Merge tag 'amd-drm-fixes-6.2-2022-12-21' of...

Merge tag 'amd-drm-fixes-6.2-2022-12-21' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-fixes-6.2-2022-12-21:

amdgpu:
- Avoid large variable on the stack
- S0ix fixes
- SMU 13.x fixes
- VCN fix
- Add missing fence reference

amdkfd:
- Fix init vm error handling
- Fix double release of compute pasid
Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20221221205828.6093-1-alexander.deucher@amd.com

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

@@ -270,8 +270,10 @@ int amdgpu_amdkfd_get_pcie_bandwidth_mbytes(struct amdgpu_device *adev, bool is_
 	(&((struct amdgpu_fpriv *)					\
 		((struct drm_file *)(drm_priv))->driver_priv)->vm)
 
+int amdgpu_amdkfd_gpuvm_set_vm_pasid(struct amdgpu_device *adev,
+				     struct file *filp, u32 pasid);
 int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev,
-					struct file *filp, u32 pasid,
+					struct file *filp,
 					void **process_info,
 					struct dma_fence **ef);
 void amdgpu_amdkfd_gpuvm_release_process_vm(struct amdgpu_device *adev,

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

@@ -1429,10 +1429,9 @@ static void amdgpu_amdkfd_gpuvm_unpin_bo(struct amdgpu_bo *bo)
 	amdgpu_bo_unreserve(bo);
 }
 
-int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev,
-					   struct file *filp, u32 pasid,
-					   void **process_info,
-					   struct dma_fence **ef)
+int amdgpu_amdkfd_gpuvm_set_vm_pasid(struct amdgpu_device *adev,
+				     struct file *filp, u32 pasid)
+
 {
 	struct amdgpu_fpriv *drv_priv;
 	struct amdgpu_vm *avm;
@@ -1443,10 +1442,6 @@ int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev,
 		return ret;
 	avm = &drv_priv->vm;
 
-	/* Already a compute VM? */
-	if (avm->process_info)
-		return -EINVAL;
-
 	/* Free the original amdgpu allocated pasid,
 	 * will be replaced with kfd allocated pasid.
 	 */
@@ -1455,14 +1450,36 @@ int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev,
 		amdgpu_vm_set_pasid(adev, avm, 0);
 	}
 
-	/* Convert VM into a compute VM */
-	ret = amdgpu_vm_make_compute(adev, avm);
+	ret = amdgpu_vm_set_pasid(adev, avm, pasid);
 	if (ret)
 		return ret;
 
-	ret = amdgpu_vm_set_pasid(adev, avm, pasid);
+	return 0;
+}
+
+int amdgpu_amdkfd_gpuvm_acquire_process_vm(struct amdgpu_device *adev,
+					   struct file *filp,
+					   void **process_info,
+					   struct dma_fence **ef)
+{
+	struct amdgpu_fpriv *drv_priv;
+	struct amdgpu_vm *avm;
+	int ret;
+
+	ret = amdgpu_file_to_fpriv(filp, &drv_priv);
 	if (ret)
 		return ret;
+	avm = &drv_priv->vm;
+
+	/* Already a compute VM? */
+	if (avm->process_info)
+		return -EINVAL;
+
+	/* Convert VM into a compute VM */
+	ret = amdgpu_vm_make_compute(adev, avm);
+	if (ret)
+		return ret;
+
 	/* Initialize KFD part of the VM and process info */
 	ret = init_kfd_vm(avm, process_info, ef);
 	if (ret)

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

@@ -3016,14 +3016,15 @@ static int amdgpu_device_ip_suspend_phase2(struct amdgpu_device *adev)
 			continue;
 		}
 
-		/* skip suspend of gfx and psp for S0ix
+		/* skip suspend of gfx/mes and psp for S0ix
 		 * gfx is in gfxoff state, so on resume it will exit gfxoff just
 		 * like at runtime. PSP is also part of the always on hardware
 		 * so no need to suspend it.
 		 */
 		if (adev->in_s0ix &&
 		    (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_PSP ||
-		     adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GFX))
+		     adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_GFX ||
+		     adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_MES))
 			continue;
 
 		/* XXX handle errors */

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

@@ -238,8 +238,10 @@ static int amdgpu_vm_sdma_update(struct amdgpu_vm_update_params *p,
 	/* Wait for PD/PT moves to be completed */
 	dma_resv_iter_begin(&cursor, bo->tbo.base.resv, DMA_RESV_USAGE_KERNEL);
 	dma_resv_for_each_fence_unlocked(&cursor, fence) {
+		dma_fence_get(fence);
 		r = drm_sched_job_add_dependency(&p->job->base, fence);
 		if (r) {
+			dma_fence_put(fence);
 			dma_resv_iter_end(&cursor);
 			return r;
 		}

drivers/gpu/drm/amd/amdgpu/mes_v11_0.c

@@ -1342,7 +1342,8 @@ static int mes_v11_0_late_init(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
-	if (!amdgpu_in_reset(adev) &&
+	/* it's only intended for use in mes_self_test case, not for s0ix and reset */
+	if (!amdgpu_in_reset(adev) && !adev->in_s0ix &&
 	    (adev->ip_versions[GC_HWIP][0] != IP_VERSION(11, 0, 3)))
 		amdgpu_mes_self_test(adev);
 

drivers/gpu/drm/amd/amdgpu/soc21.c

@@ -666,6 +666,7 @@ static int soc21_common_early_init(void *handle)
 			AMD_CG_SUPPORT_VCN_MGCG |
 			AMD_CG_SUPPORT_JPEG_MGCG;
 		adev->pg_flags = AMD_PG_SUPPORT_VCN |
+			AMD_PG_SUPPORT_VCN_DPG |
 			AMD_PG_SUPPORT_GFX_PG |
 			AMD_PG_SUPPORT_JPEG;
 		adev->external_rev_id = adev->rev_id + 0x1;

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

@@ -689,13 +689,13 @@ void kfd_process_destroy_wq(void)
 }
 
 static void kfd_process_free_gpuvm(struct kgd_mem *mem,
-			struct kfd_process_device *pdd, void *kptr)
+			struct kfd_process_device *pdd, void **kptr)
 {
 	struct kfd_dev *dev = pdd->dev;
 
-	if (kptr) {
+	if (kptr && *kptr) {
 		amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel(mem);
-		kptr = NULL;
+		*kptr = NULL;
 	}
 
 	amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(dev->adev, mem, pdd->drm_priv);
@@ -795,7 +795,7 @@ static void kfd_process_device_destroy_ib_mem(struct kfd_process_device *pdd)
 	if (!qpd->ib_kaddr || !qpd->ib_base)
 		return;
 
-	kfd_process_free_gpuvm(qpd->ib_mem, pdd, qpd->ib_kaddr);
+	kfd_process_free_gpuvm(qpd->ib_mem, pdd, &qpd->ib_kaddr);
 }
 
 struct kfd_process *kfd_create_process(struct file *filep)
@@ -1277,7 +1277,7 @@ static void kfd_process_device_destroy_cwsr_dgpu(struct kfd_process_device *pdd)
 	if (!dev->cwsr_enabled || !qpd->cwsr_kaddr || !qpd->cwsr_base)
 		return;
 
-	kfd_process_free_gpuvm(qpd->cwsr_mem, pdd, qpd->cwsr_kaddr);
+	kfd_process_free_gpuvm(qpd->cwsr_mem, pdd, &qpd->cwsr_kaddr);
 }
 
 void kfd_process_set_trap_handler(struct qcm_process_device *qpd,
@@ -1576,9 +1576,9 @@ int kfd_process_device_init_vm(struct kfd_process_device *pdd,
 	p = pdd->process;
 	dev = pdd->dev;
 
-	ret = amdgpu_amdkfd_gpuvm_acquire_process_vm(
-		dev->adev, drm_file, p->pasid,
-		&p->kgd_process_info, &p->ef);
+	ret = amdgpu_amdkfd_gpuvm_acquire_process_vm(dev->adev, drm_file,
+						     &p->kgd_process_info,
+						     &p->ef);
 	if (ret) {
 		pr_err("Failed to create process VM object\n");
 		return ret;
@@ -1593,13 +1593,19 @@ int kfd_process_device_init_vm(struct kfd_process_device *pdd,
 	if (ret)
 		goto err_init_cwsr;
 
+	ret = amdgpu_amdkfd_gpuvm_set_vm_pasid(dev->adev, drm_file, p->pasid);
+	if (ret)
+		goto err_set_pasid;
+
 	pdd->drm_file = drm_file;
 
 	return 0;
 
+err_set_pasid:
+	kfd_process_device_destroy_cwsr_dgpu(pdd);
 err_init_cwsr:
+	kfd_process_device_destroy_ib_mem(pdd);
 err_reserve_ib_mem:
-	kfd_process_device_free_bos(pdd);
 	pdd->drm_priv = NULL;
 
 	return ret;

drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu13_driver_if_v13_0_0.h

@@ -522,9 +522,9 @@ typedef enum  {
   TEMP_HOTSPOT_M,
   TEMP_MEM,
   TEMP_VR_GFX,
+  TEMP_VR_SOC,
   TEMP_VR_MEM0,
   TEMP_VR_MEM1,
-  TEMP_VR_SOC,
   TEMP_VR_U,
   TEMP_LIQUID0,
   TEMP_LIQUID1,

drivers/gpu/drm/amd/pm/swsmu/inc/smu_v13_0.h

@@ -28,6 +28,7 @@
 #define SMU13_DRIVER_IF_VERSION_INV 0xFFFFFFFF
 #define SMU13_DRIVER_IF_VERSION_YELLOW_CARP 0x04
 #define SMU13_DRIVER_IF_VERSION_ALDE 0x08
+#define SMU13_DRIVER_IF_VERSION_SMU_V13_0_0_0 0x34
 #define SMU13_DRIVER_IF_VERSION_SMU_V13_0_4 0x07
 #define SMU13_DRIVER_IF_VERSION_SMU_V13_0_5 0x04
 #define SMU13_DRIVER_IF_VERSION_SMU_V13_0_0_10 0x32

drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0.c

@@ -290,6 +290,8 @@ int smu_v13_0_check_fw_version(struct smu_context *smu)
 		smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_ALDE;
 		break;
 	case IP_VERSION(13, 0, 0):
+		smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_SMU_V13_0_0_0;
+		break;
 	case IP_VERSION(13, 0, 10):
 		smu->smc_driver_if_version = SMU13_DRIVER_IF_VERSION_SMU_V13_0_0_10;
 		break;

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

@@ -1441,12 +1441,23 @@ static void smu_v13_0_0_get_unique_id(struct smu_context *smu)
 static int smu_v13_0_0_get_fan_speed_pwm(struct smu_context *smu,
 					 uint32_t *speed)
 {
+	int ret;
+
 	if (!speed)
 		return -EINVAL;
 
-	return smu_v13_0_0_get_smu_metrics_data(smu,
-						METRICS_CURR_FANPWM,
-						speed);
+	ret = smu_v13_0_0_get_smu_metrics_data(smu,
+					       METRICS_CURR_FANPWM,
+					       speed);
+	if (ret) {
+		dev_err(smu->adev->dev, "Failed to get fan speed(PWM)!");
+		return ret;
+	}
+
+	/* Convert the PMFW output which is in percent to pwm(255) based */
+	*speed = MIN(*speed * 255 / 100, 255);
+
+	return 0;
 }
 
 static int smu_v13_0_0_get_fan_speed_rpm(struct smu_context *smu,

drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_7_ppt.c

@@ -1363,12 +1363,23 @@ static int smu_v13_0_7_populate_umd_state_clk(struct smu_context *smu)
 static int smu_v13_0_7_get_fan_speed_pwm(struct smu_context *smu,
 					 uint32_t *speed)
 {
+	int ret;
+
 	if (!speed)
 		return -EINVAL;
 
-	return smu_v13_0_7_get_smu_metrics_data(smu,
-						METRICS_CURR_FANPWM,
-						speed);
+	ret = smu_v13_0_7_get_smu_metrics_data(smu,
+					       METRICS_CURR_FANPWM,
+					       speed);
+	if (ret) {
+		dev_err(smu->adev->dev, "Failed to get fan speed(PWM)!");
+		return ret;
+	}
+
+	/* Convert the PMFW output which is in percent to pwm(255) based */
+	*speed = MIN(*speed * 255 / 100, 255);
+
+	return 0;
 }
 
 static int smu_v13_0_7_get_fan_speed_rpm(struct smu_context *smu,
@@ -1440,7 +1451,7 @@ static int smu_v13_0_7_get_power_limit(struct smu_context *smu,
 
 static int smu_v13_0_7_get_power_profile_mode(struct smu_context *smu, char *buf)
 {
-	DpmActivityMonitorCoeffIntExternal_t activity_monitor_external[PP_SMC_POWER_PROFILE_COUNT];
+	DpmActivityMonitorCoeffIntExternal_t *activity_monitor_external;
 	uint32_t i, j, size = 0;
 	int16_t workload_type = 0;
 	int result = 0;
@@ -1448,6 +1459,12 @@ static int smu_v13_0_7_get_power_profile_mode(struct smu_context *smu, char *buf
 	if (!buf)
 		return -EINVAL;
 
+	activity_monitor_external = kcalloc(PP_SMC_POWER_PROFILE_COUNT,
+					    sizeof(*activity_monitor_external),
+					    GFP_KERNEL);
+	if (!activity_monitor_external)
+		return -ENOMEM;
+
 	size += sysfs_emit_at(buf, size, "                              ");
 	for (i = 0; i <= PP_SMC_POWER_PROFILE_WINDOW3D; i++)
 		size += sysfs_emit_at(buf, size, "%-14s%s", amdgpu_pp_profile_name[i],
@@ -1460,15 +1477,17 @@ static int smu_v13_0_7_get_power_profile_mode(struct smu_context *smu, char *buf
 		workload_type = smu_cmn_to_asic_specific_index(smu,
 							       CMN2ASIC_MAPPING_WORKLOAD,
 							       i);
-		if (workload_type < 0)
-			return -EINVAL;
+		if (workload_type < 0) {
+			result = -EINVAL;
+			goto out;
+		}
 
 		result = smu_cmn_update_table(smu,
 					  SMU_TABLE_ACTIVITY_MONITOR_COEFF, workload_type,
 					  (void *)(&activity_monitor_external[i]), false);
 		if (result) {
 			dev_err(smu->adev->dev, "[%s] Failed to get activity monitor!", __func__);
-			return result;
+			goto out;
 		}
 	}
 
@@ -1496,7 +1515,10 @@ do {													\
 	PRINT_DPM_MONITOR(Fclk_BoosterFreq);
 #undef PRINT_DPM_MONITOR
 
-	return size;
+	result = size;
+out:
+	kfree(activity_monitor_external);
+	return result;
 }
 
 static int smu_v13_0_7_set_power_profile_mode(struct smu_context *smu, long *input, uint32_t size)