Merge tag 'drm-next-2024-11-29' of https://gitlab.freedesktop.org/drm/kernel

Pull drm fixes from Dave Airlie:
 "Merge window fixes, mostly amdgpu and xe, with a few other minor ones,
  all looks fairly normal,

  i915:
   - hdcp: Fix when the first read and write are retried

  xe:
   - Wake up waiters after wait condition set to true
   - Mark the preempt fence workqueue as reclaim
   - Update xe2 graphics name string
   - Fix a couple of guc submit races
   - Fix pat index usage in migrate
   - Ensure non-cached migrate pagetable bo mappings
   - Take a PM ref in the delayed snapshot capture worker

  amdgpu:
   - SMU 13.0.6 fixes
   - XGMI fixes
   - SMU 13.0.7 fixes
   - Misc code cleanups
   - Plane refcount fixes
   - DCN 4.0.1 fixes
   - DC power fixes
   - DTO fixes
   - NBIO 7.11 fixes
   - SMU 14.0.x fixes
   - Reset fixes
   - Enable DC on LoongArch
   - Sysfs hotplug warning fix
   - Misc small fixes
   - VCN 4.0.3 fix
   - Slab usage fix
   - Jpeg delayed work fix

  amdkfd:
   - wptr handling fixes

  radeon:
   - Use ttm_bo_move_null()
   - Constify struct pci_device_id
   - Fix spurious hotplug
   - HPD fix

  rockchip
   - fix 32-bit build"

* tag 'drm-next-2024-11-29' of https://gitlab.freedesktop.org/drm/kernel: (48 commits)
  drm/xe: Take PM ref in delayed snapshot capture worker
  drm/xe/migrate: use XE_BO_FLAG_PAGETABLE
  drm/xe/migrate: fix pat index usage
  drm/xe/guc_submit: fix race around suspend_pending
  drm/xe/guc_submit: fix race around pending_disable
  drm/xe: Update xe2_graphics name string
  drm/rockchip: avoid 64-bit division
  Revert "drm/radeon: Delay Connector detecting when HPD singals is unstable"
  drm/amdgpu/jpeg: cancel the jpeg worker
  drm/amdgpu: fix usage slab after free
  drm/amdgpu/vcn: reset fw_shared when VCPU buffers corrupted on vcn v4.0.3
  drm/amdgpu: Fix sysfs warning when hotplugging
  drm/amdgpu: Add sysfs interface for vcn reset mask
  drm/amdgpu/gmc7: fix wait_for_idle callers
  drm/amd/pm: Remove arcturus min power limit
  drm/amd/pm: skip setting the power source on smu v14.0.2/3
  drm/amd/pm: disable pcie speed switching on Intel platform for smu v14.0.2/3
  drm/amdkfd: Use the correct wptr size
  drm/xe: Mark preempt fence workqueue as reclaim
  drm/xe/ufence: Wake up waiters after setting ufence->signalled
  ...

Author: torvalds

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

@@ -330,6 +330,8 @@ aldebaran_mode2_restore_hwcontext(struct amdgpu_reset_control *reset_ctl,
 	}
 
 	list_for_each_entry(tmp_adev, reset_device_list, reset_list) {
+		amdgpu_set_init_level(tmp_adev,
+				AMDGPU_INIT_LEVEL_RESET_RECOVERY);
 		dev_info(tmp_adev->dev,
 			 "GPU reset succeeded, trying to resume\n");
 		r = aldebaran_mode2_restore_ip(tmp_adev);
@@ -375,6 +377,8 @@ aldebaran_mode2_restore_hwcontext(struct amdgpu_reset_control *reset_ctl,
 							tmp_adev);
 
 		if (!r) {
+			amdgpu_set_init_level(tmp_adev,
+					      AMDGPU_INIT_LEVEL_DEFAULT);
 			amdgpu_irq_gpu_reset_resume_helper(tmp_adev);
 
 			r = amdgpu_ib_ring_tests(tmp_adev);

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

@@ -839,6 +839,7 @@ struct amdgpu_mqd {
 enum amdgpu_init_lvl_id {
 	AMDGPU_INIT_LEVEL_DEFAULT,
 	AMDGPU_INIT_LEVEL_MINIMAL_XGMI,
+	AMDGPU_INIT_LEVEL_RESET_RECOVERY,
 };
 
 struct amdgpu_init_level {

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

@@ -156,6 +156,11 @@ struct amdgpu_init_level amdgpu_init_default = {
 	.hwini_ip_block_mask = AMDGPU_IP_BLK_MASK_ALL,
 };
 
+struct amdgpu_init_level amdgpu_init_recovery = {
+	.level = AMDGPU_INIT_LEVEL_RESET_RECOVERY,
+	.hwini_ip_block_mask = AMDGPU_IP_BLK_MASK_ALL,
+};
+
 /*
  * Minimal blocks needed to be initialized before a XGMI hive can be reset. This
  * is used for cases like reset on initialization where the entire hive needs to
@@ -182,6 +187,9 @@ void amdgpu_set_init_level(struct amdgpu_device *adev,
 	case AMDGPU_INIT_LEVEL_MINIMAL_XGMI:
 		adev->init_lvl = &amdgpu_init_minimal_xgmi;
 		break;
+	case AMDGPU_INIT_LEVEL_RESET_RECOVERY:
+		adev->init_lvl = &amdgpu_init_recovery;
+		break;
 	case AMDGPU_INIT_LEVEL_DEFAULT:
 		fallthrough;
 	default:
@@ -3250,7 +3258,7 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev)
 		return r;
 	}
 
-	if (!amdgpu_in_reset(adev))
+	if (!amdgpu_reset_in_recovery(adev))
 		amdgpu_ras_set_error_query_ready(adev, true);
 
 	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_GATE);
@@ -4669,8 +4677,8 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev)
 	int idx;
 	bool px;
 
-	amdgpu_fence_driver_sw_fini(adev);
 	amdgpu_device_ip_fini(adev);
+	amdgpu_fence_driver_sw_fini(adev);
 	amdgpu_ucode_release(&adev->firmware.gpu_info_fw);
 	adev->accel_working = false;
 	dma_fence_put(rcu_dereference_protected(adev->gang_submit, true));
@@ -5419,7 +5427,7 @@ int amdgpu_device_reinit_after_reset(struct amdgpu_reset_context *reset_context)
 	struct list_head *device_list_handle;
 	bool full_reset, vram_lost = false;
 	struct amdgpu_device *tmp_adev;
-	int r;
+	int r, init_level;
 
 	device_list_handle = reset_context->reset_device_list;
 
@@ -5428,10 +5436,18 @@ int amdgpu_device_reinit_after_reset(struct amdgpu_reset_context *reset_context)
 
 	full_reset = test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags);
 
+	/**
+	 * If it's reset on init, it's default init level, otherwise keep level
+	 * as recovery level.
+	 */
+	if (reset_context->method == AMD_RESET_METHOD_ON_INIT)
+			init_level = AMDGPU_INIT_LEVEL_DEFAULT;
+	else
+			init_level = AMDGPU_INIT_LEVEL_RESET_RECOVERY;
+
 	r = 0;
 	list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
-		/* After reset, it's default init level */
-		amdgpu_set_init_level(tmp_adev, AMDGPU_INIT_LEVEL_DEFAULT);
+		amdgpu_set_init_level(tmp_adev, init_level);
 		if (full_reset) {
 			/* post card */
 			amdgpu_ras_set_fed(tmp_adev, false);
@@ -5518,6 +5534,9 @@ int amdgpu_device_reinit_after_reset(struct amdgpu_reset_context *reset_context)
 
 out:
 		if (!r) {
+			/* IP init is complete now, set level as default */
+			amdgpu_set_init_level(tmp_adev,
+					      AMDGPU_INIT_LEVEL_DEFAULT);
 			amdgpu_irq_gpu_reset_resume_helper(tmp_adev);
 			r = amdgpu_ib_ring_tests(tmp_adev);
 			if (r) {

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

@@ -1778,9 +1778,11 @@ int amdgpu_gfx_sysfs_init(struct amdgpu_device *adev)
 
 void amdgpu_gfx_sysfs_fini(struct amdgpu_device *adev)
 {
-	amdgpu_gfx_sysfs_xcp_fini(adev);
-	amdgpu_gfx_sysfs_isolation_shader_fini(adev);
-	amdgpu_gfx_sysfs_reset_mask_fini(adev);
+	if (adev->dev->kobj.sd) {
+		amdgpu_gfx_sysfs_xcp_fini(adev);
+		amdgpu_gfx_sysfs_isolation_shader_fini(adev);
+		amdgpu_gfx_sysfs_reset_mask_fini(adev);
+	}
 }
 
 int amdgpu_gfx_cleaner_shader_sw_init(struct amdgpu_device *adev,

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

@@ -447,6 +447,8 @@ int amdgpu_jpeg_sysfs_reset_mask_init(struct amdgpu_device *adev)
 
 void amdgpu_jpeg_sysfs_reset_mask_fini(struct amdgpu_device *adev)
 {
-	if (adev->jpeg.num_jpeg_inst)
-		device_remove_file(adev->dev, &dev_attr_jpeg_reset_mask);
+	if (adev->dev->kobj.sd) {
+		if (adev->jpeg.num_jpeg_inst)
+			device_remove_file(adev->dev, &dev_attr_jpeg_reset_mask);
+	}
 }

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

@@ -137,7 +137,8 @@ void amdgpu_preempt_mgr_fini(struct amdgpu_device *adev)
 	if (ret)
 		return;
 
-	device_remove_file(adev->dev, &dev_attr_mem_info_preempt_used);
+	if (adev->dev->kobj.sd)
+		device_remove_file(adev->dev, &dev_attr_mem_info_preempt_used);
 
 	ttm_resource_manager_cleanup(man);
 	ttm_set_driver_manager(&adev->mman.bdev, AMDGPU_PL_PREEMPT, NULL);

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

@@ -1298,7 +1298,7 @@ int amdgpu_ras_bind_aca(struct amdgpu_device *adev, enum amdgpu_ras_block blk,
 	struct ras_manager *obj;
 
 	/* in resume phase, no need to create aca fs node */
-	if (adev->in_suspend || amdgpu_in_reset(adev))
+	if (adev->in_suspend || amdgpu_reset_in_recovery(adev))
 		return 0;
 
 	obj = get_ras_manager(adev, blk);
@@ -3610,7 +3610,7 @@ static void amdgpu_ras_event_mgr_init(struct amdgpu_device *adev)
 	ras->event_mgr = hive ? &hive->event_mgr : &ras->__event_mgr;
 
 	/* init event manager with node 0 on xgmi system */
-	if (!amdgpu_in_reset(adev)) {
+	if (!amdgpu_reset_in_recovery(adev)) {
 		if (!hive || adev->gmc.xgmi.node_id == 0)
 			ras_event_mgr_init(ras->event_mgr);
 	}
@@ -3825,7 +3825,7 @@ int amdgpu_ras_block_late_init(struct amdgpu_device *adev,
 
 	r = amdgpu_ras_feature_enable_on_boot(adev, ras_block, 1);
 	if (r) {
-		if (adev->in_suspend || amdgpu_in_reset(adev)) {
+		if (adev->in_suspend || amdgpu_reset_in_recovery(adev)) {
 			/* in resume phase, if fail to enable ras,
 			 * clean up all ras fs nodes, and disable ras */
 			goto cleanup;
@@ -3837,7 +3837,7 @@ int amdgpu_ras_block_late_init(struct amdgpu_device *adev,
 	amdgpu_persistent_edc_harvesting(adev, ras_block);
 
 	/* in resume phase, no need to create ras fs node */
-	if (adev->in_suspend || amdgpu_in_reset(adev))
+	if (adev->in_suspend || amdgpu_reset_in_recovery(adev))
 		return 0;
 
 	ras_obj = container_of(ras_block, struct amdgpu_ras_block_object, ras_comm);
@@ -3967,7 +3967,7 @@ int amdgpu_ras_late_init(struct amdgpu_device *adev)
 	amdgpu_ras_event_mgr_init(adev);
 
 	if (amdgpu_ras_aca_is_supported(adev)) {
-		if (amdgpu_in_reset(adev)) {
+		if (amdgpu_reset_in_recovery(adev)) {
 			if (amdgpu_aca_is_enabled(adev))
 				r = amdgpu_aca_reset(adev);
 			else

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

@@ -342,3 +342,8 @@ void amdgpu_reset_get_desc(struct amdgpu_reset_context *rst_ctxt, char *buf,
 		strscpy(buf, "unknown", len);
 	}
 }
+
+bool amdgpu_reset_in_recovery(struct amdgpu_device *adev)
+{
+	return (adev->init_lvl->level == AMDGPU_INIT_LEVEL_RESET_RECOVERY);
+}

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

@@ -158,4 +158,6 @@ extern struct amdgpu_reset_handler xgmi_reset_on_init_handler;
 int amdgpu_reset_do_xgmi_reset_on_init(
 	struct amdgpu_reset_context *reset_context);
 
+bool amdgpu_reset_in_recovery(struct amdgpu_device *adev);
+
 #endif

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

@@ -451,6 +451,8 @@ void amdgpu_sdma_sysfs_reset_mask_fini(struct amdgpu_device *adev)
 	if (!amdgpu_gpu_recovery)
 		return;
 
-	if (adev->sdma.num_instances)
-		device_remove_file(adev->dev, &dev_attr_sdma_reset_mask);
+	if (adev->dev->kobj.sd) {
+		if (adev->sdma.num_instances)
+			device_remove_file(adev->dev, &dev_attr_sdma_reset_mask);
+	}
 }