rlc autoload\n");

drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.h

@@ -292,10 +292,10 @@ enum AMDGPU_UCODE_ID {
 	AMDGPU_UCODE_ID_CP_MEC2_JT,
 	AMDGPU_UCODE_ID_CP_MES,
 	AMDGPU_UCODE_ID_CP_MES_DATA,
-	AMDGPU_UCODE_ID_RLC_G,
 	AMDGPU_UCODE_ID_RLC_RESTORE_LIST_CNTL,
 	AMDGPU_UCODE_ID_RLC_RESTORE_LIST_GPM_MEM,
 	AMDGPU_UCODE_ID_RLC_RESTORE_LIST_SRM_MEM,
+	AMDGPU_UCODE_ID_RLC_G,
 	AMDGPU_UCODE_ID_STORAGE,
 	AMDGPU_UCODE_ID_SMC,
 	AMDGPU_UCODE_ID_UVD,

drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c

@@ -1052,17 +1052,10 @@ static void gfx_v9_0_check_if_need_gfxoff(struct amdgpu_device *adev)
 	case CHIP_VEGA20:
 		break;
 	case CHIP_RAVEN:
-		/* Disable GFXOFF on original raven.  There are combinations
-		 * of sbios and platforms that are not stable.
-		 */
-		if (!(adev->rev_id >= 0x8 || adev->pdev->device == 0x15d8))
-			adev->pm.pp_feature &= ~PP_GFXOFF_MASK;
-		else if (!(adev->rev_id >= 0x8 || adev->pdev->device == 0x15d8)
-			 &&((adev->gfx.rlc_fw_version != 106 &&
-			     adev->gfx.rlc_fw_version < 531) ||
-			    (adev->gfx.rlc_fw_version == 53815) ||
-			    (adev->gfx.rlc_feature_version < 1) ||
-			    !adev->gfx.rlc.is_rlc_v2_1))
+		if (!(adev->rev_id >= 0x8 ||
+		      adev->pdev->device == 0x15d8) &&
+		    (adev->pm.fw_version < 0x41e2b || /* not raven1 fresh */
+		     !adev->gfx.rlc.is_rlc_v2_1)) /* without rlc save restore ucodes */
 			adev->pm.pp_feature &= ~PP_GFXOFF_MASK;
 
 		if (adev->pm.pp_feature & PP_GFXOFF_MASK)

drivers/gpu/drm/amd/powerplay/amdgpu_smu.c

@@ -866,6 +866,7 @@ static int smu_sw_init(void *handle)
 	smu->smu_baco.platform_support = false;
 
 	mutex_init(&smu->sensor_lock);
+	mutex_init(&smu->metrics_lock);
 
 	smu->watermarks_bitmap = 0;
 	smu->power_profile_mode = PP_SMC_POWER_PROFILE_BOOTUP_DEFAULT;

drivers/gpu/drm/amd/powerplay/arcturus_ppt.c

@@ -862,18 +862,21 @@ static int arcturus_get_metrics_table(struct smu_context *smu,
 	struct smu_table_context *smu_table= &smu->smu_table;
 	int ret = 0;
 
+	mutex_lock(&smu->metrics_lock);
 	if (!smu_table->metrics_time ||
 	     time_after(jiffies, smu_table->metrics_time + HZ / 1000)) {
 		ret = smu_update_table(smu, SMU_TABLE_SMU_METRICS, 0,
 				(void *)smu_table->metrics_table, false);
 		if (ret) {
 			pr_info("Failed to export SMU metrics table!\n");
+			mutex_unlock(&smu->metrics_lock);
 			return ret;
 		}
 		smu_table->metrics_time = jiffies;
 	}
 
 	memcpy(metrics_table, smu_table->metrics_table, sizeof(SmuMetrics_t));
+	mutex_unlock(&smu->metrics_lock);
 
 	return ret;
 }

drivers/gpu/drm/amd/powerplay/inc/amdgpu_smu.h

@@ -349,6 +349,7 @@ struct smu_context
 	const struct pptable_funcs	*ppt_funcs;
 	struct mutex			mutex;
 	struct mutex			sensor_lock;
+	struct mutex			metrics_lock;
 	uint64_t pool_size;
 
 	struct smu_table_context	smu_table;

drivers/gpu/drm/amd/powerplay/navi10_ppt.c

@@ -562,17 +562,20 @@ static int navi10_get_metrics_table(struct smu_context *smu,
 	struct smu_table_context *smu_table= &smu->smu_table;
 	int ret = 0;
 
+	mutex_lock(&smu->metrics_lock);
 	if (!smu_table->metrics_time || time_after(jiffies, smu_table->metrics_time + msecs_to_jiffies(100))) {
 		ret = smu_update_table(smu, SMU_TABLE_SMU_METRICS, 0,
 				(void *)smu_table->metrics_table, false);
 		if (ret) {
 			pr_info("Failed to export SMU metrics table!\n");
+			mutex_unlock(&smu->metrics_lock);
 			return ret;
 		}
 		smu_table->metrics_time = jiffies;
 	}
 
 	memcpy(metrics_table, smu_table->metrics_table, sizeof(SmuMetrics_t));
+	mutex_unlock(&smu->metrics_lock);
 
 	return ret;
 }

drivers/gpu/drm/amd/powerplay/vega20_ppt.c

@@ -1678,17 +1678,20 @@ static int vega20_get_metrics_table(struct smu_context *smu,
 	struct smu_table_context *smu_table= &smu->smu_table;
 	int ret = 0;
 
+	mutex_lock(&smu->metrics_lock);
 	if (!smu_table->metrics_time || time_after(jiffies, smu_table->metrics_time + HZ / 1000)) {
 		ret = smu_update_table(smu, SMU_TABLE_SMU_METRICS, 0,
 				(void *)smu_table->metrics_table, false);
 		if (ret) {
 			pr_info("Failed to export SMU metrics table!\n");
+			mutex_unlock(&smu->metrics_lock);
 			return ret;
 		}
 		smu_table->metrics_time = jiffies;
 	}
 
 	memcpy(metrics_table, smu_table->metrics_table, sizeof(SmuMetrics_t));
+	mutex_unlock(&smu->metrics_lock);
 
 	return ret;
 }