sync

sys/dev/pci/drm/amd/amdgpu/amdgpu_object.c

@@ -80,9 +80,10 @@ static void amdgpu_bo_user_destroy(struct ttm_buffer_object *tbo)
 static void amdgpu_bo_vm_destroy(struct ttm_buffer_object *tbo)
 {
 	struct amdgpu_device *adev = amdgpu_ttm_adev(tbo->bdev);
-	struct amdgpu_bo *bo = ttm_to_amdgpu_bo(tbo);
+	struct amdgpu_bo *shadow_bo = ttm_to_amdgpu_bo(tbo), *bo;
 	struct amdgpu_bo_vm *vmbo;
 
+	bo = shadow_bo->parent;
 	vmbo = to_amdgpu_bo_vm(bo);
 	/* in case amdgpu_device_recover_vram got NULL of bo->parent */
 	if (!list_empty(&vmbo->shadow_list)) {
@@ -693,11 +694,6 @@ int amdgpu_bo_create_vm(struct amdgpu_device *adev,
 		return r;
 
 	*vmbo_ptr = to_amdgpu_bo_vm(bo_ptr);
-	INIT_LIST_HEAD(&(*vmbo_ptr)->shadow_list);
-	/* Set destroy callback to amdgpu_bo_vm_destroy after vmbo->shadow_list
-	 * is initialized.
-	 */
-	bo_ptr->tbo.destroy = &amdgpu_bo_vm_destroy;
 	return r;
 }
 
@@ -714,6 +710,8 @@ void amdgpu_bo_add_to_shadow_list(struct amdgpu_bo_vm *vmbo)
 
 	mutex_lock(&adev->shadow_list_lock);
 	list_add_tail(&vmbo->shadow_list, &adev->shadow_list);
+	vmbo->shadow->parent = amdgpu_bo_ref(&vmbo->bo);
+	vmbo->shadow->tbo.destroy = &amdgpu_bo_vm_destroy;
 	mutex_unlock(&adev->shadow_list_lock);
 }
 

sys/dev/pci/drm/amd/amdgpu/amdgpu_vm_pt.c

@@ -564,7 +564,6 @@ int amdgpu_vm_pt_create(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 		return r;
 	}
 
-	(*vmbo)->shadow->parent = amdgpu_bo_ref(bo);
 	amdgpu_bo_add_to_shadow_list(*vmbo);
 
 	return 0;

sys/dev/pci/drm/amd/amdgpu/amdgpu_vram_mgr.c

@@ -806,7 +806,7 @@ static void amdgpu_vram_mgr_debug(struct ttm_resource_manager *man,
 {
 	struct amdgpu_vram_mgr *mgr = to_vram_mgr(man);
 	struct drm_buddy *mm = &mgr->mm;
-	struct drm_buddy_block *block;
+	struct amdgpu_vram_reservation *rsv;
 
 	drm_printf(printer, "  vis usage:%llu\n",
 		   amdgpu_vram_mgr_vis_usage(mgr));
@@ -818,8 +818,9 @@ static void amdgpu_vram_mgr_debug(struct ttm_resource_manager *man,
 	drm_buddy_print(mm, printer);
 
 	drm_printf(printer, "reserved:\n");
-	list_for_each_entry(block, &mgr->reserved_pages, link)
-		drm_buddy_block_print(mm, block, printer);
+	list_for_each_entry(rsv, &mgr->reserved_pages, blocks)
+		drm_printf(printer, "%#018llx-%#018llx: %llu\n",
+			rsv->start, rsv->start + rsv->size, rsv->size);
 	mutex_unlock(&mgr->lock);
 }
 

sys/dev/pci/drm/amd/amdgpu/vi.c

@@ -542,8 +542,15 @@ static u32 vi_get_xclk(struct amdgpu_device *adev)
 	u32 reference_clock = adev->clock.spll.reference_freq;
 	u32 tmp;
 
-	if (adev->flags & AMD_IS_APU)
-		return reference_clock;
+	if (adev->flags & AMD_IS_APU) {
+		switch (adev->asic_type) {
+		case CHIP_STONEY:
+			/* vbios says 48Mhz, but the actual freq is 100Mhz */
+			return 10000;
+		default:
+			return reference_clock;
+		}
+	}
 
 	tmp = RREG32_SMC(ixCG_CLKPIN_CNTL_2);
 	if (REG_GET_FIELD(tmp, CG_CLKPIN_CNTL_2, MUX_TCLK_TO_XCLK))

sys/dev/pci/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c

@@ -137,7 +137,7 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_2_soc = {
 	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
 	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
 	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
-	.pct_ideal_sdp_bw_after_urgent = 100.0,
+	.pct_ideal_sdp_bw_after_urgent = 90.0,
 	.pct_ideal_fabric_bw_after_urgent = 67.0,
 	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 20.0,
 	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0, // N/A, for now keep as is until DML implemented

sys/dev/pci/drm/amd/pm/swsmu/smu11/sienna_cichlid_ppt.c

@@ -2067,33 +2067,96 @@ static int sienna_cichlid_display_disable_memory_clock_switch(struct smu_context
 	return ret;
 }
 
+static void sienna_cichlid_get_override_pcie_settings(struct smu_context *smu,
+						      uint32_t *gen_speed_override,
+						      uint32_t *lane_width_override)
+{
+	struct amdgpu_device *adev = smu->adev;
+
+	*gen_speed_override = 0xff;
+	*lane_width_override = 0xff;
+
+	switch (adev->pdev->device) {
+	case 0x73A0:
+	case 0x73A1:
+	case 0x73A2:
+	case 0x73A3:
+	case 0x73AB:
+	case 0x73AE:
+		/* Bit 7:0: PCIE lane width, 1 to 7 corresponds is x1 to x32 */
+		*lane_width_override = 6;
+		break;
+	case 0x73E0:
+	case 0x73E1:
+	case 0x73E3:
+		*lane_width_override = 4;
+		break;
+	case 0x7420:
+	case 0x7421:
+	case 0x7422:
+	case 0x7423:
+	case 0x7424:
+		*lane_width_override = 3;
+		break;
+	default:
+		break;
+	}
+}
+
+#ifndef MAX
+#define MAX(a, b)	((a) > (b) ? (a) : (b))
+#endif
+
 static int sienna_cichlid_update_pcie_parameters(struct smu_context *smu,
 					 uint32_t pcie_gen_cap,
 					 uint32_t pcie_width_cap)
 {
 	struct smu_11_0_dpm_context *dpm_context = smu->smu_dpm.dpm_context;
-
-	uint32_t smu_pcie_arg;
+	struct smu_11_0_pcie_table *pcie_table = &dpm_context->dpm_tables.pcie_table;
+	uint32_t gen_speed_override, lane_width_override;
 	uint8_t *table_member1, *table_member2;
+	uint32_t min_gen_speed, max_gen_speed;
+	uint32_t min_lane_width, max_lane_width;
+	uint32_t smu_pcie_arg;
 	int ret, i;
 
 	GET_PPTABLE_MEMBER(PcieGenSpeed, &table_member1);
 	GET_PPTABLE_MEMBER(PcieLaneCount, &table_member2);
 
-	/* lclk dpm table setup */
-	for (i = 0; i < MAX_PCIE_CONF; i++) {
-		dpm_context->dpm_tables.pcie_table.pcie_gen[i] = table_member1[i];
-		dpm_context->dpm_tables.pcie_table.pcie_lane[i] = table_member2[i];
+	sienna_cichlid_get_override_pcie_settings(smu,
+						  &gen_speed_override,
+						  &lane_width_override);
+
+	/* PCIE gen speed override */
+	if (gen_speed_override != 0xff) {
+		min_gen_speed = MIN(pcie_gen_cap, gen_speed_override);
+		max_gen_speed = MIN(pcie_gen_cap, gen_speed_override);
+	} else {
+		min_gen_speed = MAX(0, table_member1[0]);
+		max_gen_speed = MIN(pcie_gen_cap, table_member1[1]);
+		min_gen_speed = min_gen_speed > max_gen_speed ?
+				max_gen_speed : min_gen_speed;
 	}
+	pcie_table->pcie_gen[0] = min_gen_speed;
+	pcie_table->pcie_gen[1] = max_gen_speed;
+
+	/* PCIE lane width override */
+	if (lane_width_override != 0xff) {
+		min_lane_width = MIN(pcie_width_cap, lane_width_override);
+		max_lane_width = MIN(pcie_width_cap, lane_width_override);
+	} else {
+		min_lane_width = MAX(1, table_member2[0]);
+		max_lane_width = MIN(pcie_width_cap, table_member2[1]);
+		min_lane_width = min_lane_width > max_lane_width ?
+				 max_lane_width : min_lane_width;
+	}
+	pcie_table->pcie_lane[0] = min_lane_width;
+	pcie_table->pcie_lane[1] = max_lane_width;
 
 	for (i = 0; i < NUM_LINK_LEVELS; i++) {
-		smu_pcie_arg = (i << 16) |
-			((table_member1[i] <= pcie_gen_cap) ?
-			 (table_member1[i] << 8) :
-			 (pcie_gen_cap << 8)) |
-			((table_member2[i] <= pcie_width_cap) ?
-			 table_member2[i] :
-			 pcie_width_cap);
+		smu_pcie_arg = (i << 16 |
+				pcie_table->pcie_gen[i] << 8 |
+				pcie_table->pcie_lane[i]);
 
 		ret = smu_cmn_send_smc_msg_with_param(smu,
 				SMU_MSG_OverridePcieParameters,
@@ -2101,11 +2164,6 @@ static int sienna_cichlid_update_pcie_parameters(struct smu_context *smu,
 				NULL);
 		if (ret)
 			return ret;
-
-		if (table_member1[i] > pcie_gen_cap)
-			dpm_context->dpm_tables.pcie_table.pcie_gen[i] = pcie_gen_cap;
-		if (table_member2[i] > pcie_width_cap)
-			dpm_context->dpm_tables.pcie_table.pcie_lane[i] = pcie_width_cap;
 	}
 
 	return 0;

sys/dev/pci/drm/amd/pm/swsmu/smu13/smu_v13_0.c

@@ -582,11 +582,11 @@ int smu_v13_0_init_power(struct smu_context *smu)
 	if (smu_power->power_context || smu_power->power_context_size != 0)
 		return -EINVAL;
 
-	smu_power->power_context = kzalloc(sizeof(struct smu_13_0_dpm_context),
+	smu_power->power_context = kzalloc(sizeof(struct smu_13_0_power_context),
 					   GFP_KERNEL);
 	if (!smu_power->power_context)
 		return -ENOMEM;
-	smu_power->power_context_size = sizeof(struct smu_13_0_dpm_context);
+	smu_power->power_context_size = sizeof(struct smu_13_0_power_context);
 
 	return 0;
 }

sys/dev/pci/drm/i915/display/intel_dp_aux.c

@@ -119,6 +119,32 @@ static u32 skl_get_aux_clock_divider(struct intel_dp *intel_dp, int index)
 	return index ? 0 : 1;
 }
 
+static int intel_dp_aux_sync_len(void)
+{
+	int precharge = 16; /* 10-16 */
+	int preamble = 16;
+
+	return precharge + preamble;
+}
+
+static int intel_dp_aux_fw_sync_len(void)
+{
+	int precharge = 10; /* 10-16 */
+	int preamble = 8;
+
+	return precharge + preamble;
+}
+
+static int g4x_dp_aux_precharge_len(void)
+{
+	int precharge_min = 10;
+	int preamble = 16;
+
+	/* HW wants the length of the extra precharge in 2us units */
+	return (intel_dp_aux_sync_len() -
+		precharge_min - preamble) / 2;
+}
+
 static u32 g4x_get_aux_send_ctl(struct intel_dp *intel_dp,
 				int send_bytes,
 				u32 aux_clock_divider)
@@ -141,7 +167,7 @@ static u32 g4x_get_aux_send_ctl(struct intel_dp *intel_dp,
 	       timeout |
 	       DP_AUX_CH_CTL_RECEIVE_ERROR |
 	       (send_bytes << DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT) |
-	       (3 << DP_AUX_CH_CTL_PRECHARGE_2US_SHIFT) |
+	       (g4x_dp_aux_precharge_len() << DP_AUX_CH_CTL_PRECHARGE_2US_SHIFT) |
 	       (aux_clock_divider << DP_AUX_CH_CTL_BIT_CLOCK_2X_SHIFT);
 }
 
@@ -165,8 +191,8 @@ static u32 skl_get_aux_send_ctl(struct intel_dp *intel_dp,
 	      DP_AUX_CH_CTL_TIME_OUT_MAX |
 	      DP_AUX_CH_CTL_RECEIVE_ERROR |
 	      (send_bytes << DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT) |
-	      DP_AUX_CH_CTL_FW_SYNC_PULSE_SKL(24) |
-	      DP_AUX_CH_CTL_SYNC_PULSE_SKL(32);
+	      DP_AUX_CH_CTL_FW_SYNC_PULSE_SKL(intel_dp_aux_fw_sync_len()) |
+	      DP_AUX_CH_CTL_SYNC_PULSE_SKL(intel_dp_aux_sync_len());
 
 	if (intel_tc_port_in_tbt_alt_mode(dig_port))
 		ret |= DP_AUX_CH_CTL_TBT_IO;

sys/dev/pci/drm/i915/gem/selftests/i915_gem_context.c

@@ -179,97 +179,108 @@ out_file:
 }
 
 struct parallel_switch {
-	struct task_struct *tsk;
+	struct kthread_worker *worker;
+	struct kthread_work work;
 	struct intel_context *ce[2];
+	int result;
 };
 
-static int __live_parallel_switch1(void *data)
+static void __live_parallel_switch1(struct kthread_work *work)
 {
-	struct parallel_switch *arg = data;
+	struct parallel_switch *arg =
+		container_of(work, typeof(*arg), work);
 	IGT_TIMEOUT(end_time);
 	unsigned long count;
 
 	count = 0;
+	arg->result = 0;
 	do {
 		struct i915_request *rq = NULL;
-		int err, n;
+		int n;
 
-		err = 0;
-		for (n = 0; !err && n < ARRAY_SIZE(arg->ce); n++) {
+		for (n = 0; !arg->result && n < ARRAY_SIZE(arg->ce); n++) {
 			struct i915_request *prev = rq;
 
 			rq = i915_request_create(arg->ce[n]);
 			if (IS_ERR(rq)) {
 				i915_request_put(prev);
-				return PTR_ERR(rq);
+				arg->result = PTR_ERR(rq);
+				break;
 			}
 
 			i915_request_get(rq);
 			if (prev) {
-				err = i915_request_await_dma_fence(rq, &prev->fence);
+				arg->result =
+					i915_request_await_dma_fence(rq,
+								     &prev->fence);
 				i915_request_put(prev);
 			}
 
 			i915_request_add(rq);
 		}
+
+		if (IS_ERR_OR_NULL(rq))
+			break;
+
 		if (i915_request_wait(rq, 0, HZ) < 0)
-			err = -ETIME;
+			arg->result = -ETIME;
+
 		i915_request_put(rq);
-		if (err)
-			return err;
 
 		count++;
-	} while (!__igt_timeout(end_time, NULL));
+	} while (!arg->result && !__igt_timeout(end_time, NULL));
 
-	pr_info("%s: %lu switches (sync)\n", arg->ce[0]->engine->name, count);
-	return 0;
+	pr_info("%s: %lu switches (sync) <%d>\n",
+		arg->ce[0]->engine->name, count, arg->result);
 }
 
-static int __live_parallel_switchN(void *data)
+static void __live_parallel_switchN(struct kthread_work *work)
 {
-	struct parallel_switch *arg = data;
+	struct parallel_switch *arg =
+		container_of(work, typeof(*arg), work);
 	struct i915_request *rq = NULL;
 	IGT_TIMEOUT(end_time);
 	unsigned long count;
 	int n;
 
 	count = 0;
+	arg->result = 0;
 	do {
-		for (n = 0; n < ARRAY_SIZE(arg->ce); n++) {
+		for (n = 0; !arg->result && n < ARRAY_SIZE(arg->ce); n++) {
 			struct i915_request *prev = rq;
-			int err = 0;
 
 			rq = i915_request_create(arg->ce[n]);
 			if (IS_ERR(rq)) {
 				i915_request_put(prev);
-				return PTR_ERR(rq);
+				arg->result = PTR_ERR(rq);
+				break;
 			}
 
 			i915_request_get(rq);
 			if (prev) {
-				err = i915_request_await_dma_fence(rq, &prev->fence);
+				arg->result =
+					i915_request_await_dma_fence(rq,
+								     &prev->fence);
 				i915_request_put(prev);
 			}
 
 			i915_request_add(rq);
-			if (err) {
-				i915_request_put(rq);
-				return err;
-			}
 		}
 
 		count++;
-	} while (!__igt_timeout(end_time, NULL));
-	i915_request_put(rq);
+	} while (!arg->result && !__igt_timeout(end_time, NULL));
 
-	pr_info("%s: %lu switches (many)\n", arg->ce[0]->engine->name, count);
-	return 0;
+	if (!IS_ERR_OR_NULL(rq))
+		i915_request_put(rq);
+
+	pr_info("%s: %lu switches (many) <%d>\n",
+		arg->ce[0]->engine->name, count, arg->result);
 }
 
 static int live_parallel_switch(void *arg)
 {
 	struct drm_i915_private *i915 = arg;
-	static int (* const func[])(void *arg) = {
+	static void (* const func[])(struct kthread_work *) = {
 		__live_parallel_switch1,
 		__live_parallel_switchN,
 		NULL,
@@ -277,7 +288,7 @@ static int live_parallel_switch(void *arg)
 	struct parallel_switch *data = NULL;
 	struct i915_gem_engines *engines;
 	struct i915_gem_engines_iter it;
-	int (* const *fn)(void *arg);
+	void (* const *fn)(struct kthread_work *);
 	struct i915_gem_context *ctx;
 	struct intel_context *ce;
 	struct file *file;
@@ -335,8 +346,10 @@ static int live_parallel_switch(void *arg)
 				continue;
 
 			ce = intel_context_create(data[m].ce[0]->engine);
-			if (IS_ERR(ce))
+			if (IS_ERR(ce)) {
+				err = PTR_ERR(ce);
 				goto out;
+			}
 
 			err = intel_context_pin(ce);
 			if (err) {
@@ -348,9 +361,24 @@ static int live_parallel_switch(void *arg)
 		}
 	}
 
+	for (n = 0; n < count; n++) {
+		struct kthread_worker *worker;
+
+		if (!data[n].ce[0])
+			continue;
+
+		worker = kthread_create_worker(0, "igt/parallel:%s",
+					       data[n].ce[0]->engine->name);
+		if (IS_ERR(worker)) {
+			err = PTR_ERR(worker);
+			goto out;
+		}
+
+		data[n].worker = worker;
+	}
+
 	for (fn = func; !err && *fn; fn++) {
 		struct igt_live_test t;
-		int n;
 
 		err = igt_live_test_begin(&t, i915, __func__, "");
 		if (err)
@@ -360,34 +388,23 @@ static int live_parallel_switch(void *arg)
 			if (!data[n].ce[0])
 				continue;
 
-			data[n].tsk = kthread_run(*fn, &data[n],
-						  "igt/parallel:%s",
-						  data[n].ce[0]->engine->name);
-			if (IS_ERR(data[n].tsk)) {
-				err = PTR_ERR(data[n].tsk);
-				break;
-			}
-			get_task_struct(data[n].tsk);
+			data[n].result = 0;
+			kthread_init_work(&data[n].work, *fn);
+			kthread_queue_work(data[n].worker, &data[n].work);
 		}
 
-		yield(); /* start all threads before we kthread_stop() */
-
 		for (n = 0; n < count; n++) {
-			int status;
-
-			if (IS_ERR_OR_NULL(data[n].tsk))
-				continue;
-
-			status = kthread_stop(data[n].tsk);
-			if (status && !err)
-				err = status;
-
-			put_task_struct(data[n].tsk);
-			data[n].tsk = NULL;
+			if (data[n].ce[0]) {
+				kthread_flush_work(&data[n].work);
+				if (data[n].result && !err)
+					err = data[n].result;
+			}
 		}
 
-		if (igt_live_test_end(&t))
-			err = -EIO;
+		if (igt_live_test_end(&t)) {
+			err = err ?: -EIO;
+			break;
+		}
 	}
 
 out:
@@ -399,6 +416,9 @@ out:
 			intel_context_unpin(data[n].ce[m]);
 			intel_context_put(data[n].ce[m]);
 		}
+
+		if (data[n].worker)
+			kthread_destroy_worker(data[n].worker);
 	}
 	kfree(data);
 out_file:

sys/dev/pci/drm/i915/gt/selftest_execlists.c

@@ -1532,8 +1532,8 @@ static int live_busywait_preempt(void *arg)
 	struct drm_i915_gem_object *obj;
 	struct i915_vma *vma;
 	enum intel_engine_id id;
-	int err = -ENOMEM;
 	u32 *map;
+	int err;
 
 	/*
 	 * Verify that even without HAS_LOGICAL_RING_PREEMPTION, we can
@@ -1541,13 +1541,17 @@ static int live_busywait_preempt(void *arg)
 	 */
 
 	ctx_hi = kernel_context(gt->i915, NULL);
-	if (!ctx_hi)
-		return -ENOMEM;
+	if (IS_ERR(ctx_hi))
+		return PTR_ERR(ctx_hi);
+
 	ctx_hi->sched.priority = I915_CONTEXT_MAX_USER_PRIORITY;
 
 	ctx_lo = kernel_context(gt->i915, NULL);
-	if (!ctx_lo)
+	if (IS_ERR(ctx_lo)) {
+		err = PTR_ERR(ctx_lo);
 		goto err_ctx_hi;
+	}
+
 	ctx_lo->sched.priority = I915_CONTEXT_MIN_USER_PRIORITY;
 
 	obj = i915_gem_object_create_internal(gt->i915, PAGE_SIZE);
@@ -3475,12 +3479,14 @@ static int random_priority(struct rnd_state *rnd)
 
 struct preempt_smoke {
 	struct intel_gt *gt;
+	struct kthread_work work;
 	struct i915_gem_context **contexts;
 	struct intel_engine_cs *engine;
 	struct drm_i915_gem_object *batch;
 	unsigned int ncontext;
 	struct rnd_state prng;
 	unsigned long count;
+	int result;
 };
 
 static struct i915_gem_context *smoke_context(struct preempt_smoke *smoke)
@@ -3540,34 +3546,31 @@ unpin:
 	return err;
 }
 
-static int smoke_crescendo_thread(void *arg)
+static void smoke_crescendo_work(struct kthread_work *work)
 {
-	struct preempt_smoke *smoke = arg;
+	struct preempt_smoke *smoke = container_of(work, typeof(*smoke), work);
 	IGT_TIMEOUT(end_time);
 	unsigned long count;
 
 	count = 0;
 	do {
 		struct i915_gem_context *ctx = smoke_context(smoke);
-		int err;
 
-		err = smoke_submit(smoke,
-				   ctx, count % I915_PRIORITY_MAX,
-				   smoke->batch);
-		if (err)
-			return err;
+		smoke->result = smoke_submit(smoke, ctx,
+					     count % I915_PRIORITY_MAX,
+					     smoke->batch);
 
 		count++;
-	} while (count < smoke->ncontext && !__igt_timeout(end_time, NULL));
+	} while (!smoke->result && count < smoke->ncontext &&
+		 !__igt_timeout(end_time, NULL));
 
 	smoke->count = count;
-	return 0;
 }
 
 static int smoke_crescendo(struct preempt_smoke *smoke, unsigned int flags)
 #define BATCH BIT(0)
 {
-	struct task_struct *tsk[I915_NUM_ENGINES] = {};
+	struct kthread_worker *worker[I915_NUM_ENGINES] = {};
 	struct preempt_smoke *arg;
 	struct intel_engine_cs *engine;
 	enum intel_engine_id id;
@@ -3578,6 +3581,8 @@ static int smoke_crescendo(struct preempt_smoke *smoke, unsigned int flags)
 	if (!arg)
 		return -ENOMEM;
 
+	memset(arg, 0, I915_NUM_ENGINES * sizeof(*arg));
+
 	for_each_engine(engine, smoke->gt, id) {
 		arg[id] = *smoke;
 		arg[id].engine = engine;
@@ -3585,31 +3590,28 @@ static int smoke_crescendo(struct preempt_smoke *smoke, unsigned int flags)
 			arg[id].batch = NULL;
 		arg[id].count = 0;
 
-		tsk[id] = kthread_run(smoke_crescendo_thread, arg,
-				      "igt/smoke:%d", id);
-		if (IS_ERR(tsk[id])) {
-			err = PTR_ERR(tsk[id]);
+		worker[id] = kthread_create_worker(0, "igt/smoke:%d", id);
+		if (IS_ERR(worker[id])) {
+			err = PTR_ERR(worker[id]);
 			break;
 		}
-		get_task_struct(tsk[id]);
-	}
 
-	yield(); /* start all threads before we kthread_stop() */
+		kthread_init_work(&arg[id].work, smoke_crescendo_work);
+		kthread_queue_work(worker[id], &arg[id].work);
+	}
 
 	count = 0;
 	for_each_engine(engine, smoke->gt, id) {
-		int status;
-
-		if (IS_ERR_OR_NULL(tsk[id]))
+		if (IS_ERR_OR_NULL(worker[id]))
 			continue;
 
-		status = kthread_stop(tsk[id]);
-		if (status && !err)
-			err = status;
+		kthread_flush_work(&arg[id].work);
+		if (arg[id].result && !err)
+			err = arg[id].result;
 
 		count += arg[id].count;
 
-		put_task_struct(tsk[id]);
+		kthread_destroy_worker(worker[id]);
 	}
 
 	pr_info("Submitted %lu crescendo:%x requests across %d engines and %d contexts\n",

sys/dev/pci/drm/i915/gt/selftest_hangcheck.c

@@ -866,10 +866,13 @@ static int igt_reset_active_engine(void *arg)
 }
 
 struct active_engine {
-	struct task_struct *task;
+	struct kthread_worker *worker;
+	struct kthread_work work;
 	struct intel_engine_cs *engine;
 	unsigned long resets;
 	unsigned int flags;
+	bool stop;
+	int result;
 };
 
 #define TEST_ACTIVE	BIT(0)
@@ -900,10 +903,10 @@ static int active_request_put(struct i915_request *rq)
 	return err;
 }
 
-static int active_engine(void *data)
+static void active_engine(struct kthread_work *work)
 {
 	I915_RND_STATE(prng);
-	struct active_engine *arg = data;
+	struct active_engine *arg = container_of(work, typeof(*arg), work);
 	struct intel_engine_cs *engine = arg->engine;
 	struct i915_request *rq[8] = {};
 	struct intel_context *ce[ARRAY_SIZE(rq)];
@@ -913,16 +916,17 @@ static int active_engine(void *data)
 	for (count = 0; count < ARRAY_SIZE(ce); count++) {
 		ce[count] = intel_context_create(engine);
 		if (IS_ERR(ce[count])) {
-			err = PTR_ERR(ce[count]);
-			pr_err("[%s] Create context #%ld failed: %d!\n", engine->name, count, err);
+			arg->result = PTR_ERR(ce[count]);
+			pr_err("[%s] Create context #%ld failed: %d!\n",
+			       engine->name, count, arg->result);
 			while (--count)
 				intel_context_put(ce[count]);
-			return err;
+			return;
 		}
 	}
 
 	count = 0;
-	while (!kthread_should_stop()) {
+	while (!READ_ONCE(arg->stop)) {
 		unsigned int idx = count++ & (ARRAY_SIZE(rq) - 1);
 		struct i915_request *old = rq[idx];
 		struct i915_request *new;
@@ -967,7 +971,7 @@ static int active_engine(void *data)
 		intel_context_put(ce[count]);
 	}
 
-	return err;
+	arg->result = err;
 }
 
 static int __igt_reset_engines(struct intel_gt *gt,
@@ -1022,7 +1026,7 @@ static int __igt_reset_engines(struct intel_gt *gt,
 
 		memset(threads, 0, sizeof(*threads) * I915_NUM_ENGINES);
 		for_each_engine(other, gt, tmp) {
-			struct task_struct *tsk;
+			struct kthread_worker *worker;
 
 			threads[tmp].resets =
 				i915_reset_engine_count(global, other);
@@ -1036,19 +1040,21 @@ static int __igt_reset_engines(struct intel_gt *gt,
 			threads[tmp].engine = other;
 			threads[tmp].flags = flags;
 
-			tsk = kthread_run(active_engine, &threads[tmp],
-					  "igt/%s", other->name);
-			if (IS_ERR(tsk)) {
-				err = PTR_ERR(tsk);
-				pr_err("[%s] Thread spawn failed: %d!\n", engine->name, err);
+			worker = kthread_create_worker(0, "igt/%s",
+						       other->name);
+			if (IS_ERR(worker)) {
+				err = PTR_ERR(worker);
+				pr_err("[%s] Worker create failed: %d!\n",
+				       engine->name, err);
 				goto unwind;
 			}
 
-			threads[tmp].task = tsk;
-			get_task_struct(tsk);
-		}
+			threads[tmp].worker = worker;
 
-		yield(); /* start all threads before we begin */
+			kthread_init_work(&threads[tmp].work, active_engine);
+			kthread_queue_work(threads[tmp].worker,
+					   &threads[tmp].work);
+		}
 
 		st_engine_heartbeat_disable_no_pm(engine);
 		GEM_BUG_ON(test_and_set_bit(I915_RESET_ENGINE + id,
@@ -1197,17 +1203,20 @@ unwind:
 		for_each_engine(other, gt, tmp) {
 			int ret;
 
-			if (!threads[tmp].task)
+			if (!threads[tmp].worker)
 				continue;
 
-			ret = kthread_stop(threads[tmp].task);
+			WRITE_ONCE(threads[tmp].stop, true);
+			kthread_flush_work(&threads[tmp].work);
+			ret = READ_ONCE(threads[tmp].result);
 			if (ret) {
 				pr_err("kthread for other engine %s failed, err=%d\n",
 				       other->name, ret);
 				if (!err)
 					err = ret;
 			}
-			put_task_struct(threads[tmp].task);
+
+			kthread_destroy_worker(threads[tmp].worker);
 
 			/* GuC based resets are not logged per engine */
 			if (!using_guc) {

sys/dev/pci/drm/i915/selftests/i915_request.c

@@ -299,9 +299,18 @@ __live_request_alloc(struct intel_context *ce)
 	return intel_context_create_request(ce);
 }
 
-static int __igt_breadcrumbs_smoketest(void *arg)
+struct smoke_thread {
+	struct kthread_worker *worker;
+	struct kthread_work work;
+	struct smoketest *t;
+	bool stop;
+	int result;
+};
+
+static void __igt_breadcrumbs_smoketest(struct kthread_work *work)
 {
-	struct smoketest *t = arg;
+	struct smoke_thread *thread = container_of(work, typeof(*thread), work);
+	struct smoketest *t = thread->t;
 	const unsigned int max_batch = min(t->ncontexts, t->max_batch) - 1;
 	const unsigned int total = 4 * t->ncontexts + 1;
 	unsigned int num_waits = 0, num_fences = 0;
@@ -320,8 +329,10 @@ static int __igt_breadcrumbs_smoketest(void *arg)
 	 */
 
 	requests = kcalloc(total, sizeof(*requests), GFP_KERNEL);
-	if (!requests)
-		return -ENOMEM;
+	if (!requests) {
+		thread->result = -ENOMEM;
+		return;
+	}
 
 	order = i915_random_order(total, &prng);
 	if (!order) {
@@ -329,7 +340,7 @@ static int __igt_breadcrumbs_smoketest(void *arg)
 		goto out_requests;
 	}
 
-	while (!kthread_should_stop()) {
+	while (!READ_ONCE(thread->stop)) {
 		struct i915_sw_fence *submit, *wait;
 		unsigned int n, count;
 
@@ -437,7 +448,7 @@ static int __igt_breadcrumbs_smoketest(void *arg)
 	kfree(order);
 out_requests:
 	kfree(requests);
-	return err;
+	thread->result = err;
 }
 
 static int mock_breadcrumbs_smoketest(void *arg)
@@ -450,7 +461,7 @@ static int mock_breadcrumbs_smoketest(void *arg)
 		.request_alloc = __mock_request_alloc
 	};
 	unsigned int ncpus = num_online_cpus();
-	struct task_struct **threads;
+	struct smoke_thread *threads;
 	unsigned int n;
 	int ret = 0;
 
@@ -479,28 +490,37 @@ static int mock_breadcrumbs_smoketest(void *arg)
 	}
 
 	for (n = 0; n < ncpus; n++) {
-		threads[n] = kthread_run(__igt_breadcrumbs_smoketest,
-					 &t, "igt/%d", n);
-		if (IS_ERR(threads[n])) {
-			ret = PTR_ERR(threads[n]);
+		struct kthread_worker *worker;
+
+		worker = kthread_create_worker(0, "igt/%d", n);
+		if (IS_ERR(worker)) {
+			ret = PTR_ERR(worker);
 			ncpus = n;
 			break;
 		}
 
-		get_task_struct(threads[n]);
+		threads[n].worker = worker;
+		threads[n].t = &t;
+		threads[n].stop = false;
+		threads[n].result = 0;
+
+		kthread_init_work(&threads[n].work,
+				  __igt_breadcrumbs_smoketest);
+		kthread_queue_work(worker, &threads[n].work);
 	}
 
-	yield(); /* start all threads before we begin */
 	drm_msleep(jiffies_to_msecs(i915_selftest.timeout_jiffies));
 
 	for (n = 0; n < ncpus; n++) {
 		int err;
 
-		err = kthread_stop(threads[n]);
+		WRITE_ONCE(threads[n].stop, true);
+		kthread_flush_work(&threads[n].work);
+		err = READ_ONCE(threads[n].result);
 		if (err < 0 && !ret)
 			ret = err;
 
-		put_task_struct(threads[n]);
+		kthread_destroy_worker(threads[n].worker);
 	}
 	pr_info("Completed %lu waits for %lu fence across %d cpus\n",
 		atomic_long_read(&t.num_waits),
@@ -1419,9 +1439,18 @@ out_free:
 	return err;
 }
 
-static int __live_parallel_engine1(void *arg)
+struct parallel_thread {
+	struct kthread_worker *worker;
+	struct kthread_work work;
+	struct intel_engine_cs *engine;
+	int result;
+};
+
+static void __live_parallel_engine1(struct kthread_work *work)
 {
-	struct intel_engine_cs *engine = arg;
+	struct parallel_thread *thread =
+		container_of(work, typeof(*thread), work);
+	struct intel_engine_cs *engine = thread->engine;
 	IGT_TIMEOUT(end_time);
 	unsigned long count;
 	int err = 0;
@@ -1452,12 +1481,14 @@ static int __live_parallel_engine1(void *arg)
 	intel_engine_pm_put(engine);
 
 	pr_info("%s: %lu request + sync\n", engine->name, count);
-	return err;
+	thread->result = err;
 }
 
-static int __live_parallel_engineN(void *arg)
+static void __live_parallel_engineN(struct kthread_work *work)
 {
-	struct intel_engine_cs *engine = arg;
+	struct parallel_thread *thread =
+		container_of(work, typeof(*thread), work);
+	struct intel_engine_cs *engine = thread->engine;
 	IGT_TIMEOUT(end_time);
 	unsigned long count;
 	int err = 0;
@@ -1479,7 +1510,7 @@ static int __live_parallel_engineN(void *arg)
 	intel_engine_pm_put(engine);
 
 	pr_info("%s: %lu requests\n", engine->name, count);
-	return err;
+	thread->result = err;
 }
 
 static bool wake_all(struct drm_i915_private *i915)
@@ -1505,9 +1536,11 @@ static int wait_for_all(struct drm_i915_private *i915)
 	return -ETIME;
 }
 
-static int __live_parallel_spin(void *arg)
+static void __live_parallel_spin(struct kthread_work *work)
 {
-	struct intel_engine_cs *engine = arg;
+	struct parallel_thread *thread =
+		container_of(work, typeof(*thread), work);
+	struct intel_engine_cs *engine = thread->engine;
 	struct igt_spinner spin;
 	struct i915_request *rq;
 	int err = 0;
@@ -1520,7 +1553,8 @@ static int __live_parallel_spin(void *arg)
 
 	if (igt_spinner_init(&spin, engine->gt)) {
 		wake_all(engine->i915);
-		return -ENOMEM;
+		thread->result = -ENOMEM;
+		return;
 	}
 
 	intel_engine_pm_get(engine);
@@ -1553,22 +1587,22 @@ static int __live_parallel_spin(void *arg)
 
 out_spin:
 	igt_spinner_fini(&spin);
-	return err;
+	thread->result = err;
 }
 
 static int live_parallel_engines(void *arg)
 {
 	struct drm_i915_private *i915 = arg;
-	static int (* const func[])(void *arg) = {
+	static void (* const func[])(struct kthread_work *) = {
 		__live_parallel_engine1,
 		__live_parallel_engineN,
 		__live_parallel_spin,
 		NULL,
 	};
 	const unsigned int nengines = num_uabi_engines(i915);
+	struct parallel_thread *threads;
 	struct intel_engine_cs *engine;
-	int (* const *fn)(void *arg);
-	struct task_struct **tsk;
+	void (* const *fn)(struct kthread_work *);
 	int err = 0;
 
 	/*
@@ -1576,8 +1610,8 @@ static int live_parallel_engines(void *arg)
 	 * tests that we load up the system maximally.
 	 */
 
-	tsk = kcalloc(nengines, sizeof(*tsk), GFP_KERNEL);
-	if (!tsk)
+	threads = kcalloc(nengines, sizeof(*threads), GFP_KERNEL);
+	if (!threads)
 		return -ENOMEM;
 
 	for (fn = func; !err && *fn; fn++) {
@@ -1594,37 +1628,44 @@ static int live_parallel_engines(void *arg)
 
 		idx = 0;
 		for_each_uabi_engine(engine, i915) {
-			tsk[idx] = kthread_run(*fn, engine,
-					       "igt/parallel:%s",
-					       engine->name);
-			if (IS_ERR(tsk[idx])) {
-				err = PTR_ERR(tsk[idx]);
+			struct kthread_worker *worker;
+
+			worker = kthread_create_worker(0, "igt/parallel:%s",
+						       engine->name);
+			if (IS_ERR(worker)) {
+				err = PTR_ERR(worker);
 				break;
 			}
-			get_task_struct(tsk[idx++]);
-		}
 
-		yield(); /* start all threads before we kthread_stop() */
+			threads[idx].worker = worker;
+			threads[idx].result = 0;
+			threads[idx].engine = engine;
+
+			kthread_init_work(&threads[idx].work, *fn);
+			kthread_queue_work(worker, &threads[idx].work);
+			idx++;
+		}
 
 		idx = 0;
 		for_each_uabi_engine(engine, i915) {
 			int status;
 
-			if (IS_ERR(tsk[idx]))
+			if (!threads[idx].worker)
 				break;
 
-			status = kthread_stop(tsk[idx]);
+			kthread_flush_work(&threads[idx].work);
+			status = READ_ONCE(threads[idx].result);
 			if (status && !err)
 				err = status;
 
-			put_task_struct(tsk[idx++]);
+			kthread_destroy_worker(threads[idx++].worker);
 		}
 
 		if (igt_live_test_end(&t))
 			err = -EIO;
 	}
 
-	kfree(tsk);
+	kfree(threads);
 	return err;
 }
 
@@ -1672,7 +1713,7 @@ static int live_breadcrumbs_smoketest(void *arg)
 	const unsigned int ncpus = num_online_cpus();
 	unsigned long num_waits, num_fences;
 	struct intel_engine_cs *engine;
-	struct task_struct **threads;
+	struct smoke_thread *threads;
 	struct igt_live_test live;
 	intel_wakeref_t wakeref;
 	struct smoketest *smoke;
@@ -1746,23 +1787,26 @@ static int live_breadcrumbs_smoketest(void *arg)
 			 smoke[idx].max_batch, engine->name);
 
 		for (n = 0; n < ncpus; n++) {
-			struct task_struct *tsk;
+			unsigned int i = idx * ncpus + n;
+			struct kthread_worker *worker;
 
-			tsk = kthread_run(__igt_breadcrumbs_smoketest,
-					  &smoke[idx], "igt/%d.%d", idx, n);
-			if (IS_ERR(tsk)) {
-				ret = PTR_ERR(tsk);
+			worker = kthread_create_worker(0, "igt/%d.%d", idx, n);
+			if (IS_ERR(worker)) {
+				ret = PTR_ERR(worker);
 				goto out_flush;
 			}
 
-			get_task_struct(tsk);
-			threads[idx * ncpus + n] = tsk;
+			threads[i].worker = worker;
+			threads[i].t = &smoke[idx];
+
+			kthread_init_work(&threads[i].work,
+					  __igt_breadcrumbs_smoketest);
+			kthread_queue_work(worker, &threads[i].work);
 		}
 
 		idx++;
 	}
 
-	yield(); /* start all threads before we begin */
 	drm_msleep(jiffies_to_msecs(i915_selftest.timeout_jiffies));
 
 out_flush:
@@ -1771,17 +1815,19 @@ out_flush:
 	num_fences = 0;
 	for_each_uabi_engine(engine, i915) {
 		for (n = 0; n < ncpus; n++) {
-			struct task_struct *tsk = threads[idx * ncpus + n];
+			unsigned int i = idx * ncpus + n;
 			int err;
 
-			if (!tsk)
+			if (!threads[i].worker)
 				continue;
 
-			err = kthread_stop(tsk);
+			WRITE_ONCE(threads[i].stop, true);
+			kthread_flush_work(&threads[i].work);
+			err = READ_ONCE(threads[i].result);
 			if (err < 0 && !ret)
 				ret = err;
 
-			put_task_struct(tsk);
+			kthread_destroy_worker(threads[i].worker);
 		}
 
 		num_waits += atomic_long_read(&smoke[idx].num_waits);
@@ -2891,9 +2937,18 @@ out:
 	return err;
 }
 
-static int p_sync0(void *arg)
+struct p_thread {
+	struct perf_stats p;
+	struct kthread_worker *worker;
+	struct kthread_work work;
+	struct intel_engine_cs *engine;
+	int result;
+};
+
+static void p_sync0(struct kthread_work *work)
 {
-	struct perf_stats *p = arg;
+	struct p_thread *thread = container_of(work, typeof(*thread), work);
+	struct perf_stats *p = &thread->p;
 	struct intel_engine_cs *engine = p->engine;
 	struct intel_context *ce;
 	IGT_TIMEOUT(end_time);
@@ -2902,13 +2957,16 @@ static int p_sync0(void *arg)
 	int err = 0;
 
 	ce = intel_context_create(engine);
-	if (IS_ERR(ce))
-		return PTR_ERR(ce);
+	if (IS_ERR(ce)) {
+		thread->result = PTR_ERR(ce);
+		return;
+	}
 
 	err = intel_context_pin(ce);
 	if (err) {
 		intel_context_put(ce);
-		return err;
+		thread->result = err;
+		return;
 	}
 
 	if (intel_engine_supports_stats(engine)) {
@@ -2958,12 +3016,13 @@ static int p_sync0(void *arg)
 
 	intel_context_unpin(ce);
 	intel_context_put(ce);
-	return err;
+	thread->result = err;
 }
 
-static int p_sync1(void *arg)
+static void p_sync1(struct kthread_work *work)
 {
-	struct perf_stats *p = arg;
+	struct p_thread *thread = container_of(work, typeof(*thread), work);
+	struct perf_stats *p = &thread->p;
 	struct intel_engine_cs *engine = p->engine;
 	struct i915_request *prev = NULL;
 	struct intel_context *ce;
@@ -2973,13 +3032,16 @@ static int p_sync1(void *arg)
 	int err = 0;
 
 	ce = intel_context_create(engine);
-	if (IS_ERR(ce))
-		return PTR_ERR(ce);
+	if (IS_ERR(ce)) {
+		thread->result = PTR_ERR(ce);
+		return;
+	}
 
 	err = intel_context_pin(ce);
 	if (err) {
 		intel_context_put(ce);
-		return err;
+		thread->result = err;
+		return;
 	}
 
 	if (intel_engine_supports_stats(engine)) {
@@ -3031,12 +3093,13 @@ static int p_sync1(void *arg)
 
 	intel_context_unpin(ce);
 	intel_context_put(ce);
-	return err;
+	thread->result = err;
 }
 
-static int p_many(void *arg)
+static void p_many(struct kthread_work *work)
 {
-	struct perf_stats *p = arg;
+	struct p_thread *thread = container_of(work, typeof(*thread), work);
+	struct perf_stats *p = &thread->p;
 	struct intel_engine_cs *engine = p->engine;
 	struct intel_context *ce;
 	IGT_TIMEOUT(end_time);
@@ -3045,13 +3108,16 @@ static int p_many(void *arg)
 	bool busy;
 
 	ce = intel_context_create(engine);
-	if (IS_ERR(ce))
-		return PTR_ERR(ce);
+	if (IS_ERR(ce)) {
+		thread->result = PTR_ERR(ce);
+		return;
+	}
 
 	err = intel_context_pin(ce);
 	if (err) {
 		intel_context_put(ce);
-		return err;
+		thread->result = err;
+		return;
 	}
 
 	if (intel_engine_supports_stats(engine)) {
@@ -3092,26 +3158,23 @@ static int p_many(void *arg)
 
 	intel_context_unpin(ce);
 	intel_context_put(ce);
-	return err;
+	thread->result = err;
 }
 
 static int perf_parallel_engines(void *arg)
 {
 	struct drm_i915_private *i915 = arg;
-	static int (* const func[])(void *arg) = {
+	static void (* const func[])(struct kthread_work *) = {
 		p_sync0,
 		p_sync1,
 		p_many,
 		NULL,
 	};
 	const unsigned int nengines = num_uabi_engines(i915);
+	void (* const *fn)(struct kthread_work *);
 	struct intel_engine_cs *engine;
-	int (* const *fn)(void *arg);
 	struct pm_qos_request qos;
-	struct {
-		struct perf_stats p;
-		struct task_struct *tsk;
-	} *engines;
+	struct p_thread *engines;
 	int err = 0;
 
 	engines = kcalloc(nengines, sizeof(*engines), GFP_KERNEL);
@@ -3134,36 +3197,45 @@ static int perf_parallel_engines(void *arg)
 
 		idx = 0;
 		for_each_uabi_engine(engine, i915) {
+			struct kthread_worker *worker;
+
 			intel_engine_pm_get(engine);
 
 			memset(&engines[idx].p, 0, sizeof(engines[idx].p));
-			engines[idx].p.engine = engine;
 
-			engines[idx].tsk = kthread_run(*fn, &engines[idx].p,
-						       "igt:%s", engine->name);
-			if (IS_ERR(engines[idx].tsk)) {
-				err = PTR_ERR(engines[idx].tsk);
+			worker = kthread_create_worker(0, "igt:%s",
+						       engine->name);
+			if (IS_ERR(worker)) {
+				err = PTR_ERR(worker);
 				intel_engine_pm_put(engine);
 				break;
 			}
-			get_task_struct(engines[idx++].tsk);
-		}
+			engines[idx].worker = worker;
+			engines[idx].result = 0;
+			engines[idx].p.engine = engine;
+			engines[idx].engine = engine;
 
-		yield(); /* start all threads before we kthread_stop() */
+			kthread_init_work(&engines[idx].work, *fn);
+			kthread_queue_work(worker, &engines[idx].work);
+			idx++;
+		}
 
 		idx = 0;
 		for_each_uabi_engine(engine, i915) {
 			int status;
 
-			if (IS_ERR(engines[idx].tsk))
+			if (!engines[idx].worker)
 				break;
 
-			status = kthread_stop(engines[idx].tsk);
+			kthread_flush_work(&engines[idx].work);
+			status = READ_ONCE(engines[idx].result);
 			if (status && !err)
 				err = status;
 
 			intel_engine_pm_put(engine);
-			put_task_struct(engines[idx++].tsk);
+
+			kthread_destroy_worker(engines[idx].worker);
+			idx++;
 		}
 
 		if (igt_live_test_end(&t))