1 files changed, 228 insertions, 164 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 25e1f5ed7ead..cfa411c12072 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -37,6 +37,7 @@
 #include <linux/pci-p2pdma.h>
 
 #include <drm/drm_atomic_helper.h>
+#include <drm/drm_fb_helper.h>
 #include <drm/drm_probe_helper.h>
 #include <drm/amdgpu_drm.h>
 #include <linux/vgaarb.h>
@@ -1568,7 +1569,7 @@ static int amdgpu_device_check_arguments(struct amdgpu_device *adev)
  * @pdev: pci dev pointer
  * @state: vga_switcheroo state
  *
- * Callback for the switcheroo driver.  Suspends or resumes the
+ * Callback for the switcheroo driver.  Suspends or resumes
  * the asics before or after it is powered up using ACPI methods.
  */
 static void amdgpu_switcheroo_set_state(struct pci_dev *pdev,
@@ -1915,6 +1916,16 @@ static void amdgpu_device_enable_virtual_display(struct amdgpu_device *adev)
 	}
 }
 
+void amdgpu_device_set_sriov_virtual_display(struct amdgpu_device *adev)
+{
+	if (amdgpu_sriov_vf(adev) && !adev->enable_virtual_display) {
+		adev->mode_info.num_crtc = 1;
+		adev->enable_virtual_display = true;
+		DRM_INFO("virtual_display:%d, num_crtc:%d\n",
+			 adev->enable_virtual_display, adev->mode_info.num_crtc);
+	}
+}
+
 /**
  * amdgpu_device_parse_gpu_info_fw - parse gpu info firmware
  *
@@ -2397,7 +2408,7 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev)
 			adev->ip_blocks[i].status.hw = true;
 
 			/* right after GMC hw init, we create CSA */
-			if (amdgpu_mcbp || amdgpu_sriov_vf(adev)) {
+			if (amdgpu_mcbp) {
 				r = amdgpu_allocate_static_csa(adev, &adev->virt.csa_obj,
 								AMDGPU_GEM_DOMAIN_VRAM,
 								AMDGPU_CSA_SIZE);
@@ -2459,19 +2470,26 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev)
 	 */
 	if (adev->gmc.xgmi.num_physical_nodes > 1) {
 		if (amdgpu_xgmi_add_device(adev) == 0) {
-			struct amdgpu_hive_info *hive = amdgpu_get_xgmi_hive(adev);
+			if (!amdgpu_sriov_vf(adev)) {
+				struct amdgpu_hive_info *hive = amdgpu_get_xgmi_hive(adev);
+
+				if (WARN_ON(!hive)) {
+					r = -ENOENT;
+					goto init_failed;
+				}
+
+				if (!hive->reset_domain ||
+				    !amdgpu_reset_get_reset_domain(hive->reset_domain)) {
+					r = -ENOENT;
+					amdgpu_put_xgmi_hive(hive);
+					goto init_failed;
+				}
 
-			if (!hive->reset_domain ||
-			    !amdgpu_reset_get_reset_domain(hive->reset_domain)) {
-				r = -ENOENT;
+				/* Drop the early temporary reset domain we created for device */
+				amdgpu_reset_put_reset_domain(adev->reset_domain);
+				adev->reset_domain = hive->reset_domain;
 				amdgpu_put_xgmi_hive(hive);
-				goto init_failed;
 			}
-
-			/* Drop the early temporary reset domain we created for device */
-			amdgpu_reset_put_reset_domain(adev->reset_domain);
-			adev->reset_domain = hive->reset_domain;
-			amdgpu_put_xgmi_hive(hive);
 		}
 	}
 
@@ -2926,6 +2944,14 @@ static int amdgpu_device_ip_suspend_phase1(struct amdgpu_device *adev)
 	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE);
 	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE);
 
+	/*
+	 * Per PMFW team's suggestion, driver needs to handle gfxoff
+	 * and df cstate features disablement for gpu reset(e.g. Mode1Reset)
+	 * scenario. Add the missing df cstate disablement here.
+	 */
+	if (amdgpu_dpm_set_df_cstate(adev, DF_CSTATE_DISALLOW))
+		dev_warn(adev->dev, "Failed to disallow df cstate");
+
 	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
 		if (!adev->ip_blocks[i].status.valid)
 			continue;
@@ -3200,6 +3226,15 @@ static int amdgpu_device_ip_resume_phase2(struct amdgpu_device *adev)
 			return r;
 		}
 		adev->ip_blocks[i].status.hw = true;
+
+		if (adev->in_s0ix && adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_SMC) {
+			/* disable gfxoff for IP resume. The gfxoff will be re-enabled in
+			 * amdgpu_device_resume() after IP resume.
+			 */
+			amdgpu_gfx_off_ctrl(adev, false);
+			DRM_DEBUG("will disable gfxoff for re-initializing other blocks\n");
+		}
+
 	}
 
 	return 0;
@@ -3328,8 +3363,7 @@ bool amdgpu_device_asic_has_dc_support(enum amd_asic_type asic_type)
  */
 bool amdgpu_device_has_dc_support(struct amdgpu_device *adev)
 {
-	if (amdgpu_sriov_vf(adev) ||
-	    adev->enable_virtual_display ||
+	if (adev->enable_virtual_display ||
 	    (adev->harvest_ip_mask & AMD_HARVEST_IP_DMU_MASK))
 		return false;
 
@@ -3510,6 +3544,7 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	adev->gmc.gart_size = 512 * 1024 * 1024;
 	adev->accel_working = false;
 	adev->num_rings = 0;
+	RCU_INIT_POINTER(adev->gang_submit, dma_fence_get_stub());
 	adev->mman.buffer_funcs = NULL;
 	adev->mman.buffer_funcs_ring = NULL;
 	adev->vm_manager.vm_pte_funcs = NULL;
@@ -3588,6 +3623,8 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	INIT_WORK(&adev->xgmi_reset_work, amdgpu_device_xgmi_reset_func);
 
 	adev->gfx.gfx_off_req_count = 1;
+	adev->gfx.gfx_off_residency = 0;
+	adev->gfx.gfx_off_entrycount = 0;
 	adev->pm.ac_power = power_supply_is_system_supplied() > 0;
 
 	atomic_set(&adev->throttling_logging_enabled, 1);
@@ -3976,8 +4013,7 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev)
 
 	amdgpu_gart_dummy_page_fini(adev);
 
-	if (drm_dev_is_unplugged(adev_to_drm(adev)))
-		amdgpu_device_unmap_mmio(adev);
+	amdgpu_device_unmap_mmio(adev);
 
 }
 
@@ -3990,6 +4026,7 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev)
 	release_firmware(adev->firmware.gpu_info_fw);
 	adev->firmware.gpu_info_fw = NULL;
 	adev->accel_working = false;
+	dma_fence_put(rcu_dereference_protected(adev->gang_submit, true));
 
 	amdgpu_reset_fini(adev);
 
@@ -4038,15 +4075,18 @@ void amdgpu_device_fini_sw(struct amdgpu_device *adev)
  * at suspend time.
  *
  */
-static void amdgpu_device_evict_resources(struct amdgpu_device *adev)
+static int amdgpu_device_evict_resources(struct amdgpu_device *adev)
 {
+	int ret;
+
 	/* No need to evict vram on APUs for suspend to ram or s2idle */
 	if ((adev->in_s3 || adev->in_s0ix) && (adev->flags & AMD_IS_APU))
-		return;
+		return 0;
 
-	if (amdgpu_ttm_evict_resources(adev, TTM_PL_VRAM))
+	ret = amdgpu_ttm_evict_resources(adev, TTM_PL_VRAM);
+	if (ret)
 		DRM_WARN("evicting device resources failed\n");
-
+	return ret;
 }
 
 /*
@@ -4096,7 +4136,9 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon)
 	if (!adev->in_s0ix)
 		amdgpu_amdkfd_suspend(adev, adev->in_runpm);
 
-	amdgpu_device_evict_resources(adev);
+	r = amdgpu_device_evict_resources(adev);
+	if (r)
+		return r;
 
 	amdgpu_fence_driver_hw_fini(adev);
 
@@ -4144,21 +4186,15 @@ int amdgpu_device_resume(struct drm_device *dev, bool fbcon)
 
 	r = amdgpu_device_ip_resume(adev);
 
-	/* no matter what r is, always need to properly release full GPU */
-	if (amdgpu_sriov_vf(adev)) {
-		amdgpu_virt_init_data_exchange(adev);
-		amdgpu_virt_release_full_gpu(adev, true);
-	}
-
 	if (r) {
 		dev_err(adev->dev, "amdgpu_device_ip_resume failed (%d).\n", r);
-		return r;
+		goto exit;
 	}
 	amdgpu_fence_driver_hw_init(adev);
 
 	r = amdgpu_device_ip_late_init(adev);
 	if (r)
-		return r;
+		goto exit;
 
 	queue_delayed_work(system_wq, &adev->delayed_init_work,
 			   msecs_to_jiffies(AMDGPU_RESUME_MS));
@@ -4166,12 +4202,28 @@ int amdgpu_device_resume(struct drm_device *dev, bool fbcon)
 	if (!adev->in_s0ix) {
 		r = amdgpu_amdkfd_resume(adev, adev->in_runpm);
 		if (r)
-			return r;
+			goto exit;
+	}
+
+exit:
+	if (amdgpu_sriov_vf(adev)) {
+		amdgpu_virt_init_data_exchange(adev);
+		amdgpu_virt_release_full_gpu(adev, true);
 	}
 
+	if (r)
+		return r;
+
 	/* Make sure IB tests flushed */
 	flush_delayed_work(&adev->delayed_init_work);
 
+	if (adev->in_s0ix) {
+		/* re-enable gfxoff after IP resume. This re-enables gfxoff after
+		 * it was disabled for IP resume in amdgpu_device_ip_resume_phase2().
+		 */
+		amdgpu_gfx_off_ctrl(adev, true);
+		DRM_DEBUG("will enable gfxoff for the mission mode\n");
+	}
 	if (fbcon)
 		drm_fb_helper_set_suspend_unlocked(adev_to_drm(adev)->fb_helper, false);
 
@@ -4179,25 +4231,27 @@ int amdgpu_device_resume(struct drm_device *dev, bool fbcon)
 
 	amdgpu_ras_resume(adev);
 
-	/*
-	 * Most of the connector probing functions try to acquire runtime pm
-	 * refs to ensure that the GPU is powered on when connector polling is
-	 * performed. Since we're calling this from a runtime PM callback,
-	 * trying to acquire rpm refs will cause us to deadlock.
-	 *
-	 * Since we're guaranteed to be holding the rpm lock, it's safe to
-	 * temporarily disable the rpm helpers so this doesn't deadlock us.
-	 */
+	if (adev->mode_info.num_crtc) {
+		/*
+		 * Most of the connector probing functions try to acquire runtime pm
+		 * refs to ensure that the GPU is powered on when connector polling is
+		 * performed. Since we're calling this from a runtime PM callback,
+		 * trying to acquire rpm refs will cause us to deadlock.
+		 *
+		 * Since we're guaranteed to be holding the rpm lock, it's safe to
+		 * temporarily disable the rpm helpers so this doesn't deadlock us.
+		 */
 #ifdef CONFIG_PM
-	dev->dev->power.disable_depth++;
+		dev->dev->power.disable_depth++;
 #endif
-	if (!amdgpu_device_has_dc_support(adev))
-		drm_helper_hpd_irq_event(dev);
-	else
-		drm_kms_helper_hotplug_event(dev);
+		if (!adev->dc_enabled)
+			drm_helper_hpd_irq_event(dev);
+		else
+			drm_kms_helper_hotplug_event(dev);
 #ifdef CONFIG_PM
-	dev->dev->power.disable_depth--;
+		dev->dev->power.disable_depth--;
 #endif
+	}
 	adev->in_suspend = false;
 
 	if (amdgpu_acpi_smart_shift_update(dev, AMDGPU_SS_DEV_D0))
@@ -4542,14 +4596,19 @@ bool amdgpu_device_has_job_running(struct amdgpu_device *adev)
  */
 bool amdgpu_device_should_recover_gpu(struct amdgpu_device *adev)
 {
-	if (!amdgpu_device_ip_check_soft_reset(adev)) {
-		dev_info(adev->dev, "Timeout, but no hardware hang detected.\n");
-		return false;
-	}
 
 	if (amdgpu_gpu_recovery == 0)
 		goto disabled;
 
+	/* Skip soft reset check in fatal error mode */
+	if (!amdgpu_ras_is_poison_mode_supported(adev))
+		return true;
+
+	if (!amdgpu_device_ip_check_soft_reset(adev)) {
+		dev_info(adev->dev,"Timeout, but no hardware hang detected.\n");
+		return false;
+	}
+
 	if (amdgpu_sriov_vf(adev))
 		return true;
 
@@ -4674,7 +4733,7 @@ int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev,
 		if (!need_full_reset)
 			need_full_reset = amdgpu_device_ip_need_full_reset(adev);
 
-		if (!need_full_reset) {
+		if (!need_full_reset && amdgpu_gpu_recovery) {
 			amdgpu_device_ip_pre_soft_reset(adev);
 			r = amdgpu_device_ip_soft_reset(adev);
 			amdgpu_device_ip_post_soft_reset(adev);
@@ -4770,6 +4829,7 @@ int amdgpu_do_asic_reset(struct list_head *device_list_handle,
 	struct amdgpu_device *tmp_adev = NULL;
 	bool need_full_reset, skip_hw_reset, vram_lost = false;
 	int r = 0;
+	bool gpu_reset_for_dev_remove = 0;
 
 	/* Try reset handler method first */
 	tmp_adev = list_first_entry(device_list_handle, struct amdgpu_device,
@@ -4789,6 +4849,10 @@ int amdgpu_do_asic_reset(struct list_head *device_list_handle,
 		test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags);
 	skip_hw_reset = test_bit(AMDGPU_SKIP_HW_RESET, &reset_context->flags);
 
+	gpu_reset_for_dev_remove =
+		test_bit(AMDGPU_RESET_FOR_DEVICE_REMOVE, &reset_context->flags) &&
+			test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags);
+
 	/*
 	 * ASIC reset has to be done on all XGMI hive nodes ASAP
 	 * to allow proper links negotiation in FW (within 1 sec)
@@ -4833,6 +4897,18 @@ int amdgpu_do_asic_reset(struct list_head *device_list_handle,
 		amdgpu_ras_intr_cleared();
 	}
 
+	/* Since the mode1 reset affects base ip blocks, the
+	 * phase1 ip blocks need to be resumed. Otherwise there
+	 * will be a BIOS signature error and the psp bootloader
+	 * can't load kdb on the next amdgpu install.
+	 */
+	if (gpu_reset_for_dev_remove) {
+		list_for_each_entry(tmp_adev, device_list_handle, reset_list)
+			amdgpu_device_ip_resume_phase1(tmp_adev);
+
+		goto end;
+	}
+
 	list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
 		if (need_full_reset) {
 			/* post card */
@@ -4975,6 +5051,8 @@ static void amdgpu_device_resume_display_audio(struct amdgpu_device *adev)
 		pm_runtime_enable(&(p->dev));
 		pm_runtime_resume(&(p->dev));
 	}
+
+	pci_dev_put(p);
 }
 
 static int amdgpu_device_suspend_display_audio(struct amdgpu_device *adev)
@@ -5013,6 +5091,7 @@ static int amdgpu_device_suspend_display_audio(struct amdgpu_device *adev)
 
 		if (expires < ktime_get_mono_fast_ns()) {
 			dev_warn(adev->dev, "failed to suspend display audio\n");
+			pci_dev_put(p);
 			/* TODO: abort the succeeding gpu reset? */
 			return -ETIMEDOUT;
 		}
@@ -5020,96 +5099,10 @@ static int amdgpu_device_suspend_display_audio(struct amdgpu_device *adev)
 
 	pm_runtime_disable(&(p->dev));
 
+	pci_dev_put(p);
 	return 0;
 }
 
-static void amdgpu_device_recheck_guilty_jobs(
-	struct amdgpu_device *adev, struct list_head *device_list_handle,
-	struct amdgpu_reset_context *reset_context)
-{
-	int i, r = 0;
-
-	for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
-		struct amdgpu_ring *ring = adev->rings[i];
-		int ret = 0;
-		struct drm_sched_job *s_job;
-
-		if (!ring || !ring->sched.thread)
-			continue;
-
-		s_job = list_first_entry_or_null(&ring->sched.pending_list,
-				struct drm_sched_job, list);
-		if (s_job == NULL)
-			continue;
-
-		/* clear job's guilty and depend the folowing step to decide the real one */
-		drm_sched_reset_karma(s_job);
-		drm_sched_resubmit_jobs_ext(&ring->sched, 1);
-
-		if (!s_job->s_fence->parent) {
-			DRM_WARN("Failed to get a HW fence for job!");
-			continue;
-		}
-
-		ret = dma_fence_wait_timeout(s_job->s_fence->parent, false, ring->sched.timeout);
-		if (ret == 0) { /* timeout */
-			DRM_ERROR("Found the real bad job! ring:%s, job_id:%llx\n",
-						ring->sched.name, s_job->id);
-
-
-			amdgpu_fence_driver_isr_toggle(adev, true);
-
-			/* Clear this failed job from fence array */
-			amdgpu_fence_driver_clear_job_fences(ring);
-
-			amdgpu_fence_driver_isr_toggle(adev, false);
-
-			/* Since the job won't signal and we go for
-			 * another resubmit drop this parent pointer
-			 */
-			dma_fence_put(s_job->s_fence->parent);
-			s_job->s_fence->parent = NULL;
-
-			/* set guilty */
-			drm_sched_increase_karma(s_job);
-retry:
-			/* do hw reset */
-			if (amdgpu_sriov_vf(adev)) {
-				amdgpu_virt_fini_data_exchange(adev);
-				r = amdgpu_device_reset_sriov(adev, false);
-				if (r)
-					adev->asic_reset_res = r;
-			} else {
-				clear_bit(AMDGPU_SKIP_HW_RESET,
-					  &reset_context->flags);
-				r = amdgpu_do_asic_reset(device_list_handle,
-							 reset_context);
-				if (r && r == -EAGAIN)
-					goto retry;
-			}
-
-			/*
-			 * add reset counter so that the following
-			 * resubmitted job could flush vmid
-			 */
-			atomic_inc(&adev->gpu_reset_counter);
-			continue;
-		}
-
-		/* got the hw fence, signal finished fence */
-		atomic_dec(ring->sched.score);
-		dma_fence_get(&s_job->s_fence->finished);
-		dma_fence_signal(&s_job->s_fence->finished);
-		dma_fence_put(&s_job->s_fence->finished);
-
-		/* remove node from list and free the job */
-		spin_lock(&ring->sched.job_list_lock);
-		list_del_init(&s_job->list);
-		spin_unlock(&ring->sched.job_list_lock);
-		ring->sched.ops->free_job(s_job);
-	}
-}
-
 static inline void amdgpu_device_stop_pending_resets(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
@@ -5130,7 +5123,6 @@ static inline void amdgpu_device_stop_pending_resets(struct amdgpu_device *adev)
 
 }
 
-
 /**
  * amdgpu_device_gpu_recover - reset the asic and recover scheduler
  *
@@ -5153,7 +5145,11 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 	int i, r = 0;
 	bool need_emergency_restart = false;
 	bool audio_suspended = false;
-	int tmp_vram_lost_counter;
+	bool gpu_reset_for_dev_remove = false;
+
+	gpu_reset_for_dev_remove =
+			test_bit(AMDGPU_RESET_FOR_DEVICE_REMOVE, &reset_context->flags) &&
+				test_bit(AMDGPU_NEED_FULL_RESET, &reset_context->flags);
 
 	/*
 	 * Special case: RAS triggered and full reset isn't supported
@@ -5188,8 +5184,11 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 	 */
 	INIT_LIST_HEAD(&device_list);
 	if (!amdgpu_sriov_vf(adev) && (adev->gmc.xgmi.num_physical_nodes > 1)) {
-		list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head)
+		list_for_each_entry(tmp_adev, &hive->device_list, gmc.xgmi.head) {
 			list_add_tail(&tmp_adev->reset_list, &device_list);
+			if (gpu_reset_for_dev_remove && adev->shutdown)
+				tmp_adev->shutdown = true;
+		}
 		if (!list_is_first(&adev->reset_list, &device_list))
 			list_rotate_to_front(&adev->reset_list, &device_list);
 		device_list_handle = &device_list;
@@ -5272,6 +5271,10 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 
 retry:	/* Rest of adevs pre asic reset from XGMI hive. */
 	list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
+		if (gpu_reset_for_dev_remove) {
+			/* Workaroud for ASICs need to disable SMC first */
+			amdgpu_device_smu_fini_early(tmp_adev);
+		}
 		r = amdgpu_device_pre_asic_reset(tmp_adev, reset_context);
 		/*TODO Should we stop ?*/
 		if (r) {
@@ -5287,7 +5290,6 @@ retry:	/* Rest of adevs pre asic reset from XGMI hive. */
 		amdgpu_device_stop_pending_resets(tmp_adev);
 	}
 
-	tmp_vram_lost_counter = atomic_read(&((adev)->vram_lost_counter));
 	/* Actual ASIC resets if needed.*/
 	/* Host driver will handle XGMI hive reset for SRIOV */
 	if (amdgpu_sriov_vf(adev)) {
@@ -5302,6 +5304,9 @@ retry:	/* Rest of adevs pre asic reset from XGMI hive. */
 		r = amdgpu_do_asic_reset(device_list_handle, reset_context);
 		if (r && r == -EAGAIN)
 			goto retry;
+
+		if (!r && gpu_reset_for_dev_remove)
+			goto recover_end;
 	}
 
 skip_hw_reset:
@@ -5309,32 +5314,16 @@ skip_hw_reset:
 	/* Post ASIC reset for all devs .*/
 	list_for_each_entry(tmp_adev, device_list_handle, reset_list) {
 
-		/*
-		 * Sometimes a later bad compute job can block a good gfx job as gfx
-		 * and compute ring share internal GC HW mutually. We add an additional
-		 * guilty jobs recheck step to find the real guilty job, it synchronously
-		 * submits and pends for the first job being signaled. If it gets timeout,
-		 * we identify it as a real guilty job.
-		 */
-		if (amdgpu_gpu_recovery == 2 &&
-			!(tmp_vram_lost_counter < atomic_read(&adev->vram_lost_counter)))
-			amdgpu_device_recheck_guilty_jobs(
-				tmp_adev, device_list_handle, reset_context);
-
 		for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
 			struct amdgpu_ring *ring = tmp_adev->rings[i];
 
 			if (!ring || !ring->sched.thread)
 				continue;
 
-			/* No point to resubmit jobs if we didn't HW reset*/
-			if (!tmp_adev->asic_reset_res && !job_signaled)
-				drm_sched_resubmit_jobs(&ring->sched);
-
-			drm_sched_start(&ring->sched, !tmp_adev->asic_reset_res);
+			drm_sched_start(&ring->sched, true);
 		}
 
-		if (adev->enable_mes)
+		if (adev->enable_mes && adev->ip_versions[GC_HWIP][0] != IP_VERSION(11, 0, 3))
 			amdgpu_mes_self_test(tmp_adev);
 
 		if (!drm_drv_uses_atomic_modeset(adev_to_drm(tmp_adev)) && !job_signaled) {
@@ -5373,8 +5362,11 @@ skip_sched_resume:
 			amdgpu_device_resume_display_audio(tmp_adev);
 
 		amdgpu_device_unset_mp1_state(tmp_adev);
+
+		amdgpu_ras_set_error_query_ready(tmp_adev, true);
 	}
 
+recover_end:
 	tmp_adev = list_first_entry(device_list_handle, struct amdgpu_device,
 					    reset_list);
 	amdgpu_device_unlock_reset_domain(tmp_adev->reset_domain);
@@ -5557,9 +5549,9 @@ bool amdgpu_device_is_peer_accessible(struct amdgpu_device *adev,
 		~*peer_adev->dev->dma_mask : ~((1ULL << 32) - 1);
 	resource_size_t aper_limit =
 		adev->gmc.aper_base + adev->gmc.aper_size - 1;
-	bool p2p_access = !adev->gmc.xgmi.connected_to_cpu &&
-			  !(pci_p2pdma_distance_many(adev->pdev,
-					&peer_adev->dev, 1, true) < 0);
+	bool p2p_access =
+		!adev->gmc.xgmi.connected_to_cpu &&
+		!(pci_p2pdma_distance(adev->pdev, peer_adev->dev, false) < 0);
 
 	return pcie_p2p && p2p_access && (adev->gmc.visible_vram_size &&
 		adev->gmc.real_vram_size == adev->gmc.visible_vram_size &&
@@ -5783,8 +5775,6 @@ void amdgpu_pci_resume(struct pci_dev *pdev)
 		if (!ring || !ring->sched.thread)
 			continue;
 
-
-		drm_sched_resubmit_jobs(&ring->sched);
 		drm_sched_start(&ring->sched, true);
 	}
 
@@ -5942,3 +5932,77 @@ void amdgpu_device_pcie_port_wreg(struct amdgpu_device *adev,
 	(void)RREG32(data);
 	spin_unlock_irqrestore(&adev->pcie_idx_lock, flags);
 }
+
+/**
+ * amdgpu_device_switch_gang - switch to a new gang
+ * @adev: amdgpu_device pointer
+ * @gang: the gang to switch to
+ *
+ * Try to switch to a new gang.
+ * Returns: NULL if we switched to the new gang or a reference to the current
+ * gang leader.
+ */
+struct dma_fence *amdgpu_device_switch_gang(struct amdgpu_device *adev,
+					    struct dma_fence *gang)
+{
+	struct dma_fence *old = NULL;
+
+	do {
+		dma_fence_put(old);
+		rcu_read_lock();
+		old = dma_fence_get_rcu_safe(&adev->gang_submit);
+		rcu_read_unlock();
+
+		if (old == gang)
+			break;
+
+		if (!dma_fence_is_signaled(old))
+			return old;
+
+	} while (cmpxchg((struct dma_fence __force **)&adev->gang_submit,
+			 old, gang) != old);
+
+	dma_fence_put(old);
+	return NULL;
+}
+
+bool amdgpu_device_has_display_hardware(struct amdgpu_device *adev)
+{
+	switch (adev->asic_type) {
+#ifdef CONFIG_DRM_AMDGPU_SI
+	case CHIP_HAINAN:
+#endif
+	case CHIP_TOPAZ:
+		/* chips with no display hardware */
+		return false;
+#ifdef CONFIG_DRM_AMDGPU_SI
+	case CHIP_TAHITI:
+	case CHIP_PITCAIRN:
+	case CHIP_VERDE:
+	case CHIP_OLAND:
+#endif
+#ifdef CONFIG_DRM_AMDGPU_CIK
+	case CHIP_BONAIRE:
+	case CHIP_HAWAII:
+	case CHIP_KAVERI:
+	case CHIP_KABINI:
+	case CHIP_MULLINS:
+#endif
+	case CHIP_TONGA:
+	case CHIP_FIJI:
+	case CHIP_POLARIS10:
+	case CHIP_POLARIS11:
+	case CHIP_POLARIS12:
+	case CHIP_VEGAM:
+	case CHIP_CARRIZO:
+	case CHIP_STONEY:
+		/* chips with display hardware */
+		return true;
+	default:
+		/* IP discovery */
+		if (!adev->ip_versions[DCE_HWIP][0] ||
+		    (adev->harvest_ip_mask & AMD_HARVEST_IP_DMU_MASK))
+			return false;
+		return true;
+	}
+}