1 files changed, 284 insertions, 141 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 559dc24ef436..545beebcf43e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -64,9 +64,11 @@
 #include "amdgpu_xgmi.h"
 #include "amdgpu_ras.h"
 #include "amdgpu_pmu.h"
+#include "amdgpu_fru_eeprom.h"
 
 #include <linux/suspend.h>
 #include <drm/task_barrier.h>
+#include <linux/pm_runtime.h>
 
 MODULE_FIRMWARE("amdgpu/vega10_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/vega12_gpu_info.bin");
@@ -138,6 +140,72 @@ static DEVICE_ATTR(pcie_replay_count, S_IRUGO,
 static void amdgpu_device_get_pcie_info(struct amdgpu_device *adev);
 
 /**
+ * DOC: product_name
+ *
+ * The amdgpu driver provides a sysfs API for reporting the product name
+ * for the device
+ * The file serial_number is used for this and returns the product name
+ * as returned from the FRU.
+ * NOTE: This is only available for certain server cards
+ */
+
+static ssize_t amdgpu_device_get_product_name(struct device *dev,
+		struct device_attribute *attr, char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = ddev->dev_private;
+
+	return snprintf(buf, PAGE_SIZE, "%s\n", adev->product_name);
+}
+
+static DEVICE_ATTR(product_name, S_IRUGO,
+		amdgpu_device_get_product_name, NULL);
+
+/**
+ * DOC: product_number
+ *
+ * The amdgpu driver provides a sysfs API for reporting the part number
+ * for the device
+ * The file serial_number is used for this and returns the part number
+ * as returned from the FRU.
+ * NOTE: This is only available for certain server cards
+ */
+
+static ssize_t amdgpu_device_get_product_number(struct device *dev,
+		struct device_attribute *attr, char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = ddev->dev_private;
+
+	return snprintf(buf, PAGE_SIZE, "%s\n", adev->product_number);
+}
+
+static DEVICE_ATTR(product_number, S_IRUGO,
+		amdgpu_device_get_product_number, NULL);
+
+/**
+ * DOC: serial_number
+ *
+ * The amdgpu driver provides a sysfs API for reporting the serial number
+ * for the device
+ * The file serial_number is used for this and returns the serial number
+ * as returned from the FRU.
+ * NOTE: This is only available for certain server cards
+ */
+
+static ssize_t amdgpu_device_get_serial_number(struct device *dev,
+		struct device_attribute *attr, char *buf)
+{
+	struct drm_device *ddev = dev_get_drvdata(dev);
+	struct amdgpu_device *adev = ddev->dev_private;
+
+	return snprintf(buf, PAGE_SIZE, "%s\n", adev->serial);
+}
+
+static DEVICE_ATTR(serial_number, S_IRUGO,
+		amdgpu_device_get_serial_number, NULL);
+
+/**
  * amdgpu_device_supports_boco - Is the device a dGPU with HG/PX power control
  *
  * @dev: drm_device pointer
@@ -231,10 +299,10 @@ void amdgpu_device_vram_access(struct amdgpu_device *adev, loff_t pos,
 }
 
 /*
- * MMIO register access helper functions.
+ * device register access helper functions.
  */
 /**
- * amdgpu_mm_rreg - read a memory mapped IO register
+ * amdgpu_device_rreg - read a register
  *
  * @adev: amdgpu_device pointer
  * @reg: dword aligned register offset
@@ -242,25 +310,19 @@ void amdgpu_device_vram_access(struct amdgpu_device *adev, loff_t pos,
  *
  * Returns the 32 bit value from the offset specified.
  */
-uint32_t amdgpu_mm_rreg(struct amdgpu_device *adev, uint32_t reg,
-			uint32_t acc_flags)
+uint32_t amdgpu_device_rreg(struct amdgpu_device *adev, uint32_t reg,
+			    uint32_t acc_flags)
 {
 	uint32_t ret;
 
-	if ((acc_flags & AMDGPU_REGS_KIQ) || (!(acc_flags & AMDGPU_REGS_NO_KIQ) && amdgpu_sriov_runtime(adev)))
+	if (!(acc_flags & AMDGPU_REGS_NO_KIQ) && amdgpu_sriov_runtime(adev))
 		return amdgpu_kiq_rreg(adev, reg);
 
-	if ((reg * 4) < adev->rmmio_size && !(acc_flags & AMDGPU_REGS_IDX))
+	if ((reg * 4) < adev->rmmio_size)
 		ret = readl(((void __iomem *)adev->rmmio) + (reg * 4));
-	else {
-		unsigned long flags;
-
-		spin_lock_irqsave(&adev->mmio_idx_lock, flags);
-		writel((reg * 4), ((void __iomem *)adev->rmmio) + (mmMM_INDEX * 4));
-		ret = readl(((void __iomem *)adev->rmmio) + (mmMM_DATA * 4));
-		spin_unlock_irqrestore(&adev->mmio_idx_lock, flags);
-	}
-	trace_amdgpu_mm_rreg(adev->pdev->device, reg, ret);
+	else
+		ret = adev->pcie_rreg(adev, (reg * 4));
+	trace_amdgpu_device_rreg(adev->pdev->device, reg, ret);
 	return ret;
 }
 
@@ -306,28 +368,19 @@ void amdgpu_mm_wreg8(struct amdgpu_device *adev, uint32_t offset, uint8_t value)
 		BUG();
 }
 
-void static inline amdgpu_mm_wreg_mmio(struct amdgpu_device *adev, uint32_t reg, uint32_t v, uint32_t acc_flags)
+void static inline amdgpu_device_wreg_no_kiq(struct amdgpu_device *adev, uint32_t reg,
+					     uint32_t v, uint32_t acc_flags)
 {
-	trace_amdgpu_mm_wreg(adev->pdev->device, reg, v);
+	trace_amdgpu_device_wreg(adev->pdev->device, reg, v);
 
-	if ((reg * 4) < adev->rmmio_size && !(acc_flags & AMDGPU_REGS_IDX))
+	if ((reg * 4) < adev->rmmio_size)
 		writel(v, ((void __iomem *)adev->rmmio) + (reg * 4));
-	else {
-		unsigned long flags;
-
-		spin_lock_irqsave(&adev->mmio_idx_lock, flags);
-		writel((reg * 4), ((void __iomem *)adev->rmmio) + (mmMM_INDEX * 4));
-		writel(v, ((void __iomem *)adev->rmmio) + (mmMM_DATA * 4));
-		spin_unlock_irqrestore(&adev->mmio_idx_lock, flags);
-	}
-
-	if (adev->asic_type >= CHIP_VEGA10 && reg == 1 && adev->last_mm_index == 0x5702C) {
-		udelay(500);
-	}
+	else
+		adev->pcie_wreg(adev, (reg * 4), v);
 }
 
 /**
- * amdgpu_mm_wreg - write to a memory mapped IO register
+ * amdgpu_device_wreg - write to a register
  *
  * @adev: amdgpu_device pointer
  * @reg: dword aligned register offset
@@ -336,17 +389,13 @@ void static inline amdgpu_mm_wreg_mmio(struct amdgpu_device *adev, uint32_t reg,
  *
  * Writes the value specified to the offset specified.
  */
-void amdgpu_mm_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v,
-		    uint32_t acc_flags)
+void amdgpu_device_wreg(struct amdgpu_device *adev, uint32_t reg, uint32_t v,
+			uint32_t acc_flags)
 {
-	if (adev->asic_type >= CHIP_VEGA10 && reg == 0) {
-		adev->last_mm_index = v;
-	}
-
-	if ((acc_flags & AMDGPU_REGS_KIQ) || (!(acc_flags & AMDGPU_REGS_NO_KIQ) && amdgpu_sriov_runtime(adev)))
+	if (!(acc_flags & AMDGPU_REGS_NO_KIQ) && amdgpu_sriov_runtime(adev))
 		return amdgpu_kiq_wreg(adev, reg, v);
 
-	amdgpu_mm_wreg_mmio(adev, reg, v, acc_flags);
+	amdgpu_device_wreg_no_kiq(adev, reg, v, acc_flags);
 }
 
 /*
@@ -365,7 +414,7 @@ void amdgpu_mm_wreg_mmio_rlc(struct amdgpu_device *adev, uint32_t reg, uint32_t
 			return adev->gfx.rlc.funcs->rlcg_wreg(adev, reg, v);
 	}
 
-	amdgpu_mm_wreg_mmio(adev, reg, v, acc_flags);
+	amdgpu_device_wreg_no_kiq(adev, reg, v, acc_flags);
 }
 
 /**
@@ -397,20 +446,12 @@ u32 amdgpu_io_rreg(struct amdgpu_device *adev, u32 reg)
  */
 void amdgpu_io_wreg(struct amdgpu_device *adev, u32 reg, u32 v)
 {
-	if (adev->asic_type >= CHIP_VEGA10 && reg == 0) {
-		adev->last_mm_index = v;
-	}
-
 	if ((reg * 4) < adev->rio_mem_size)
 		iowrite32(v, adev->rio_mem + (reg * 4));
 	else {
 		iowrite32((reg * 4), adev->rio_mem + (mmMM_INDEX * 4));
 		iowrite32(v, adev->rio_mem + (mmMM_DATA * 4));
 	}
-
-	if (adev->asic_type >= CHIP_VEGA10 && reg == 1 && adev->last_mm_index == 0x5702C) {
-		udelay(500);
-	}
 }
 
 /**
@@ -1126,6 +1167,8 @@ static int amdgpu_device_check_arguments(struct amdgpu_device *adev)
 
 	adev->firmware.load_type = amdgpu_ucode_get_load_type(adev, amdgpu_fw_load_type);
 
+	amdgpu_gmc_tmz_set(adev);
+
 	return 0;
 }
 
@@ -1147,7 +1190,7 @@ static void amdgpu_switcheroo_set_state(struct pci_dev *pdev, enum vga_switchero
 		return;
 
 	if (state == VGA_SWITCHEROO_ON) {
-		pr_info("amdgpu: switched on\n");
+		pr_info("switched on\n");
 		/* don't suspend or resume card normally */
 		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
 
@@ -1161,7 +1204,7 @@ static void amdgpu_switcheroo_set_state(struct pci_dev *pdev, enum vga_switchero
 		dev->switch_power_state = DRM_SWITCH_POWER_ON;
 		drm_kms_helper_poll_enable(dev);
 	} else {
-		pr_info("amdgpu: switched off\n");
+		pr_info("switched off\n");
 		drm_kms_helper_poll_disable(dev);
 		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
 		amdgpu_device_suspend(dev, true);
@@ -1731,9 +1774,28 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev)
 	amdgpu_amdkfd_device_probe(adev);
 
 	if (amdgpu_sriov_vf(adev)) {
+		/* handle vbios stuff prior full access mode for new handshake */
+		if (adev->virt.req_init_data_ver == 1) {
+			if (!amdgpu_get_bios(adev)) {
+				DRM_ERROR("failed to get vbios\n");
+				return -EINVAL;
+			}
+
+			r = amdgpu_atombios_init(adev);
+			if (r) {
+				dev_err(adev->dev, "amdgpu_atombios_init failed\n");
+				amdgpu_vf_error_put(adev, AMDGIM_ERROR_VF_ATOMBIOS_INIT_FAIL, 0, 0);
+				return r;
+			}
+		}
+	}
+
+	/* we need to send REQ_GPU here for legacy handshaker otherwise the vbios
+	 * will not be prepared by host for this VF */
+	if (amdgpu_sriov_vf(adev) && adev->virt.req_init_data_ver < 1) {
 		r = amdgpu_virt_request_full_gpu(adev, true);
 		if (r)
-			return -EAGAIN;
+			return r;
 	}
 
 	adev->pm.pp_feature = amdgpu_pp_feature_mask;
@@ -1763,6 +1825,10 @@ static int amdgpu_device_ip_early_init(struct amdgpu_device *adev)
 		}
 		/* get the vbios after the asic_funcs are set up */
 		if (adev->ip_blocks[i].version->type == AMD_IP_BLOCK_TYPE_COMMON) {
+			/* skip vbios handling for new handshake */
+			if (amdgpu_sriov_vf(adev) && adev->virt.req_init_data_ver == 1)
+				continue;
+
 			/* Read BIOS */
 			if (!amdgpu_get_bios(adev))
 				return -EINVAL;
@@ -1889,6 +1955,12 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev)
 	if (r)
 		return r;
 
+	if (amdgpu_sriov_vf(adev) && adev->virt.req_init_data_ver > 0) {
+		r = amdgpu_virt_request_full_gpu(adev, true);
+		if (r)
+			return -EAGAIN;
+	}
+
 	for (i = 0; i < adev->num_ip_blocks; i++) {
 		if (!adev->ip_blocks[i].status.valid)
 			continue;
@@ -1975,6 +2047,8 @@ static int amdgpu_device_ip_init(struct amdgpu_device *adev)
 		amdgpu_xgmi_add_device(adev);
 	amdgpu_amdkfd_device_init(adev);
 
+	amdgpu_fru_get_product_info(adev);
+
 init_failed:
 	if (amdgpu_sriov_vf(adev))
 		amdgpu_virt_release_full_gpu(adev, true);
@@ -2008,8 +2082,24 @@ static void amdgpu_device_fill_reset_magic(struct amdgpu_device *adev)
  */
 static bool amdgpu_device_check_vram_lost(struct amdgpu_device *adev)
 {
-	return !!memcmp(adev->gart.ptr, adev->reset_magic,
-			AMDGPU_RESET_MAGIC_NUM);
+	if (memcmp(adev->gart.ptr, adev->reset_magic,
+			AMDGPU_RESET_MAGIC_NUM))
+		return true;
+
+	if (!adev->in_gpu_reset)
+		return false;
+
+	/*
+	 * For all ASICs with baco/mode1 reset, the VRAM is
+	 * always assumed to be lost.
+	 */
+	switch (amdgpu_asic_reset_method(adev)) {
+	case AMD_RESET_METHOD_BACO:
+	case AMD_RESET_METHOD_MODE1:
+		return true;
+	default:
+		return false;
+	}
 }
 
 /**
@@ -2155,6 +2245,8 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev)
 		adev->ip_blocks[i].status.late_initialized = true;
 	}
 
+	amdgpu_ras_set_error_query_ready(adev, true);
+
 	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_GATE);
 	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_GATE);
 
@@ -2187,7 +2279,8 @@ static int amdgpu_device_ip_late_init(struct amdgpu_device *adev)
 				if (gpu_instance->adev->flags & AMD_IS_APU)
 					continue;
 
-				r = amdgpu_xgmi_set_pstate(gpu_instance->adev, 0);
+				r = amdgpu_xgmi_set_pstate(gpu_instance->adev,
+						AMDGPU_XGMI_PSTATE_MIN);
 				if (r) {
 					DRM_ERROR("pstate setting failed (%d).\n", r);
 					break;
@@ -2340,6 +2433,8 @@ static int amdgpu_device_ip_suspend_phase1(struct amdgpu_device *adev)
 {
 	int i, r;
 
+	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE);
+	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE);
 
 	for (i = adev->num_ip_blocks - 1; i >= 0; i--) {
 		if (!adev->ip_blocks[i].status.valid)
@@ -2767,12 +2862,12 @@ static int amdgpu_device_get_job_timeout_settings(struct amdgpu_device *adev)
 	 * By default timeout for non compute jobs is 10000.
 	 * And there is no timeout enforced on compute jobs.
 	 * In SR-IOV or passthrough mode, timeout for compute
-	 * jobs are 10000 by default.
+	 * jobs are 60000 by default.
 	 */
 	adev->gfx_timeout = msecs_to_jiffies(10000);
 	adev->sdma_timeout = adev->video_timeout = adev->gfx_timeout;
 	if (amdgpu_sriov_vf(adev) || amdgpu_passthrough(adev))
-		adev->compute_timeout = adev->gfx_timeout;
+		adev->compute_timeout =  msecs_to_jiffies(60000);
 	else
 		adev->compute_timeout = MAX_SCHEDULE_TIMEOUT;
 
@@ -2823,6 +2918,14 @@ static int amdgpu_device_get_job_timeout_settings(struct amdgpu_device *adev)
 	return ret;
 }
 
+static const struct attribute *amdgpu_dev_attributes[] = {
+	&dev_attr_product_name.attr,
+	&dev_attr_product_number.attr,
+	&dev_attr_serial_number.attr,
+	&dev_attr_pcie_replay_count.attr,
+	NULL
+};
+
 /**
  * amdgpu_device_init - initialize the driver
  *
@@ -2924,9 +3027,6 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	INIT_LIST_HEAD(&adev->shadow_list);
 	mutex_init(&adev->shadow_list_lock);
 
-	INIT_LIST_HEAD(&adev->ring_lru_list);
-	spin_lock_init(&adev->ring_lru_list_lock);
-
 	INIT_DELAYED_WORK(&adev->delayed_init_work,
 			  amdgpu_device_delayed_init_work_handler);
 	INIT_DELAYED_WORK(&adev->gfx.gfx_off_delay_work,
@@ -2935,7 +3035,7 @@ 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->pm.ac_power = power_supply_is_system_supplied() > 0 ? true : false;
+	adev->pm.ac_power = power_supply_is_system_supplied() > 0;
 
 	/* Registers mapping */
 	/* TODO: block userspace mapping of io register */
@@ -2984,18 +3084,8 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 	if (amdgpu_mes && adev->asic_type >= CHIP_NAVI10)
 		adev->enable_mes = true;
 
-	if (amdgpu_discovery && adev->asic_type >= CHIP_NAVI10) {
-		r = amdgpu_discovery_init(adev);
-		if (r) {
-			dev_err(adev->dev, "amdgpu_discovery_init failed\n");
-			return r;
-		}
-	}
-
-	/* early init functions */
-	r = amdgpu_device_ip_early_init(adev);
-	if (r)
-		return r;
+	/* detect hw virtualization here */
+	amdgpu_detect_virtualization(adev);
 
 	r = amdgpu_device_get_job_timeout_settings(adev);
 	if (r) {
@@ -3003,6 +3093,11 @@ int amdgpu_device_init(struct amdgpu_device *adev,
 		return r;
 	}
 
+	/* early init functions */
+	r = amdgpu_device_ip_early_init(adev);
+	if (r)
+		return r;
+
 	/* doorbell bar mapping and doorbell index init*/
 	amdgpu_device_doorbell_init(adev);
 
@@ -3109,14 +3204,13 @@ fence_driver_init:
 		goto failed;
 	}
 
-	DRM_DEBUG("SE %d, SH per SE %d, CU per SH %d, active_cu_number %d\n",
+	dev_info(adev->dev,
+		"SE %d, SH per SE %d, CU per SH %d, active_cu_number %d\n",
 			adev->gfx.config.max_shader_engines,
 			adev->gfx.config.max_sh_per_se,
 			adev->gfx.config.max_cu_per_sh,
 			adev->gfx.cu_info.number);
 
-	amdgpu_ctx_init_sched(adev);
-
 	adev->accel_working = true;
 
 	amdgpu_vm_check_compute_bug(adev);
@@ -3181,9 +3275,9 @@ fence_driver_init:
 	queue_delayed_work(system_wq, &adev->delayed_init_work,
 			   msecs_to_jiffies(AMDGPU_RESUME_MS));
 
-	r = device_create_file(adev->dev, &dev_attr_pcie_replay_count);
+	r = sysfs_create_files(&adev->dev->kobj, amdgpu_dev_attributes);
 	if (r) {
-		dev_err(adev->dev, "Could not create pcie_replay_count");
+		dev_err(adev->dev, "Could not create amdgpu device attr\n");
 		return r;
 	}
 
@@ -3266,9 +3360,10 @@ void amdgpu_device_fini(struct amdgpu_device *adev)
 	adev->rmmio = NULL;
 	amdgpu_device_doorbell_fini(adev);
 
-	device_remove_file(adev->dev, &dev_attr_pcie_replay_count);
 	if (adev->ucode_sysfs_en)
 		amdgpu_ucode_sysfs_fini(adev);
+
+	sysfs_remove_files(&adev->dev->kobj, amdgpu_dev_attributes);
 	if (IS_ENABLED(CONFIG_PERF_EVENTS))
 		amdgpu_pmu_fini(adev);
 	if (amdgpu_discovery && adev->asic_type >= CHIP_NAVI10)
@@ -3354,15 +3449,12 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon)
 		}
 	}
 
-	amdgpu_device_set_pg_state(adev, AMD_PG_STATE_UNGATE);
-	amdgpu_device_set_cg_state(adev, AMD_CG_STATE_UNGATE);
-
-	amdgpu_amdkfd_suspend(adev, !fbcon);
-
 	amdgpu_ras_suspend(adev);
 
 	r = amdgpu_device_ip_suspend_phase1(adev);
 
+	amdgpu_amdkfd_suspend(adev, !fbcon);
+
 	/* evict vram memory */
 	amdgpu_bo_evict_vram(adev);
 
@@ -3739,6 +3831,8 @@ 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)
@@ -3833,6 +3927,8 @@ static int amdgpu_device_pre_asic_reset(struct amdgpu_device *adev,
 	int i, r = 0;
 	bool need_full_reset  = *need_full_reset_arg;
 
+	amdgpu_debugfs_wait_dump(adev);
+
 	/* block all schedulers and reset given job's ring */
 	for (i = 0; i < AMDGPU_MAX_RINGS; ++i) {
 		struct amdgpu_ring *ring = adev->rings[i];
@@ -4037,6 +4133,64 @@ static void amdgpu_device_unlock_adev(struct amdgpu_device *adev)
 	mutex_unlock(&adev->lock_reset);
 }
 
+static void amdgpu_device_resume_display_audio(struct amdgpu_device *adev)
+{
+	struct pci_dev *p = NULL;
+
+	p = pci_get_domain_bus_and_slot(pci_domain_nr(adev->pdev->bus),
+			adev->pdev->bus->number, 1);
+	if (p) {
+		pm_runtime_enable(&(p->dev));
+		pm_runtime_resume(&(p->dev));
+	}
+}
+
+static int amdgpu_device_suspend_display_audio(struct amdgpu_device *adev)
+{
+	enum amd_reset_method reset_method;
+	struct pci_dev *p = NULL;
+	u64 expires;
+
+	/*
+	 * For now, only BACO and mode1 reset are confirmed
+	 * to suffer the audio issue without proper suspended.
+	 */
+	reset_method = amdgpu_asic_reset_method(adev);
+	if ((reset_method != AMD_RESET_METHOD_BACO) &&
+	     (reset_method != AMD_RESET_METHOD_MODE1))
+		return -EINVAL;
+
+	p = pci_get_domain_bus_and_slot(pci_domain_nr(adev->pdev->bus),
+			adev->pdev->bus->number, 1);
+	if (!p)
+		return -ENODEV;
+
+	expires = pm_runtime_autosuspend_expiration(&(p->dev));
+	if (!expires)
+		/*
+		 * If we cannot get the audio device autosuspend delay,
+		 * a fixed 4S interval will be used. Considering 3S is
+		 * the audio controller default autosuspend delay setting.
+		 * 4S used here is guaranteed to cover that.
+		 */
+		expires = ktime_get_mono_fast_ns() + NSEC_PER_SEC * 4ULL;
+
+	while (!pm_runtime_status_suspended(&(p->dev))) {
+		if (!pm_runtime_suspend(&(p->dev)))
+			break;
+
+		if (expires < ktime_get_mono_fast_ns()) {
+			dev_warn(adev->dev, "failed to suspend display audio\n");
+			/* TODO: abort the succeeding gpu reset? */
+			return -ETIMEDOUT;
+		}
+	}
+
+	pm_runtime_disable(&(p->dev));
+
+	return 0;
+}
+
 /**
  * amdgpu_device_gpu_recover - reset the asic and recover scheduler
  *
@@ -4052,7 +4206,8 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 			      struct amdgpu_job *job)
 {
 	struct list_head device_list, *device_list_handle =  NULL;
-	bool need_full_reset, job_signaled;
+	bool need_full_reset = false;
+	bool job_signaled = false;
 	struct amdgpu_hive_info *hive = NULL;
 	struct amdgpu_device *tmp_adev = NULL;
 	int i, r = 0;
@@ -4060,6 +4215,7 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 	bool use_baco =
 		(amdgpu_asic_reset_method(adev) == AMD_RESET_METHOD_BACO) ?
 		true : false;
+	bool audio_suspended = false;
 
 	/*
 	 * Flush RAM to disk so that after reboot
@@ -4073,16 +4229,9 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 		emergency_restart();
 	}
 
-	need_full_reset = job_signaled = false;
-	INIT_LIST_HEAD(&device_list);
-
 	dev_info(adev->dev, "GPU %s begin!\n",
 		(in_ras_intr && !use_baco) ? "jobs stop":"reset");
 
-	cancel_delayed_work_sync(&adev->delayed_init_work);
-
-	hive = amdgpu_get_xgmi_hive(adev, false);
-
 	/*
 	 * Here we trylock to avoid chain of resets executing from
 	 * either trigger by jobs on different adevs in XGMI hive or jobs on
@@ -4090,39 +4239,25 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 	 * We always reset all schedulers for device and all devices for XGMI
 	 * hive so that should take care of them too.
 	 */
-
+	hive = amdgpu_get_xgmi_hive(adev, true);
 	if (hive && !mutex_trylock(&hive->reset_lock)) {
 		DRM_INFO("Bailing on TDR for s_job:%llx, hive: %llx as another already in progress",
 			  job ? job->base.id : -1, hive->hive_id);
+		mutex_unlock(&hive->hive_lock);
 		return 0;
 	}
 
-	/* Start with adev pre asic reset first for soft reset check.*/
-	if (!amdgpu_device_lock_adev(adev, !hive)) {
-		DRM_INFO("Bailing on TDR for s_job:%llx, as another already in progress",
-			  job ? job->base.id : -1);
-		return 0;
-	}
-
-	/* Block kfd: SRIOV would do it separately */
-	if (!amdgpu_sriov_vf(adev))
-                amdgpu_amdkfd_pre_reset(adev);
-
-	/* Build list of devices to reset */
-	if  (adev->gmc.xgmi.num_physical_nodes > 1) {
-		if (!hive) {
-			/*unlock kfd: SRIOV would do it separately */
-			if (!amdgpu_sriov_vf(adev))
-		                amdgpu_amdkfd_post_reset(adev);
-			amdgpu_device_unlock_adev(adev);
+	/*
+	 * Build list of devices to reset.
+	 * In case we are in XGMI hive mode, resort the device list
+	 * to put adev in the 1st position.
+	 */
+	INIT_LIST_HEAD(&device_list);
+	if (adev->gmc.xgmi.num_physical_nodes > 1) {
+		if (!hive)
 			return -ENODEV;
-		}
-
-		/*
-		 * In case we are in XGMI hive mode device reset is done for all the
-		 * nodes in the hive to retrain all XGMI links and hence the reset
-		 * sequence is executed in loop on all nodes.
-		 */
+		if (!list_is_first(&adev->gmc.xgmi.head, &hive->device_list))
+			list_rotate_to_front(&adev->gmc.xgmi.head, &hive->device_list);
 		device_list_handle = &hive->device_list;
 	} else {
 		list_add_tail(&adev->gmc.xgmi.head, &device_list);
@@ -4131,19 +4266,40 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 
 	/* block all schedulers and reset given job's ring */
 	list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) {
-		if (tmp_adev != adev) {
-			amdgpu_device_lock_adev(tmp_adev, false);
-			if (!amdgpu_sriov_vf(tmp_adev))
-			                amdgpu_amdkfd_pre_reset(tmp_adev);
+		if (!amdgpu_device_lock_adev(tmp_adev, !hive)) {
+			DRM_INFO("Bailing on TDR for s_job:%llx, as another already in progress",
+				  job ? job->base.id : -1);
+			mutex_unlock(&hive->hive_lock);
+			return 0;
 		}
 
 		/*
+		 * Try to put the audio codec into suspend state
+		 * before gpu reset started.
+		 *
+		 * Due to the power domain of the graphics device
+		 * is shared with AZ power domain. Without this,
+		 * we may change the audio hardware from behind
+		 * the audio driver's back. That will trigger
+		 * some audio codec errors.
+		 */
+		if (!amdgpu_device_suspend_display_audio(tmp_adev))
+			audio_suspended = true;
+
+		amdgpu_ras_set_error_query_ready(tmp_adev, false);
+
+		cancel_delayed_work_sync(&tmp_adev->delayed_init_work);
+
+		if (!amdgpu_sriov_vf(tmp_adev))
+			amdgpu_amdkfd_pre_reset(tmp_adev);
+
+		/*
 		 * Mark these ASICs to be reseted as untracked first
 		 * And add them back after reset completed
 		 */
 		amdgpu_unregister_gpu_instance(tmp_adev);
 
-		amdgpu_fbdev_set_suspend(adev, 1);
+		amdgpu_fbdev_set_suspend(tmp_adev, 1);
 
 		/* disable ras on ALL IPs */
 		if (!(in_ras_intr && !use_baco) &&
@@ -4163,7 +4319,6 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 		}
 	}
 
-
 	if (in_ras_intr && !use_baco)
 		goto skip_sched_resume;
 
@@ -4174,30 +4329,14 @@ int amdgpu_device_gpu_recover(struct amdgpu_device *adev,
 	 * job->base holds a reference to parent fence
 	 */
 	if (job && job->base.s_fence->parent &&
-	    dma_fence_is_signaled(job->base.s_fence->parent))
+	    dma_fence_is_signaled(job->base.s_fence->parent)) {
 		job_signaled = true;
-
-	if (job_signaled) {
 		dev_info(adev->dev, "Guilty job already signaled, skipping HW reset");
 		goto skip_hw_reset;
 	}
 
-
-	/* Guilty job will be freed after this*/
-	r = amdgpu_device_pre_asic_reset(adev, job, &need_full_reset);
-	if (r) {
-		/*TODO Should we stop ?*/
-		DRM_ERROR("GPU pre asic reset failed with err, %d for drm dev, %s ",
-			  r, adev->ddev->unique);
-		adev->asic_reset_res = r;
-	}
-
 retry:	/* Rest of adevs pre asic reset from XGMI hive. */
 	list_for_each_entry(tmp_adev, device_list_handle, gmc.xgmi.head) {
-
-		if (tmp_adev == adev)
-			continue;
-
 		r = amdgpu_device_pre_asic_reset(tmp_adev,
 						 NULL,
 						 &need_full_reset);
@@ -4259,11 +4398,15 @@ skip_sched_resume:
 		/*unlock kfd: SRIOV would do it separately */
 		if (!(in_ras_intr && !use_baco) && !amdgpu_sriov_vf(tmp_adev))
 	                amdgpu_amdkfd_post_reset(tmp_adev);
+		if (audio_suspended)
+			amdgpu_device_resume_display_audio(tmp_adev);
 		amdgpu_device_unlock_adev(tmp_adev);
 	}
 
-	if (hive)
+	if (hive) {
 		mutex_unlock(&hive->reset_lock);
+		mutex_unlock(&hive->hive_lock);
+	}
 
 	if (r)
 		dev_info(adev->dev, "GPU reset end with ret = %d\n", r);