Merge tag 'amd-drm-fixes-5.16-2021-11-10' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-fixes-5.16-2021-11-10:

amdgpu:
- Don't allow partial copy from user for DC debugfs
- SRIOV fixes
- GFX9 CSB pin count fix
- Various IP version check fixes
- DP 2.0 fixes
- Limit DCN1 MPO fix to DCN1

amdkfd:
- SVM fixes
- Reset fixes

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20211110222536.7527-1-alexander.deucher@amd.com

Author: airlied

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

@@ -297,7 +297,7 @@ void amdgpu_amdkfd_ras_poison_consumption_handler(struct kgd_dev *kgd);
 void amdgpu_amdkfd_gpuvm_init_mem_limits(void);
 void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev,
 				struct amdgpu_vm *vm);
-void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo);
+void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo);
 void amdgpu_amdkfd_reserve_system_mem(uint64_t size);
 #else
 static inline
@@ -312,7 +312,7 @@ void amdgpu_amdkfd_gpuvm_destroy_cb(struct amdgpu_device *adev,
 }
 
 static inline
-void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo)
+void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo)
 {
 }
 #endif

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

@@ -207,7 +207,7 @@ static void unreserve_mem_limit(struct amdgpu_device *adev,
 	spin_unlock(&kfd_mem_limit.mem_limit_lock);
 }
 
-void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo)
+void amdgpu_amdkfd_release_notify(struct amdgpu_bo *bo)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev);
 	u32 domain = bo->preferred_domains;
@@ -219,6 +219,8 @@ void amdgpu_amdkfd_unreserve_memory_limit(struct amdgpu_bo *bo)
 	}
 
 	unreserve_mem_limit(adev, amdgpu_bo_size(bo), domain, sg);
+
+	kfree(bo->kfd_bo);
 }
 
 
@@ -734,14 +736,19 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem,
 		}
 
 		/* Add BO to VM internal data structures */
+		ret = amdgpu_bo_reserve(bo[i], false);
+		if (ret) {
+			pr_debug("Unable to reserve BO during memory attach");
+			goto unwind;
+		}
 		attachment[i]->bo_va = amdgpu_vm_bo_add(adev, vm, bo[i]);
+		amdgpu_bo_unreserve(bo[i]);
 		if (unlikely(!attachment[i]->bo_va)) {
 			ret = -ENOMEM;
 			pr_err("Failed to add BO object to VM. ret == %d\n",
 			       ret);
 			goto unwind;
 		}
-
 		attachment[i]->va = va;
 		attachment[i]->pte_flags = get_pte_flags(adev, mem);
 		attachment[i]->adev = adev;
@@ -757,7 +764,9 @@ static int kfd_mem_attach(struct amdgpu_device *adev, struct kgd_mem *mem,
 		if (!attachment[i])
 			continue;
 		if (attachment[i]->bo_va) {
+			amdgpu_bo_reserve(bo[i], true);
 			amdgpu_vm_bo_rmv(adev, attachment[i]->bo_va);
+			amdgpu_bo_unreserve(bo[i]);
 			list_del(&attachment[i]->list);
 		}
 		if (bo[i])
@@ -1568,12 +1577,12 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
 	pr_debug("Release VA 0x%llx - 0x%llx\n", mem->va,
 		mem->va + bo_size * (1 + mem->aql_queue));
 
-	ret = unreserve_bo_and_vms(&ctx, false, false);
-
 	/* Remove from VM internal data structures */
 	list_for_each_entry_safe(entry, tmp, &mem->attachments, list)
 		kfd_mem_detach(entry);
 
+	ret = unreserve_bo_and_vms(&ctx, false, false);
+
 	/* Free the sync object */
 	amdgpu_sync_free(&mem->sync);
 
@@ -1600,9 +1609,13 @@ int amdgpu_amdkfd_gpuvm_free_memory_of_gpu(
 	drm_vma_node_revoke(&mem->bo->tbo.base.vma_node, drm_priv);
 	if (mem->dmabuf)
 		dma_buf_put(mem->dmabuf);
-	drm_gem_object_put(&mem->bo->tbo.base);
 	mutex_destroy(&mem->lock);
-	kfree(mem);
+
+	/* If this releases the last reference, it will end up calling
+	 * amdgpu_amdkfd_release_notify and kfree the mem struct. That's why
+	 * this needs to be the last call here.
+	 */
+	drm_gem_object_put(&mem->bo->tbo.base);
 
 	return ret;
 }

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

@@ -3167,11 +3167,21 @@ bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type)
 {
 	switch (asic_type) {
 #if defined(CONFIG_DRM_AMD_DC)
-#if defined(CONFIG_DRM_AMD_DC_SI)
 	case CHIP_TAHITI:
 	case CHIP_PITCAIRN:
 	case CHIP_VERDE:
 	case CHIP_OLAND:
+		/*
+		 * We have systems in the wild with these ASICs that require
+		 * LVDS and VGA support which is not supported with DC.
+		 *
+		 * Fallback to the non-DC driver here by default so as not to
+		 * cause regressions.
+		 */
+#if defined(CONFIG_DRM_AMD_DC_SI)
+		return amdgpu_dc > 0;
+#else
+		return false;
 #endif
 	case CHIP_BONAIRE:
 	case CHIP_KAVERI:
@@ -4283,8 +4293,6 @@ static int amdgpu_device_reset_sriov(struct amdgpu_device *adev,
 	if (r)
 		return r;
 
-	amdgpu_amdkfd_pre_reset(adev);
-
 	/* Resume IP prior to SMC */
 	r = amdgpu_device_ip_reinit_early_sriov(adev);
 	if (r)
@@ -5020,8 +5028,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 
 		cancel_delayed_work_sync(&tmp_adev->delayed_init_work);
 
-		if (!amdgpu_sriov_vf(tmp_adev))
-			amdgpu_amdkfd_pre_reset(tmp_adev);
+		amdgpu_amdkfd_pre_reset(tmp_adev);
 
 		/*
 		 * Mark these ASICs to be reseted as untracked first

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

@@ -867,7 +867,8 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev)
 		case IP_VERSION(2, 0, 2):
 		case IP_VERSION(2, 2, 0):
 			amdgpu_device_ip_block_add(adev, &vcn_v2_0_ip_block);
-			amdgpu_device_ip_block_add(adev, &jpeg_v2_0_ip_block);
+			if (!amdgpu_sriov_vf(adev))
+				amdgpu_device_ip_block_add(adev, &jpeg_v2_0_ip_block);
 			break;
 		case IP_VERSION(2, 0, 3):
 			break;
@@ -881,6 +882,7 @@ static int amdgpu_discovery_set_mm_ip_blocks(struct amdgpu_device *adev)
 			break;
 		case IP_VERSION(3, 0, 0):
 		case IP_VERSION(3, 0, 16):
+		case IP_VERSION(3, 0, 64):
 		case IP_VERSION(3, 1, 1):
 		case IP_VERSION(3, 0, 2):
 			amdgpu_device_ip_block_add(adev, &vcn_v3_0_ip_block);

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

@@ -1274,7 +1274,7 @@ void amdgpu_bo_release_notify(struct ttm_buffer_object *bo)
 	abo = ttm_to_amdgpu_bo(bo);
 
 	if (abo->kfd_bo)
-		amdgpu_amdkfd_unreserve_memory_limit(abo);
+		amdgpu_amdkfd_release_notify(abo);
 
 	/* We only remove the fence if the resv has individualized. */
 	WARN_ON_ONCE(bo->type == ttm_bo_type_kernel

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

@@ -134,6 +134,7 @@ int amdgpu_vcn_sw_init(struct amdgpu_device *adev)
 			adev->vcn.indirect_sram = true;
 		break;
 	case IP_VERSION(3, 0, 0):
+	case IP_VERSION(3, 0, 64):
 		if (adev->ip_versions[GC_HWIP][0] == IP_VERSION(10, 3, 0))
 			fw_name = FIRMWARE_SIENNA_CICHLID;
 		else

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

@@ -806,9 +806,9 @@ static void amdgpu_xgmi_reset_ras_error_count(struct amdgpu_device *adev)
 		for (i = 0; i < ARRAY_SIZE(xgmi23_pcs_err_status_reg_aldebaran); i++)
 			pcs_clear_status(adev,
 					 xgmi23_pcs_err_status_reg_aldebaran[i]);
-		for (i = 0; i < ARRAY_SIZE(xgmi23_pcs_err_status_reg_aldebaran); i++)
+		for (i = 0; i < ARRAY_SIZE(xgmi3x16_pcs_err_status_reg_aldebaran); i++)
 			pcs_clear_status(adev,
-					 xgmi23_pcs_err_status_reg_aldebaran[i]);
+					 xgmi3x16_pcs_err_status_reg_aldebaran[i]);
 		for (i = 0; i < ARRAY_SIZE(walf_pcs_err_status_reg_aldebaran); i++)
 			pcs_clear_status(adev,
 					 walf_pcs_err_status_reg_aldebaran[i]);

drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c

@@ -8249,6 +8249,9 @@ static int gfx_v10_0_update_gfx_clock_gating(struct amdgpu_device *adev,
 static void gfx_v10_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 {
 	u32 reg, data;
+
+	amdgpu_gfx_off_ctrl(adev, false);
+
 	/* not for *_SOC15 */
 	reg = SOC15_REG_OFFSET(GC, 0, mmRLC_SPM_MC_CNTL);
 	if (amdgpu_sriov_is_pp_one_vf(adev))
@@ -8263,6 +8266,8 @@ static void gfx_v10_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 		WREG32_SOC15_NO_KIQ(GC, 0, mmRLC_SPM_MC_CNTL, data);
 	else
 		WREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL, data);
+
+	amdgpu_gfx_off_ctrl(adev, true);
 }
 
 static bool gfx_v10_0_check_rlcg_range(struct amdgpu_device *adev,

drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c

@@ -3575,12 +3575,16 @@ static void gfx_v7_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 {
 	u32 data;
 
+	amdgpu_gfx_off_ctrl(adev, false);
+
 	data = RREG32(mmRLC_SPM_VMID);
 
 	data &= ~RLC_SPM_VMID__RLC_SPM_VMID_MASK;
 	data |= (vmid & RLC_SPM_VMID__RLC_SPM_VMID_MASK) << RLC_SPM_VMID__RLC_SPM_VMID__SHIFT;
 
 	WREG32(mmRLC_SPM_VMID, data);
+
+	amdgpu_gfx_off_ctrl(adev, true);
 }
 
 static void gfx_v7_0_enable_cgcg(struct amdgpu_device *adev, bool enable)

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c

@@ -5624,6 +5624,8 @@ static void gfx_v8_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 {
 	u32 data;
 
+	amdgpu_gfx_off_ctrl(adev, false);
+
 	if (amdgpu_sriov_is_pp_one_vf(adev))
 		data = RREG32_NO_KIQ(mmRLC_SPM_VMID);
 	else
@@ -5636,6 +5638,8 @@ static void gfx_v8_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 		WREG32_NO_KIQ(mmRLC_SPM_VMID, data);
 	else
 		WREG32(mmRLC_SPM_VMID, data);
+
+	amdgpu_gfx_off_ctrl(adev, true);
 }
 
 static const struct amdgpu_rlc_funcs iceland_rlc_funcs = {