Merge tag 'amd-drm-next-6.9-2024-02-09' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-next-6.9-2024-02-09:

amdgpu:
- Validate DMABuf imports in compute VMs
- Add RAS ACA framework
- PSP 13 fixes
- Misc code cleanups
- Replay fixes
- Atom interpretor PS, WS bounds checking
- DML2 fixes
- Audio fixes
- DCN 3.5 Z state fixes
- Remove deprecated ida_simple usage
- UBSAN fixes
- RAS fixes
- Enable seq64 infrastructure
- DC color block enablement
- Documentation updates
- DC documentation updates
- DMCUB updates
- S3 fixes
- VCN 4.0.5 fixes
- DP MST fixes
- SR-IOV fixes

amdkfd:
- Validate DMABuf imports in compute VMs
- SVM fixes
- Trap handler updates

radeon:
- Atom interpretor PS, WS bounds checking
- Misc code cleanups

UAPI:
- Bump KFD version so UMDs know that the fixes that enable the management of
  VA mappings in compute VMs using the GEM_VA ioctl for DMABufs exported from KFD are present
- Add INFO query for input power.  This matches the existing INFO query for average
  power.  Used in gaming HUDs, etc.
  Example userspace: https://github.com/Umio-Yasuno/libdrm-amdgpu-sys-rs/tree/input_power

From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240209221459.5453-1-alexander.deucher@amd.com

88 files changed, 3118 insertions, 784 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile
index 260e32ef7bae..4c989da4d2f3 100644
--- a/drivers/gpu/drm/amd/amdgpu/Makefile
+++ b/drivers/gpu/drm/amd/amdgpu/Makefile
@@ -80,7 +80,7 @@ amdgpu-y += amdgpu_device.o amdgpu_doorbell_mgr.o amdgpu_kms.o \
 	amdgpu_umc.o smu_v11_0_i2c.o amdgpu_fru_eeprom.o amdgpu_rap.o \
 	amdgpu_fw_attestation.o amdgpu_securedisplay.o \
 	amdgpu_eeprom.o amdgpu_mca.o amdgpu_psp_ta.o amdgpu_lsdma.o \
-	amdgpu_ring_mux.o amdgpu_xcp.o amdgpu_seq64.o
+	amdgpu_ring_mux.o amdgpu_xcp.o amdgpu_seq64.o amdgpu_aca.o
 
 amdgpu-$(CONFIG_PROC_FS) += amdgpu_fdinfo.o
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 3d8a48f46b01..312dfaec7b4a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -107,6 +107,7 @@
 #include "amdgpu_smuio.h"
 #include "amdgpu_fdinfo.h"
 #include "amdgpu_mca.h"
+#include "amdgpu_aca.h"
 #include "amdgpu_ras.h"
 #include "amdgpu_xcp.h"
 #include "amdgpu_seq64.h"
@@ -114,14 +115,12 @@
 
 #define MAX_GPU_INSTANCE		64
 
-struct amdgpu_gpu_instance
-{
+struct amdgpu_gpu_instance {
 	struct amdgpu_device		*adev;
 	int				mgpu_fan_enabled;
 };
 
-struct amdgpu_mgpu_info
-{
+struct amdgpu_mgpu_info {
 	struct amdgpu_gpu_instance	gpu_ins[MAX_GPU_INSTANCE];
 	struct mutex			mutex;
 	uint32_t			num_gpu;
@@ -140,8 +139,7 @@ enum amdgpu_ss {
 	AMDGPU_SS_DRV_UNLOAD
 };
 
-struct amdgpu_watchdog_timer
-{
+struct amdgpu_watchdog_timer {
 	bool timeout_fatal_disable;
 	uint32_t period; /* maxCycles = (1 << period), the number of cycles before a timeout */
 };
@@ -1045,6 +1043,9 @@ struct amdgpu_device {
 	/* MCA */
 	struct amdgpu_mca               mca;
 
+	/* ACA */
+	struct amdgpu_aca		aca;
+
 	struct amdgpu_ip_block          ip_blocks[AMDGPU_MAX_IP_NUM];
 	uint32_t		        harvest_ip_mask;
 	int				num_ip_blocks;
@@ -1078,6 +1079,8 @@ struct amdgpu_device {
 	bool				in_s3;
 	bool				in_s4;
 	bool				in_s0ix;
+	/* indicate amdgpu suspension status */
+	bool				suspend_complete;
 
 	enum pp_mp1_state               mp1_state;
 	struct amdgpu_doorbell_index doorbell_index;
@@ -1329,6 +1332,7 @@ int emu_soc_asic_init(struct amdgpu_device *adev);
 #define WREG32_FIELD_OFFSET(reg, offset, field, val)	\
 	WREG32(mm##reg + offset, (RREG32(mm##reg + offset) & ~REG_FIELD_MASK(reg, field)) | (val) << REG_FIELD_SHIFT(reg, field))
 
+#define AMDGPU_GET_REG_FIELD(x, h, l) (((x) & GENMASK_ULL(h, l)) >> (l))
 /*
  * BIOS helpers.
  */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c
new file mode 100644
index 000000000000..493982f94649
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.c
@@ -0,0 +1,879 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include <linux/list.h>
+#include "amdgpu.h"
+#include "amdgpu_aca.h"
+#include "amdgpu_ras.h"
+
+#define ACA_BANK_HWID(type, hwid, mcatype) [ACA_HWIP_TYPE_##type] = {hwid, mcatype}
+
+typedef int bank_handler_t(struct aca_handle *handle, struct aca_bank *bank, enum aca_error_type type, void *data);
+
+struct aca_banks {
+	int nr_banks;
+	struct list_head list;
+};
+
+struct aca_hwip {
+	int hwid;
+	int mcatype;
+};
+
+static struct aca_hwip aca_hwid_mcatypes[ACA_HWIP_TYPE_COUNT] = {
+	ACA_BANK_HWID(SMU,	0x01,	0x01),
+	ACA_BANK_HWID(PCS_XGMI, 0x50,	0x00),
+	ACA_BANK_HWID(UMC,	0x96,	0x00),
+};
+
+static void aca_banks_init(struct aca_banks *banks)
+{
+	if (!banks)
+		return;
+
+	memset(banks, 0, sizeof(*banks));
+	INIT_LIST_HEAD(&banks->list);
+}
+
+static int aca_banks_add_bank(struct aca_banks *banks, struct aca_bank *bank)
+{
+	struct aca_bank_node *node;
+
+	if (!bank)
+		return -EINVAL;
+
+	node = kvzalloc(sizeof(*node), GFP_KERNEL);
+	if (!node)
+		return -ENOMEM;
+
+	memcpy(&node->bank, bank, sizeof(*bank));
+
+	INIT_LIST_HEAD(&node->node);
+	list_add_tail(&node->node, &banks->list);
+
+	banks->nr_banks++;
+
+	return 0;
+}
+
+static void aca_banks_release(struct aca_banks *banks)
+{
+	struct aca_bank_node *node, *tmp;
+
+	list_for_each_entry_safe(node, tmp, &banks->list, node) {
+		list_del(&node->node);
+		kvfree(node);
+	}
+}
+
+static int aca_smu_get_valid_aca_count(struct amdgpu_device *adev, enum aca_error_type type, u32 *count)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+	const struct aca_smu_funcs *smu_funcs = aca->smu_funcs;
+
+	if (!count)
+		return -EINVAL;
+
+	if (!smu_funcs || !smu_funcs->get_valid_aca_count)
+		return -EOPNOTSUPP;
+
+	return smu_funcs->get_valid_aca_count(adev, type, count);
+}
+
+static struct aca_regs_dump {
+	const char *name;
+	int reg_idx;
+} aca_regs[] = {
+	{"CONTROL",		ACA_REG_IDX_CTL},
+	{"STATUS",		ACA_REG_IDX_STATUS},
+	{"ADDR",		ACA_REG_IDX_ADDR},
+	{"MISC",		ACA_REG_IDX_MISC0},
+	{"CONFIG",		ACA_REG_IDX_CONFG},
+	{"IPID",		ACA_REG_IDX_IPID},
+	{"SYND",		ACA_REG_IDX_SYND},
+	{"DESTAT",		ACA_REG_IDX_DESTAT},
+	{"DEADDR",		ACA_REG_IDX_DEADDR},
+	{"CONTROL_MASK",	ACA_REG_IDX_CTL_MASK},
+};
+
+static void aca_smu_bank_dump(struct amdgpu_device *adev, int idx, int total, struct aca_bank *bank)
+{
+	int i;
+
+	dev_info(adev->dev, HW_ERR "Accelerator Check Architecture events logged\n");
+	/* plus 1 for output format, e.g: ACA[08/08]: xxxx */
+	for (i = 0; i < ARRAY_SIZE(aca_regs); i++)
+		dev_info(adev->dev, HW_ERR "ACA[%02d/%02d].%s=0x%016llx\n",
+			 idx + 1, total, aca_regs[i].name, bank->regs[aca_regs[i].reg_idx]);
+}
+
+static int aca_smu_get_valid_aca_banks(struct amdgpu_device *adev, enum aca_error_type type,
+				       int start, int count,
+				       struct aca_banks *banks)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+	const struct aca_smu_funcs *smu_funcs = aca->smu_funcs;
+	struct aca_bank bank;
+	int i, max_count, ret;
+
+	if (!count)
+		return 0;
+
+	if (!smu_funcs || !smu_funcs->get_valid_aca_bank)
+		return -EOPNOTSUPP;
+
+	switch (type) {
+	case ACA_ERROR_TYPE_UE:
+		max_count = smu_funcs->max_ue_bank_count;
+		break;
+	case ACA_ERROR_TYPE_CE:
+		max_count = smu_funcs->max_ce_bank_count;
+		break;
+	case ACA_ERROR_TYPE_DEFERRED:
+	default:
+		return -EINVAL;
+	}
+
+	if (start + count >= max_count)
+		return -EINVAL;
+
+	count = min_t(int, count, max_count);
+	for (i = 0; i < count; i++) {
+		memset(&bank, 0, sizeof(bank));
+		ret = smu_funcs->get_valid_aca_bank(adev, type, start + i, &bank);
+		if (ret)
+			return ret;
+
+		aca_smu_bank_dump(adev, i, count, &bank);
+
+		ret = aca_banks_add_bank(banks, &bank);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static bool aca_bank_hwip_is_matched(struct aca_bank *bank, enum aca_hwip_type type)
+{
+
+	struct aca_hwip *hwip;
+	int hwid, mcatype;
+	u64 ipid;
+
+	if (!bank || type == ACA_HWIP_TYPE_UNKNOW)
+		return false;
+
+	hwip = &aca_hwid_mcatypes[type];
+	if (!hwip->hwid)
+		return false;
+
+	ipid = bank->regs[ACA_REG_IDX_IPID];
+	hwid = ACA_REG__IPID__HARDWAREID(ipid);
+	mcatype = ACA_REG__IPID__MCATYPE(ipid);
+
+	return hwip->hwid == hwid && hwip->mcatype == mcatype;
+}
+
+static bool aca_bank_is_valid(struct aca_handle *handle, struct aca_bank *bank, enum aca_error_type type)
+{
+	const struct aca_bank_ops *bank_ops = handle->bank_ops;
+
+	if (!aca_bank_hwip_is_matched(bank, handle->hwip))
+		return false;
+
+	if (!bank_ops->aca_bank_is_valid)
+		return true;
+
+	return bank_ops->aca_bank_is_valid(handle, bank, type, handle->data);
+}
+
+static struct aca_bank_error *new_bank_error(struct aca_error *aerr, struct aca_bank_info *info)
+{
+	struct aca_bank_error *bank_error;
+
+	bank_error = kvzalloc(sizeof(*bank_error), GFP_KERNEL);
+	if (!bank_error)
+		return NULL;
+
+	INIT_LIST_HEAD(&bank_error->node);
+	memcpy(&bank_error->info, info, sizeof(*info));
+
+	mutex_lock(&aerr->lock);
+	list_add_tail(&bank_error->node, &aerr->list);
+	mutex_unlock(&aerr->lock);
+
+	return bank_error;
+}
+
+static struct aca_bank_error *find_bank_error(struct aca_error *aerr, struct aca_bank_info *info)
+{
+	struct aca_bank_error *bank_error = NULL;
+	struct aca_bank_info *tmp_info;
+	bool found = false;
+
+	mutex_lock(&aerr->lock);
+	list_for_each_entry(bank_error, &aerr->list, node) {
+		tmp_info = &bank_error->info;
+		if (tmp_info->socket_id == info->socket_id &&
+		    tmp_info->die_id == info->die_id) {
+			found = true;
+			goto out_unlock;
+		}
+	}
+
+out_unlock:
+	mutex_unlock(&aerr->lock);
+
+	return found ? bank_error : NULL;
+}
+
+static void aca_bank_error_remove(struct aca_error *aerr, struct aca_bank_error *bank_error)
+{
+	if (!aerr || !bank_error)
+		return;
+
+	list_del(&bank_error->node);
+	aerr->nr_errors--;
+
+	kvfree(bank_error);
+}
+
+static struct aca_bank_error *get_bank_error(struct aca_error *aerr, struct aca_bank_info *info)
+{
+	struct aca_bank_error *bank_error;
+
+	if (!aerr || !info)
+		return NULL;
+
+	bank_error = find_bank_error(aerr, info);
+	if (bank_error)
+		return bank_error;
+
+	return new_bank_error(aerr, info);
+}
+
+static int aca_log_errors(struct aca_handle *handle, enum aca_error_type type,
+			  struct aca_bank_report *report)
+{
+	struct aca_error_cache *error_cache = &handle->error_cache;
+	struct aca_bank_error *bank_error;
+	struct aca_error *aerr;
+
+	if (!handle || !report)
+		return -EINVAL;
+
+	if (!report->count[type])
+		return 0;
+
+	aerr = &error_cache->errors[type];
+	bank_error = get_bank_error(aerr, &report->info);
+	if (!bank_error)
+		return -ENOMEM;
+
+	bank_error->count[type] += report->count[type];
+
+	return 0;
+}
+
+static int aca_generate_bank_report(struct aca_handle *handle, struct aca_bank *bank,
+				    enum aca_error_type type, struct aca_bank_report *report)
+{
+	const struct aca_bank_ops *bank_ops = handle->bank_ops;
+
+	if (!bank || !report)
+		return -EINVAL;
+
+	if (!bank_ops->aca_bank_generate_report)
+		return -EOPNOTSUPP;
+
+	memset(report, 0, sizeof(*report));
+	return bank_ops->aca_bank_generate_report(handle, bank, type,
+						  report, handle->data);
+}
+
+static int handler_aca_log_bank_error(struct aca_handle *handle, struct aca_bank *bank,
+				      enum aca_error_type type, void *data)
+{
+	struct aca_bank_report report;
+	int ret;
+
+	ret = aca_generate_bank_report(handle, bank, type, &report);
+	if (ret)
+		return ret;
+
+	if (!report.count[type])
+		return 0;
+
+	ret = aca_log_errors(handle, type, &report);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+static int aca_dispatch_bank(struct aca_handle_manager *mgr, struct aca_bank *bank,
+			     enum aca_error_type type, bank_handler_t handler, void *data)
+{
+	struct aca_handle *handle;
+	int ret;
+
+	if (list_empty(&mgr->list))
+		return 0;
+
+	list_for_each_entry(handle, &mgr->list, node) {
+		if (!aca_bank_is_valid(handle, bank, type))
+			continue;
+
+		ret = handler(handle, bank, type, data);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int aca_dispatch_banks(struct aca_handle_manager *mgr, struct aca_banks *banks,
+			      enum aca_error_type type, bank_handler_t handler, void *data)
+{
+	struct aca_bank_node *node;
+	struct aca_bank *bank;
+	int ret;
+
+	if (!mgr || !banks)
+		return -EINVAL;
+
+	/* pre check to avoid unnecessary operations */
+	if (list_empty(&mgr->list) || list_empty(&banks->list))
+		return 0;
+
+	list_for_each_entry(node, &banks->list, node) {
+		bank = &node->bank;
+
+		ret = aca_dispatch_bank(mgr, bank, type, handler, data);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static int aca_banks_update(struct amdgpu_device *adev, enum aca_error_type type,
+			    bank_handler_t handler, void *data)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+	struct aca_banks banks;
+	u32 count = 0;
+	int ret;
+
+	if (list_empty(&aca->mgr.list))
+		return 0;
+
+	/* NOTE: pmfw is only support UE and CE */
+	if (type == ACA_ERROR_TYPE_DEFERRED)
+		type = ACA_ERROR_TYPE_CE;
+
+	ret = aca_smu_get_valid_aca_count(adev, type, &count);
+	if (ret)
+		return ret;
+
+	if (!count)
+		return 0;
+
+	aca_banks_init(&banks);
+
+	ret = aca_smu_get_valid_aca_banks(adev, type, 0, count, &banks);
+	if (ret)
+		goto err_release_banks;
+
+	if (list_empty(&banks.list)) {
+		ret = 0;
+		goto err_release_banks;
+	}
+
+	ret = aca_dispatch_banks(&aca->mgr, &banks, type,
+				 handler, data);
+	if (ret)
+		goto err_release_banks;
+
+err_release_banks:
+	aca_banks_release(&banks);
+
+	return ret;
+}
+
+static int aca_log_aca_error_data(struct aca_bank_error *bank_error, enum aca_error_type type, struct ras_err_data *err_data)
+{
+	struct aca_bank_info *info;
+	struct amdgpu_smuio_mcm_config_info mcm_info;
+	u64 count;
+
+	if (type >= ACA_ERROR_TYPE_COUNT)
+		return -EINVAL;
+
+	count = bank_error->count[type];
+	if (!count)
+		return 0;
+
+	info = &bank_error->info;
+	mcm_info.die_id = info->die_id;
+	mcm_info.socket_id = info->socket_id;
+
+	switch (type) {
+	case ACA_ERROR_TYPE_UE:
+		amdgpu_ras_error_statistic_ue_count(err_data, &mcm_info, NULL, count);
+		break;
+	case ACA_ERROR_TYPE_CE:
+		amdgpu_ras_error_statistic_ce_count(err_data, &mcm_info, NULL, count);
+		break;
+	case ACA_ERROR_TYPE_DEFERRED:
+	default:
+		break;
+	}
+
+	return 0;
+}
+
+static int aca_log_aca_error(struct aca_handle *handle, enum aca_error_type type, struct ras_err_data *err_data)
+{
+	struct aca_error_cache *error_cache = &handle->error_cache;
+	struct aca_error *aerr = &error_cache->errors[type];
+	struct aca_bank_error *bank_error, *tmp;
+
+	mutex_lock(&aerr->lock);
+
+	if (list_empty(&aerr->list))
+		goto out_unlock;
+
+	list_for_each_entry_safe(bank_error, tmp, &aerr->list, node) {
+		aca_log_aca_error_data(bank_error, type, err_data);
+		aca_bank_error_remove(aerr, bank_error);
+	}
+
+out_unlock:
+	mutex_unlock(&aerr->lock);
+
+	return 0;
+}
+
+static int __aca_get_error_data(struct amdgpu_device *adev, struct aca_handle *handle, enum aca_error_type type,
+				struct ras_err_data *err_data)
+{
+	int ret;
+
+	/* udpate aca bank to aca source error_cache first */
+	ret = aca_banks_update(adev, type, handler_aca_log_bank_error, NULL);
+	if (ret)
+		return ret;
+
+	return aca_log_aca_error(handle, type, err_data);
+}
+
+static bool aca_handle_is_valid(struct aca_handle *handle)
+{
+	if (!handle->mask || !list_empty(&handle->node))
+		return false;
+
+	return true;
+}
+
+int amdgpu_aca_get_error_data(struct amdgpu_device *adev, struct aca_handle *handle,
+			      enum aca_error_type type, void *data)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)data;
+
+	if (!handle || !err_data)
+		return -EINVAL;
+
+	if (aca_handle_is_valid(handle))
+		return -EOPNOTSUPP;
+
+	if (!(BIT(type) & handle->mask))
+		return  0;
+
+	return __aca_get_error_data(adev, handle, type, err_data);
+}
+
+static void aca_error_init(struct aca_error *aerr, enum aca_error_type type)
+{
+	mutex_init(&aerr->lock);
+	INIT_LIST_HEAD(&aerr->list);
+	aerr->type = type;
+	aerr->nr_errors = 0;
+}
+
+static void aca_init_error_cache(struct aca_handle *handle)
+{
+	struct aca_error_cache *error_cache = &handle->error_cache;
+	int type;
+
+	for (type = ACA_ERROR_TYPE_UE; type < ACA_ERROR_TYPE_COUNT; type++)
+		aca_error_init(&error_cache->errors[type], type);
+}
+
+static void aca_error_fini(struct aca_error *aerr)
+{
+	struct aca_bank_error *bank_error, *tmp;
+
+	mutex_lock(&aerr->lock);
+	list_for_each_entry_safe(bank_error, tmp, &aerr->list, node)
+		aca_bank_error_remove(aerr, bank_error);
+
+	mutex_destroy(&aerr->lock);
+}
+
+static void aca_fini_error_cache(struct aca_handle *handle)
+{
+	struct aca_error_cache *error_cache = &handle->error_cache;
+	int type;
+
+	for (type = ACA_ERROR_TYPE_UE; type < ACA_ERROR_TYPE_COUNT; type++)
+		aca_error_fini(&error_cache->errors[type]);
+}
+
+static int add_aca_handle(struct amdgpu_device *adev, struct aca_handle_manager *mgr, struct aca_handle *handle,
+			  const char *name, const struct aca_info *ras_info, void *data)
+{
+	memset(handle, 0, sizeof(*handle));
+
+	handle->adev = adev;
+	handle->mgr = mgr;
+	handle->name = name;
+	handle->hwip = ras_info->hwip;
+	handle->mask = ras_info->mask;
+	handle->bank_ops = ras_info->bank_ops;
+	handle->data = data;
+	aca_init_error_cache(handle);
+
+	INIT_LIST_HEAD(&handle->node);
+	list_add_tail(&handle->node, &mgr->list);
+	mgr->nr_handles++;
+
+	return 0;
+}
+
+static ssize_t aca_sysfs_read(struct device *dev,
+			      struct device_attribute *attr, char *buf)
+{
+	struct aca_handle *handle = container_of(attr, struct aca_handle, aca_attr);
+
+	/* NOTE: the aca cache will be auto cleared once read,
+	 * So the driver should unify the query entry point, forward request to ras query interface directly */
+	return amdgpu_ras_aca_sysfs_read(dev, attr, handle, buf, handle->data);
+}
+
+static int add_aca_sysfs(struct amdgpu_device *adev, struct aca_handle *handle)
+{
+	struct device_attribute *aca_attr = &handle->aca_attr;
+
+	snprintf(handle->attr_name, sizeof(handle->attr_name) - 1, "aca_%s", handle->name);
+	aca_attr->show = aca_sysfs_read;
+	aca_attr->attr.name = handle->attr_name;
+	aca_attr->attr.mode = S_IRUGO;
+	sysfs_attr_init(&aca_attr->attr);
+
+	return sysfs_add_file_to_group(&adev->dev->kobj,
+				       &aca_attr->attr,
+				       "ras");
+}
+
+int amdgpu_aca_add_handle(struct amdgpu_device *adev, struct aca_handle *handle,
+			  const char *name, const struct aca_info *ras_info, void *data)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+	int ret;
+
+	if (!amdgpu_aca_is_enabled(adev))
+		return 0;
+
+	ret = add_aca_handle(adev, &aca->mgr, handle, name, ras_info, data);
+	if (ret)
+		return ret;
+
+	return add_aca_sysfs(adev, handle);
+}
+
+static void remove_aca_handle(struct aca_handle *handle)
+{
+	struct aca_handle_manager *mgr = handle->mgr;
+
+	aca_fini_error_cache(handle);
+	list_del(&handle->node);
+	mgr->nr_handles--;
+}
+
+static void remove_aca_sysfs(struct aca_handle *handle)
+{
+	struct amdgpu_device *adev = handle->adev;
+	struct device_attribute *aca_attr = &handle->aca_attr;
+
+	if (adev->dev->kobj.sd)
+		sysfs_remove_file_from_group(&adev->dev->kobj,
+					     &aca_attr->attr,
+					     "ras");
+}
+
+void amdgpu_aca_remove_handle(struct aca_handle *handle)
+{
+	if (!handle || list_empty(&handle->node))
+		return;
+
+	remove_aca_sysfs(handle);
+	remove_aca_handle(handle);
+}
+
+static int aca_manager_init(struct aca_handle_manager *mgr)
+{
+	INIT_LIST_HEAD(&mgr->list);
+	mgr->nr_handles = 0;
+
+	return 0;
+}
+
+static void aca_manager_fini(struct aca_handle_manager *mgr)
+{
+	struct aca_handle *handle, *tmp;
+
+	list_for_each_entry_safe(handle, tmp, &mgr->list, node)
+		amdgpu_aca_remove_handle(handle);
+}
+
+bool amdgpu_aca_is_enabled(struct amdgpu_device *adev)
+{
+	return adev->aca.is_enabled;
+}
+
+int amdgpu_aca_init(struct amdgpu_device *adev)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+	int ret;
+
+	ret = aca_manager_init(&aca->mgr);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
+void amdgpu_aca_fini(struct amdgpu_device *adev)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+
+	aca_manager_fini(&aca->mgr);
+}
+
+int amdgpu_aca_reset(struct amdgpu_device *adev)
+{
+	amdgpu_aca_fini(adev);
+
+	return amdgpu_aca_init(adev);
+}
+
+void amdgpu_aca_set_smu_funcs(struct amdgpu_device *adev, const struct aca_smu_funcs *smu_funcs)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+
+	WARN_ON(aca->smu_funcs);
+	aca->smu_funcs = smu_funcs;
+}
+
+int aca_bank_info_decode(struct aca_bank *bank, struct aca_bank_info *info)
+{
+	u64 ipid;
+	u32 instidhi, instidlo;
+
+	if (!bank || !info)
+		return -EINVAL;
+
+	ipid = bank->regs[ACA_REG_IDX_IPID];
+	info->hwid = ACA_REG__IPID__HARDWAREID(ipid);
+	info->mcatype = ACA_REG__IPID__MCATYPE(ipid);
+	/*
+	 * Unfied DieID Format: SAASS. A:AID, S:Socket.
+	 * Unfied DieID[4:4] = InstanceId[0:0]
+	 * Unfied DieID[0:3] = InstanceIdHi[0:3]
+	 */
+	instidhi = ACA_REG__IPID__INSTANCEIDHI(ipid);
+	instidlo = ACA_REG__IPID__INSTANCEIDLO(ipid);
+	info->die_id = ((instidhi >> 2) & 0x03);
+	info->socket_id = ((instidlo & 0x1) << 2) | (instidhi & 0x03);
+
+	return 0;
+}
+
+static int aca_bank_get_error_code(struct amdgpu_device *adev, struct aca_bank *bank)
+{
+	int error_code;
+
+	switch (amdgpu_ip_version(adev, MP1_HWIP, 0)) {
+	case IP_VERSION(13, 0, 6):
+		if (!(adev->flags & AMD_IS_APU) && adev->pm.fw_version >= 0x00555600) {
+			error_code = ACA_REG__SYND__ERRORINFORMATION(bank->regs[ACA_REG_IDX_SYND]);
+			return error_code & 0xff;
+		}
+		break;
+	default:
+		break;
+	}
+
+	/* NOTE: the true error code is encoded in status.errorcode[0:7] */
+	error_code = ACA_REG__STATUS__ERRORCODE(bank->regs[ACA_REG_IDX_STATUS]);
+
+	return error_code & 0xff;
+}
+
+int aca_bank_check_error_codes(struct amdgpu_device *adev, struct aca_bank *bank, int *err_codes, int size)
+{
+	int i, error_code;
+
+	if (!bank || !err_codes)
+		return -EINVAL;
+
+	error_code = aca_bank_get_error_code(adev, bank);
+	for (i = 0; i < size; i++) {
+		if (err_codes[i] == error_code)
+			return 0;
+	}
+
+	return -EINVAL;
+}
+
+int amdgpu_aca_smu_set_debug_mode(struct amdgpu_device *adev, bool en)
+{
+	struct amdgpu_aca *aca = &adev->aca;
+	const struct aca_smu_funcs *smu_funcs = aca->smu_funcs;
+
+	if (!smu_funcs || !smu_funcs->set_debug_mode)
+		return -EOPNOTSUPP;
+
+	return smu_funcs->set_debug_mode(adev, en);
+}
+
+#if defined(CONFIG_DEBUG_FS)
+static int amdgpu_aca_smu_debug_mode_set(void *data, u64 val)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)data;
+	int ret;
+
+	ret = amdgpu_ras_set_aca_debug_mode(adev, val ? true : false);
+	if (ret)
+		return ret;
+
+	dev_info(adev->dev, "amdgpu set smu aca debug mode %s success\n", val ? "on" : "off");
+
+	return 0;
+}
+
+static void aca_dump_entry(struct seq_file *m, struct aca_bank *bank, enum aca_error_type type, int idx)
+{
+	struct aca_bank_info info;
+	int i, ret;
+
+	ret = aca_bank_info_decode(bank, &info);
+	if (ret)
+		return;
+
+	seq_printf(m, "aca entry[%d].type: %s\n", idx, type ==  ACA_ERROR_TYPE_UE ? "UE" : "CE");
+	seq_printf(m, "aca entry[%d].info: socketid:%d aid:%d hwid:0x%03x mcatype:0x%04x\n",
+		   idx, info.socket_id, info.die_id, info.hwid, info.mcatype);
+
+	for (i = 0; i < ARRAY_SIZE(aca_regs); i++)
+		seq_printf(m, "aca entry[%d].regs[%d]: 0x%016llx\n", idx, aca_regs[i].reg_idx, bank->regs[aca_regs[i].reg_idx]);
+}
+
+struct aca_dump_context {
+	struct seq_file *m;
+	int idx;
+};
+
+static int handler_aca_bank_dump(struct aca_handle *handle, struct aca_bank *bank,
+				 enum aca_error_type type, void *data)
+{
+	struct aca_dump_context *ctx = (struct aca_dump_context *)data;
+
+	aca_dump_entry(ctx->m, bank, type, ctx->idx++);
+
+	return handler_aca_log_bank_error(handle, bank, type, NULL);
+}
+
+static int aca_dump_show(struct seq_file *m, enum aca_error_type type)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)m->private;
+	struct aca_dump_context context = {
+		.m = m,
+		.idx = 0,
+	};
+
+	return aca_banks_update(adev, type, handler_aca_bank_dump, (void *)&context);
+}
+
+static int aca_dump_ce_show(struct seq_file *m, void *unused)
+{
+	return aca_dump_show(m, ACA_ERROR_TYPE_CE);
+}
+
+static int aca_dump_ce_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, aca_dump_ce_show, inode->i_private);
+}
+
+static const struct file_operations aca_ce_dump_debug_fops = {
+	.owner = THIS_MODULE,
+	.open = aca_dump_ce_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+};
+
+static int aca_dump_ue_show(struct seq_file *m, void *unused)
+{
+	return aca_dump_show(m, ACA_ERROR_TYPE_UE);
+}
+
+static int aca_dump_ue_open(struct inode *inode, struct file *file)
+{
+	return single_open(file, aca_dump_ue_show, inode->i_private);
+}
+
+static const struct file_operations aca_ue_dump_debug_fops = {
+	.owner = THIS_MODULE,
+	.open = aca_dump_ue_open,
+	.read = seq_read,
+	.llseek = seq_lseek,
+	.release = single_release,
+};
+
+DEFINE_DEBUGFS_ATTRIBUTE(aca_debug_mode_fops, NULL, amdgpu_aca_smu_debug_mode_set, "%llu\n");
+#endif
+
+void amdgpu_aca_smu_debugfs_init(struct amdgpu_device *adev, struct dentry *root)
+{
+#if defined(CONFIG_DEBUG_FS)
+	if (!root || adev->ip_versions[MP1_HWIP][0] != IP_VERSION(13, 0, 6))
+		return;
+
+	debugfs_create_file("aca_debug_mode", 0200, root, adev, &aca_debug_mode_fops);
+	debugfs_create_file("aca_ue_dump", 0400, root, adev, &aca_ue_dump_debug_fops);
+	debugfs_create_file("aca_ce_dump", 0400, root, adev, &aca_ce_dump_debug_fops);
+#endif
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h
new file mode 100644
index 000000000000..2da50e095883
--- /dev/null
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_aca.h
@@ -0,0 +1,202 @@
+/*
+ * Copyright 2023 Advanced Micro Devices, Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef __AMDGPU_ACA_H__
+#define __AMDGPU_ACA_H__
+
+#include <linux/list.h>
+
+#define ACA_MAX_REGS_COUNT	(16)
+
+#define ACA_REG_FIELD(x, h, l)			(((x) & GENMASK_ULL(h, l)) >> l)
+#define ACA_REG__STATUS__VAL(x)			ACA_REG_FIELD(x, 63, 63)
+#define ACA_REG__STATUS__OVERFLOW(x)		ACA_REG_FIELD(x, 62, 62)
+#define ACA_REG__STATUS__UC(x)			ACA_REG_FIELD(x, 61, 61)
+#define ACA_REG__STATUS__EN(x)			ACA_REG_FIELD(x, 60, 60)
+#define ACA_REG__STATUS__MISCV(x)		ACA_REG_FIELD(x, 59, 59)
+#define ACA_REG__STATUS__ADDRV(x)		ACA_REG_FIELD(x, 58, 58)
+#define ACA_REG__STATUS__PCC(x)			ACA_REG_FIELD(x, 57, 57)
+#define ACA_REG__STATUS__ERRCOREIDVAL(x)	ACA_REG_FIELD(x, 56, 56)
+#define ACA_REG__STATUS__TCC(x)			ACA_REG_FIELD(x, 55, 55)
+#define ACA_REG__STATUS__SYNDV(x)		ACA_REG_FIELD(x, 53, 53)
+#define ACA_REG__STATUS__CECC(x)		ACA_REG_FIELD(x, 46, 46)
+#define ACA_REG__STATUS__UECC(x)		ACA_REG_FIELD(x, 45, 45)
+#define ACA_REG__STATUS__DEFERRED(x)		ACA_REG_FIELD(x, 44, 44)
+#define ACA_REG__STATUS__POISON(x)		ACA_REG_FIELD(x, 43, 43)
+#define ACA_REG__STATUS__SCRUB(x)		ACA_REG_FIELD(x, 40, 40)
+#define ACA_REG__STATUS__ERRCOREID(x)		ACA_REG_FIELD(x, 37, 32)
+#define ACA_REG__STATUS__ADDRLSB(x)		ACA_REG_FIELD(x, 29, 24)
+#define ACA_REG__STATUS__ERRORCODEEXT(x)	ACA_REG_FIELD(x, 21, 16)
+#define ACA_REG__STATUS__ERRORCODE(x)		ACA_REG_FIELD(x, 15, 0)
+
+#define ACA_REG__IPID__MCATYPE(x)		ACA_REG_FIELD(x, 63, 48)
+#define ACA_REG__IPID__INSTANCEIDHI(x)		ACA_REG_FIELD(x, 47, 44)
+#define ACA_REG__IPID__HARDWAREID(x)		ACA_REG_FIELD(x, 43, 32)
+#define ACA_REG__IPID__INSTANCEIDLO(x)		ACA_REG_FIELD(x, 31, 0)
+
+#define ACA_REG__MISC0__VALID(x)		ACA_REG_FIELD(x, 63, 63)
+#define ACA_REG__MISC0__OVRFLW(x)		ACA_REG_FIELD(x, 48, 48)
+#define ACA_REG__MISC0__ERRCNT(x)		ACA_REG_FIELD(x, 43, 32)
+
+#define ACA_REG__SYND__ERRORINFORMATION(x)	ACA_REG_FIELD(x, 17, 0)
+
+/* NOTE: The following codes refers to the smu header file */
+#define ACA_EXTERROR_CODE_CE			0x3a
+#define ACA_EXTERROR_CODE_FAULT			0x3b
+
+#define ACA_ERROR_UE_MASK		BIT_MASK(ACA_ERROR_TYPE_UE)
+#define ACA_ERROR_CE_MASK		BIT_MASK(ACA_ERROR_TYPE_CE)
+#define ACA_ERROR_DEFERRED_MASK		BIT_MASK(ACA_ERROR_TYPE_DEFERRED)
+
+enum aca_reg_idx {
+	ACA_REG_IDX_CTL			= 0,
+	ACA_REG_IDX_STATUS		= 1,
+	ACA_REG_IDX_ADDR		= 2,
+	ACA_REG_IDX_MISC0		= 3,
+	ACA_REG_IDX_CONFG		= 4,
+	ACA_REG_IDX_IPID		= 5,
+	ACA_REG_IDX_SYND		= 6,
+	ACA_REG_IDX_DESTAT		= 8,
+	ACA_REG_IDX_DEADDR		= 9,
+	ACA_REG_IDX_CTL_MASK		= 10,
+	ACA_REG_IDX_COUNT		= 16,
+};
+
+enum aca_hwip_type {
+	ACA_HWIP_TYPE_UNKNOW = -1,
+	ACA_HWIP_TYPE_PSP = 0,
+	ACA_HWIP_TYPE_UMC,
+	ACA_HWIP_TYPE_SMU,
+	ACA_HWIP_TYPE_PCS_XGMI,
+	ACA_HWIP_TYPE_COUNT,
+};
+
+enum aca_error_type {
+	ACA_ERROR_TYPE_INVALID = -1,
+	ACA_ERROR_TYPE_UE = 0,
+	ACA_ERROR_TYPE_CE,
+	ACA_ERROR_TYPE_DEFERRED,
+	ACA_ERROR_TYPE_COUNT
+};
+
+struct aca_bank {
+	u64 regs[ACA_MAX_REGS_COUNT];
+};
+
+struct aca_bank_node {
+	struct aca_bank bank;
+	struct list_head node;
+};
+
+struct aca_bank_info {
+	int die_id;
+	int socket_id;
+	int hwid;
+	int mcatype;
+};
+
+struct aca_bank_report {
+	struct aca_bank_info info;
+	u64 count[ACA_ERROR_TYPE_COUNT];
+};
+
+struct aca_bank_error {
+	struct list_head node;
+	struct aca_bank_info info;
+	u64 count[ACA_ERROR_TYPE_COUNT];
+};
+
+struct aca_error {
+	struct list_head list;
+	struct mutex lock;
+	enum aca_error_type type;
+	int nr_errors;
+};
+
+struct aca_handle_manager {
+	struct list_head list;
+	int nr_handles;
+};
+
+struct aca_error_cache {
+	struct aca_error errors[ACA_ERROR_TYPE_COUNT];
+};
+
+struct aca_handle {
+	struct list_head node;
+	enum aca_hwip_type hwip;
+	struct amdgpu_device *adev;
+	struct aca_handle_manager *mgr;
+	struct aca_error_cache error_cache;
+	const struct aca_bank_ops *bank_ops;
+	struct device_attribute aca_attr;
+	char attr_name[64];
+	const char *name;
+	u32 mask;
+	void *data;
+};
+
+struct aca_bank_ops {
+	int (*aca_bank_generate_report)(struct aca_handle *handle, struct aca_bank *bank, enum aca_error_type type,
+					struct aca_bank_report *report, void *data);
+	bool (*aca_bank_is_valid)(struct aca_handle *handle, struct aca_bank *bank, enum aca_error_type type,
+				  void *data);
+};
+
+struct aca_smu_funcs {
+	int max_ue_bank_count;
+	int max_ce_bank_count;
+	int (*set_debug_mode)(struct amdgpu_device *adev, bool enable);
+	int (*get_valid_aca_count)(struct amdgpu_device *adev, enum aca_error_type type, u32 *count);
+	int (*get_valid_aca_bank)(struct amdgpu_device *adev, enum aca_error_type type, int idx, struct aca_bank *bank);
+};
+
+struct amdgpu_aca {
+	struct aca_handle_manager mgr;
+	const struct aca_smu_funcs *smu_funcs;
+	bool is_enabled;
+};
+
+struct aca_info {
+	enum aca_hwip_type hwip;
+	const struct aca_bank_ops *bank_ops;
+	u32 mask;
+};
+
+int amdgpu_aca_init(struct amdgpu_device *adev);
+void amdgpu_aca_fini(struct amdgpu_device *adev);
+int amdgpu_aca_reset(struct amdgpu_device *adev);
+void amdgpu_aca_set_smu_funcs(struct amdgpu_device *adev, const struct aca_smu_funcs *smu_funcs);
+bool amdgpu_aca_is_enabled(struct amdgpu_device *adev);
+
+int aca_bank_info_decode(struct aca_bank *bank, struct aca_bank_info *info);
+int aca_bank_check_error_codes(struct amdgpu_device *adev, struct aca_bank *bank, int *err_codes, int size);
+
+int amdgpu_aca_add_handle(struct amdgpu_device *adev, struct aca_handle *handle,
+			  const char *name, const struct aca_info *aca_info, void *data);
+void amdgpu_aca_remove_handle(struct aca_handle *handle);
+int amdgpu_aca_get_error_data(struct amdgpu_device *adev, struct aca_handle *handle,
+				     enum aca_error_type type, void *data);
+int amdgpu_aca_smu_set_debug_mode(struct amdgpu_device *adev, bool en);
+void amdgpu_aca_smu_debugfs_init(struct amdgpu_device *adev, struct dentry *root);
+#endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
index 41db030ddc4e..190039f14c30 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c
@@ -742,9 +742,10 @@ void amdgpu_amdkfd_debug_mem_fence(struct amdgpu_device *adev)
 	amdgpu_device_flush_hdp(adev, NULL);
 }
 
-void amdgpu_amdkfd_ras_poison_consumption_handler(struct amdgpu_device *adev, bool reset)
+void amdgpu_amdkfd_ras_poison_consumption_handler(struct amdgpu_device *adev,
+	enum amdgpu_ras_block block, bool reset)
 {
-	amdgpu_umc_poison_handler(adev, reset);
+	amdgpu_umc_poison_handler(adev, block, reset);
 }
 
 int amdgpu_amdkfd_send_close_event_drain_irq(struct amdgpu_device *adev,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
index 27c61c535e29..e60f63ccf79a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h
@@ -193,6 +193,9 @@ struct amdgpu_amdkfd_fence *to_amdgpu_amdkfd_fence(struct dma_fence *f);
 int amdgpu_amdkfd_remove_fence_on_pt_pd_bos(struct amdgpu_bo *bo);
 int amdgpu_amdkfd_evict_userptr(struct mmu_interval_notifier *mni,
 				unsigned long cur_seq, struct kgd_mem *mem);
+int amdgpu_amdkfd_bo_validate_and_fence(struct amdgpu_bo *bo,
+					uint32_t domain,
+					struct dma_fence *fence);
 #else
 static inline
 bool amdkfd_fence_check_mm(struct dma_fence *f, struct mm_struct *mm)
@@ -218,6 +221,13 @@ int amdgpu_amdkfd_evict_userptr(struct mmu_interval_notifier *mni,
 {
 	return 0;
 }
+static inline
+int amdgpu_amdkfd_bo_validate_and_fence(struct amdgpu_bo *bo,
+					uint32_t domain,
+					struct dma_fence *fence)
+{
+	return 0;
+}
 #endif
 /* Shared API */
 int amdgpu_amdkfd_alloc_gtt_mem(struct amdgpu_device *adev, size_t size,
@@ -326,7 +336,7 @@ void amdgpu_amdkfd_debug_mem_fence(struct amdgpu_device *adev);
 int amdgpu_amdkfd_get_tile_config(struct amdgpu_device *adev,
 				struct tile_config *config);
 void amdgpu_amdkfd_ras_poison_consumption_handler(struct amdgpu_device *adev,
-				bool reset);
+			enum amdgpu_ras_block block, bool reset);
 bool amdgpu_amdkfd_bo_mapped_to_dev(struct amdgpu_device *adev, struct kgd_mem *mem);
 void amdgpu_amdkfd_block_mmu_notifications(void *p);
 int amdgpu_amdkfd_criu_resume(void *p);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
index 231fd927dcfb..5cd84f72bf26 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c
@@ -426,9 +426,9 @@ validate_fail:
 	return ret;
 }
 
-static int amdgpu_amdkfd_bo_validate_and_fence(struct amdgpu_bo *bo,
-					       uint32_t domain,
-					       struct dma_fence *fence)
+int amdgpu_amdkfd_bo_validate_and_fence(struct amdgpu_bo *bo,
+					uint32_t domain,
+					struct dma_fence *fence)
 {
 	int ret = amdgpu_bo_reserve(bo, false);
 
@@ -464,13 +464,15 @@ static int amdgpu_amdkfd_validate_vm_bo(void *_unused, struct amdgpu_bo *bo)
  * again. Page directories are only updated after updating page
  * tables.
  */
-static int vm_validate_pt_pd_bos(struct amdgpu_vm *vm)
+static int vm_validate_pt_pd_bos(struct amdgpu_vm *vm,
+				 struct ww_acquire_ctx *ticket)
 {
 	struct amdgpu_bo *pd = vm->root.bo;
 	struct amdgpu_device *adev = amdgpu_ttm_adev(pd->tbo.bdev);
 	int ret;
 
-	ret = amdgpu_vm_validate_pt_bos(adev, vm, amdgpu_amdkfd_validate_vm_bo, NULL);
+	ret = amdgpu_vm_validate(adev, vm, ticket,
+				 amdgpu_amdkfd_validate_vm_bo, NULL);
 	if (ret) {
 		pr_err("failed to validate PT BOs\n");
 		return ret;
@@ -1310,14 +1312,15 @@ update_gpuvm_pte_failed:
 	return ret;
 }
 
-static int process_validate_vms(struct amdkfd_process_info *process_info)
+static int process_validate_vms(struct amdkfd_process_info *process_info,
+				struct ww_acquire_ctx *ticket)
 {
 	struct amdgpu_vm *peer_vm;
 	int ret;
 
 	list_for_each_entry(peer_vm, &process_info->vm_list_head,
 			    vm_list_node) {
-		ret = vm_validate_pt_pd_bos(peer_vm);
+		ret = vm_validate_pt_pd_bos(peer_vm, ticket);
 		if (ret)
 			return ret;
 	}
@@ -1402,7 +1405,7 @@ static int init_kfd_vm(struct amdgpu_vm *vm, void **process_info,
 	ret = amdgpu_bo_reserve(vm->root.bo, true);
 	if (ret)
 		goto reserve_pd_fail;
-	ret = vm_validate_pt_pd_bos(vm);
+	ret = vm_validate_pt_pd_bos(vm, NULL);
 	if (ret) {
 		pr_err("validate_pt_pd_bos() failed\n");
 		goto validate_pd_fail;
@@ -2043,7 +2046,7 @@ int amdgpu_amdkfd_gpuvm_map_memory_to_gpu(
 	    bo->tbo.resource->mem_type == TTM_PL_SYSTEM)
 		is_invalid_userptr = true;
 
-	ret = vm_validate_pt_pd_bos(avm);
+	ret = vm_validate_pt_pd_bos(avm, NULL);
 	if (unlikely(ret))
 		goto out_unreserve;
 
@@ -2136,7 +2139,7 @@ int amdgpu_amdkfd_gpuvm_unmap_memory_from_gpu(
 		goto unreserve_out;
 	}
 
-	ret = vm_validate_pt_pd_bos(avm);
+	ret = vm_validate_pt_pd_bos(avm, NULL);
 	if (unlikely(ret))
 		goto unreserve_out;
 
@@ -2634,7 +2637,7 @@ static int validate_invalid_user_pages(struct amdkfd_process_info *process_info)
 		}
 	}
 
-	ret = process_validate_vms(process_info);
+	ret = process_validate_vms(process_info, NULL);
 	if (ret)
 		goto unreserve_out;
 
@@ -2894,11 +2897,6 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence __rcu *
 
 	amdgpu_sync_create(&sync_obj);
 
-	/* Validate PDs and PTs */
-	ret = process_validate_vms(process_info);
-	if (ret)
-		goto validate_map_fail;
-
 	/* Validate BOs and map them to GPUVM (update VM page tables). */
 	list_for_each_entry(mem, &process_info->kfd_bo_list,
 			    validate_list) {
@@ -2949,6 +2947,13 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence __rcu *
 	if (failed_size)
 		pr_debug("0x%lx/0x%lx in system\n", failed_size, total_size);
 
+	/* Validate PDs, PTs and evicted DMABuf imports last. Otherwise BO
+	 * validations above would invalidate DMABuf imports again.
+	 */
+	ret = process_validate_vms(process_info, &exec.ticket);
+	if (ret)
+		goto validate_map_fail;
+
 	/* Update mappings not managed by KFD */
 	list_for_each_entry(peer_vm, &process_info->vm_list_head,
 			vm_list_node) {
@@ -3020,7 +3025,7 @@ int amdgpu_amdkfd_gpuvm_restore_process_bos(void *info, struct dma_fence __rcu *
 				   &process_info->eviction_fence->base,
 				   DMA_RESV_USAGE_BOOKKEEP);
 	}
-	/* Attach eviction fence to PD / PT BOs */
+	/* Attach eviction fence to PD / PT BOs and DMABuf imports */
 	list_for_each_entry(peer_vm, &process_info->vm_list_head,
 			    vm_list_node) {
 		struct amdgpu_bo *bo = peer_vm->root.bo;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
index dce9e7d5e4ec..52b12c1718eb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c
@@ -1018,7 +1018,8 @@ int amdgpu_atombios_get_clock_dividers(struct amdgpu_device *adev,
 		if (clock_type == COMPUTE_ENGINE_PLL_PARAM) {
 			args.v3.ulClockParams = cpu_to_le32((clock_type << 24) | clock);
 
-			amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+			amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args,
+				sizeof(args));
 
 			dividers->post_div = args.v3.ucPostDiv;
 			dividers->enable_post_div = (args.v3.ucCntlFlag &
@@ -1038,7 +1039,8 @@ int amdgpu_atombios_get_clock_dividers(struct amdgpu_device *adev,
 			if (strobe_mode)
 				args.v5.ucInputFlag = ATOM_PLL_INPUT_FLAG_PLL_STROBE_MODE_EN;
 
-			amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+			amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args,
+				sizeof(args));
 
 			dividers->post_div = args.v5.ucPostDiv;
 			dividers->enable_post_div = (args.v5.ucCntlFlag &
@@ -1056,7 +1058,8 @@ int amdgpu_atombios_get_clock_dividers(struct amdgpu_device *adev,
 		/* fusion */
 		args.v4.ulClock = cpu_to_le32(clock);	/* 10 khz */
 
-		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args,
+			sizeof(args));
 
 		dividers->post_divider = dividers->post_div = args.v4.ucPostDiv;
 		dividers->real_clock = le32_to_cpu(args.v4.ulClock);
@@ -1067,7 +1070,8 @@ int amdgpu_atombios_get_clock_dividers(struct amdgpu_device *adev,
 		args.v6_in.ulClock.ulComputeClockFlag = clock_type;
 		args.v6_in.ulClock.ulClockFreq = cpu_to_le32(clock);	/* 10 khz */
 
-		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args,
+			sizeof(args));
 
 		dividers->whole_fb_div = le16_to_cpu(args.v6_out.ulFbDiv.usFbDiv);
 		dividers->frac_fb_div = le16_to_cpu(args.v6_out.ulFbDiv.usFbDivFrac);
@@ -1109,7 +1113,8 @@ int amdgpu_atombios_get_memory_pll_dividers(struct amdgpu_device *adev,
 			if (strobe_mode)
 				args.ucInputFlag |= MPLL_INPUT_FLAG_STROBE_MODE_EN;
 
-			amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+			amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args,
+				sizeof(args));
 
 			mpll_param->clkfrac = le16_to_cpu(args.ulFbDiv.usFbDivFrac);
 			mpll_param->clkf = le16_to_cpu(args.ulFbDiv.usFbDiv);
@@ -1151,7 +1156,8 @@ void amdgpu_atombios_set_engine_dram_timings(struct amdgpu_device *adev,
 	if (mem_clock)
 		args.sReserved.ulClock = cpu_to_le32(mem_clock & SET_CLOCK_FREQ_MASK);
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args,
+		sizeof(args));
 }
 
 void amdgpu_atombios_get_default_voltages(struct amdgpu_device *adev,
@@ -1205,7 +1211,8 @@ int amdgpu_atombios_get_max_vddc(struct amdgpu_device *adev, u8 voltage_type,
 		args.v2.ucVoltageMode = 0;
 		args.v2.usVoltageLevel = 0;
 
-		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args,
+			sizeof(args));
 
 		*voltage = le16_to_cpu(args.v2.usVoltageLevel);
 		break;
@@ -1214,7 +1221,8 @@ int amdgpu_atombios_get_max_vddc(struct amdgpu_device *adev, u8 voltage_type,
 		args.v3.ucVoltageMode = ATOM_GET_VOLTAGE_LEVEL;
 		args.v3.usVoltageLevel = cpu_to_le16(voltage_id);
 
-		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args,
+			sizeof(args));
 
 		*voltage = le16_to_cpu(args.v3.usVoltageLevel);
 		break;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
index fb2681dd6b33..6857c586ded7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.c
@@ -941,5 +941,6 @@ int amdgpu_atomfirmware_asic_init(struct amdgpu_device *adev, bool fb_reset)
 		return -EINVAL;
 	}
 
-	return amdgpu_atom_execute_table(ctx, ATOM_CMD_INIT, (uint32_t *)&asic_init_ps_v2_1);
+	return amdgpu_atom_execute_table(ctx, ATOM_CMD_INIT, (uint32_t *)&asic_init_ps_v2_1,
+		sizeof(asic_init_ps_v2_1));
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h
index c7eb2caec65a..649b5530d8ae 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atomfirmware.h
@@ -36,7 +36,7 @@ int amdgpu_atomfirmware_get_clock_info(struct amdgpu_device *adev);
 int amdgpu_atomfirmware_get_gfx_info(struct amdgpu_device *adev);
 bool amdgpu_atomfirmware_mem_ecc_supported(struct amdgpu_device *adev);
 bool amdgpu_atomfirmware_sram_ecc_supported(struct amdgpu_device *adev);
-bool amdgpu_atomfirmware_ras_rom_addr(struct amdgpu_device *adev, uint8_t* i2c_address);
+bool amdgpu_atomfirmware_ras_rom_addr(struct amdgpu_device *adev, uint8_t *i2c_address);
 bool amdgpu_atomfirmware_mem_training_supported(struct amdgpu_device *adev);
 bool amdgpu_atomfirmware_dynamic_boot_config_supported(struct amdgpu_device *adev);
 int amdgpu_atomfirmware_get_fw_reserved_fb_size(struct amdgpu_device *adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index 6adeddfb3d56..0a4b09709cfb 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -952,10 +952,10 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 	p->bytes_moved = 0;
 	p->bytes_moved_vis = 0;
 
-	r = amdgpu_vm_validate_pt_bos(p->adev, &fpriv->vm,
-				      amdgpu_cs_bo_validate, p);
+	r = amdgpu_vm_validate(p->adev, &fpriv->vm, NULL,
+			       amdgpu_cs_bo_validate, p);
 	if (r) {
-		DRM_ERROR("amdgpu_vm_validate_pt_bos() failed.\n");
+		DRM_ERROR("amdgpu_vm_validate() failed.\n");
 		goto out_free_user_pages;
 	}
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c
index 796fa6f1420b..b5ad56690a9d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_csa.c
@@ -30,7 +30,7 @@ uint64_t amdgpu_csa_vaddr(struct amdgpu_device *adev)
 {
 	uint64_t addr = adev->vm_manager.max_pfn << AMDGPU_GPU_PAGE_SHIFT;
 
-	addr -= AMDGPU_VA_RESERVED_SIZE;
+	addr -= AMDGPU_VA_RESERVED_CSA_SIZE;
 	addr = amdgpu_gmc_sign_extend(addr);
 
 	return addr;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index fdde7488d0ed..d534e192e260 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -96,6 +96,9 @@ MODULE_FIRMWARE("amdgpu/navi12_gpu_info.bin");
 #define AMDGPU_RESUME_MS		2000
 #define AMDGPU_MAX_RETRY_LIMIT		2
 #define AMDGPU_RETRY_SRIOV_RESET(r) ((r) == -EBUSY || (r) == -ETIMEDOUT || (r) == -EINVAL)
+#define AMDGPU_PCIE_INDEX_FALLBACK (0x38 >> 2)
+#define AMDGPU_PCIE_INDEX_HI_FALLBACK (0x44 >> 2)
+#define AMDGPU_PCIE_DATA_FALLBACK (0x3C >> 2)
 
 static const struct drm_driver amdgpu_kms_driver;
 
@@ -781,12 +784,22 @@ u32 amdgpu_device_indirect_rreg_ext(struct amdgpu_device *adev,
 	void __iomem *pcie_index_hi_offset;
 	void __iomem *pcie_data_offset;
 
-	pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
-	pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
-	if ((reg_addr >> 32) && (adev->nbio.funcs->get_pcie_index_hi_offset))
-		pcie_index_hi = adev->nbio.funcs->get_pcie_index_hi_offset(adev);
-	else
+	if (unlikely(!adev->nbio.funcs)) {
+		pcie_index = AMDGPU_PCIE_INDEX_FALLBACK;
+		pcie_data = AMDGPU_PCIE_DATA_FALLBACK;
+	} else {
+		pcie_index = adev->nbio.funcs->get_pcie_index_offset(adev);
+		pcie_data = adev->nbio.funcs->get_pcie_data_offset(adev);
+	}
+
+	if (reg_addr >> 32) {
+		if (unlikely(!adev->nbio.funcs))
+			pcie_index_hi = AMDGPU_PCIE_INDEX_HI_FALLBACK;
+		else
+			pcie_index_hi = adev->nbio.funcs->get_pcie_index_hi_offset(adev);
+	} else {
 		pcie_index_hi = 0;
+	}
 
 	spin_lock_irqsave(&adev->pcie_idx_lock, flags);
 	pcie_index_offset = (void __iomem *)adev->rmmio + pcie_index * 4;
@@ -1218,8 +1231,6 @@ static int amdgpu_device_asic_init(struct amdgpu_device *adev)
 	    amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(11, 0, 0)) {
 		amdgpu_psp_wait_for_bootloader(adev);
 		ret = amdgpu_atomfirmware_asic_init(adev, true);
-		/* TODO: check the return val and stop device initialization if boot fails */
-		amdgpu_psp_query_boot_status(adev);
 		return ret;
 	} else {
 		return amdgpu_atom_asic_init(adev->mode_info.atom_context);
@@ -1442,6 +1453,10 @@ int amdgpu_device_resize_fb_bar(struct amdgpu_device *adev)
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
+	/* PCI_EXT_CAP_ID_VNDR extended capability is located at 0x100 */
+	if (!pci_find_ext_capability(adev->pdev, PCI_EXT_CAP_ID_VNDR))
+		DRM_WARN("System can't access extended configuration space,please check!!\n");
+
 	/* skip if the bios has already enabled large BAR */
 	if (adev->gmc.real_vram_size &&
 	    (pci_resource_len(adev->pdev, 0) >= adev->gmc.real_vram_size))
@@ -5680,6 +5695,7 @@ retry:	/* Rest of adevs pre asic reset from XGMI hive. */
 		/* Aldebaran and gfx_11_0_3 support ras in SRIOV, so need resume ras during reset */
 		if (amdgpu_ip_version(adev, GC_HWIP, 0) ==
 			    IP_VERSION(9, 4, 2) ||
+		    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3) ||
 		    amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(11, 0, 3))
 			amdgpu_ras_resume(adev);
 	} else {
@@ -6101,6 +6117,20 @@ pci_ers_result_t amdgpu_pci_slot_reset(struct pci_dev *pdev)
 	struct amdgpu_reset_context reset_context;
 	u32 memsize;
 	struct list_head device_list;
+	struct amdgpu_hive_info *hive;
+	int hive_ras_recovery = 0;
+	struct amdgpu_ras *ras;
+
+	/* PCI error slot reset should be skipped During RAS recovery */
+	hive = amdgpu_get_xgmi_hive(adev);
+	if (hive) {
+		hive_ras_recovery = atomic_read(&hive->ras_recovery);
+		amdgpu_put_xgmi_hive(hive);
+	}
+	ras = amdgpu_ras_get_context(adev);
+	if ((amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3)) &&
+		 ras && (atomic_read(&ras->in_recovery) || hive_ras_recovery))
+		return PCI_ERS_RESULT_RECOVERED;
 
 	DRM_INFO("PCI error: slot reset callback!!\n");
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
index c7d60dd0fb97..118288b64487 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_discovery.c
@@ -27,6 +27,7 @@
 #include "amdgpu_discovery.h"
 #include "soc15_hw_ip.h"
 #include "discovery.h"
+#include "amdgpu_ras.h"
 
 #include "soc15.h"
 #include "gfx_v9_0.h"
@@ -98,6 +99,7 @@
 #define FIRMWARE_IP_DISCOVERY "amdgpu/ip_discovery.bin"
 MODULE_FIRMWARE(FIRMWARE_IP_DISCOVERY);
 
+#define mmIP_DISCOVERY_VERSION  0x16A00
 #define mmRCC_CONFIG_MEMSIZE	0xde3
 #define mmMP0_SMN_C2PMSG_33	0x16061
 #define mmMM_INDEX		0x0
@@ -518,7 +520,9 @@ static int amdgpu_discovery_init(struct amdgpu_device *adev)
 out:
 	kfree(adev->mman.discovery_bin);
 	adev->mman.discovery_bin = NULL;
-
+	if ((amdgpu_discovery != 2) &&
+	    (RREG32(mmIP_DISCOVERY_VERSION) == 4))
+		amdgpu_ras_query_boot_status(adev, 4);
 	return r;
 }
 
@@ -1278,11 +1282,10 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev)
 				 *     0b10 : encode is disabled
 				 *     0b01 : decode is disabled
 				 */
-				adev->vcn.vcn_config[adev->vcn.num_vcn_inst] =
-					ip->revision & 0xc0;
-				ip->revision &= ~0xc0;
 				if (adev->vcn.num_vcn_inst <
 				    AMDGPU_MAX_VCN_INSTANCES) {
+					adev->vcn.vcn_config[adev->vcn.num_vcn_inst] =
+						ip->revision & 0xc0;
 					adev->vcn.num_vcn_inst++;
 					adev->vcn.inst_mask |=
 						(1U << ip->instance_number);
@@ -1293,6 +1296,7 @@ static int amdgpu_discovery_reg_base_init(struct amdgpu_device *adev)
 						adev->vcn.num_vcn_inst + 1,
 						AMDGPU_MAX_VCN_INSTANCES);
 				}
+				ip->revision &= ~0xc0;
 			}
 			if (le16_to_cpu(ip->hw_id) == SDMA0_HWID ||
 			    le16_to_cpu(ip->hw_id) == SDMA1_HWID ||
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
index decbbe3d4f06..055ba2ea4c12 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_dma_buf.c
@@ -377,6 +377,10 @@ amdgpu_dma_buf_move_notify(struct dma_buf_attachment *attach)
 	struct amdgpu_vm_bo_base *bo_base;
 	int r;
 
+	/* FIXME: This should be after the "if", but needs a fix to make sure
+	 * DMABuf imports are initialized in the right VM list.
+	 */
+	amdgpu_vm_bo_invalidate(adev, bo, false);
 	if (!bo->tbo.resource || bo->tbo.resource->mem_type == TTM_PL_SYSTEM)
 		return;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 971acf01bea6..161ecf9b4174 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -366,7 +366,7 @@ module_param_named(aspm, amdgpu_aspm, int, 0444);
  * Setting the value to 0 disables this functionality.
  * Setting the value to -2 is auto enabled with power down when displays are attached.
  */
-MODULE_PARM_DESC(runpm, "PX runtime pm (2 = force enable with BAMACO, 1 = force enable with BACO, 0 = disable, -1 = auto, -2 = autowith displays)");
+MODULE_PARM_DESC(runpm, "PX runtime pm (2 = force enable with BAMACO, 1 = force enable with BACO, 0 = disable, -1 = auto, -2 = auto with displays)");
 module_param_named(runpm, amdgpu_runtime_pm, int, 0444);
 
 /**
@@ -593,7 +593,7 @@ module_param_named(timeout_period, amdgpu_watchdog_timer.period, uint, 0644);
 #ifdef CONFIG_DRM_AMDGPU_SI
 
 #if IS_ENABLED(CONFIG_DRM_RADEON) || IS_ENABLED(CONFIG_DRM_RADEON_MODULE)
-int amdgpu_si_support = 0;
+int amdgpu_si_support;
 MODULE_PARM_DESC(si_support, "SI support (1 = enabled, 0 = disabled (default))");
 #else
 int amdgpu_si_support = 1;
@@ -612,7 +612,7 @@ module_param_named(si_support, amdgpu_si_support, int, 0444);
 #ifdef CONFIG_DRM_AMDGPU_CIK
 
 #if IS_ENABLED(CONFIG_DRM_RADEON) || IS_ENABLED(CONFIG_DRM_RADEON_MODULE)
-int amdgpu_cik_support = 0;
+int amdgpu_cik_support;
 MODULE_PARM_DESC(cik_support, "CIK support (1 = enabled, 0 = disabled (default))");
 #else
 int amdgpu_cik_support = 1;
@@ -2476,6 +2476,7 @@ static int amdgpu_pmops_suspend(struct device *dev)
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(drm_dev);
 
+	adev->suspend_complete = false;
 	if (amdgpu_acpi_is_s0ix_active(adev))
 		adev->in_s0ix = true;
 	else if (amdgpu_acpi_is_s3_active(adev))
@@ -2490,6 +2491,7 @@ static int amdgpu_pmops_suspend_noirq(struct device *dev)
 	struct drm_device *drm_dev = dev_get_drvdata(dev);
 	struct amdgpu_device *adev = drm_to_adev(drm_dev);
 
+	adev->suspend_complete = true;
 	if (amdgpu_acpi_should_gpu_reset(adev))
 		return amdgpu_asic_reset(adev);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index 49a5f1c73b3e..22aeee8adb71 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -187,7 +187,34 @@ static int amdgpu_gem_object_open(struct drm_gem_object *obj,
 	else
 		++bo_va->ref_count;
 	amdgpu_bo_unreserve(abo);
-	return 0;
+
+	/* Validate and add eviction fence to DMABuf imports with dynamic
+	 * attachment in compute VMs. Re-validation will be done by
+	 * amdgpu_vm_validate. Fences are on the reservation shared with the
+	 * export, which is currently required to be validated and fenced
+	 * already by amdgpu_amdkfd_gpuvm_restore_process_bos.
+	 *
+	 * Nested locking below for the case that a GEM object is opened in
+	 * kfd_mem_export_dmabuf. Since the lock below is only taken for imports,
+	 * but not for export, this is a different lock class that cannot lead to
+	 * circular lock dependencies.
+	 */
+	if (!vm->is_compute_context || !vm->process_info)
+		return 0;
+	if (!obj->import_attach ||
+	    !dma_buf_is_dynamic(obj->import_attach->dmabuf))
+		return 0;
+	mutex_lock_nested(&vm->process_info->lock, 1);
+	if (!WARN_ON(!vm->process_info->eviction_fence)) {
+		r = amdgpu_amdkfd_bo_validate_and_fence(abo, AMDGPU_GEM_DOMAIN_GTT,
+							&vm->process_info->eviction_fence->base);
+		if (r)
+			dev_warn(adev->dev, "%d: validate_and_fence failed: %d\n",
+				 vm->task_info.pid, r);
+	}
+	mutex_unlock(&vm->process_info->lock);
+
+	return r;
 }
 
 static void amdgpu_gem_object_close(struct drm_gem_object *obj,
@@ -682,10 +709,10 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 	uint64_t vm_size;
 	int r = 0;
 
-	if (args->va_address < AMDGPU_VA_RESERVED_SIZE) {
+	if (args->va_address < AMDGPU_VA_RESERVED_BOTTOM) {
 		dev_dbg(dev->dev,
 			"va_address 0x%llx is in reserved area 0x%llx\n",
-			args->va_address, AMDGPU_VA_RESERVED_SIZE);
+			args->va_address, AMDGPU_VA_RESERVED_BOTTOM);
 		return -EINVAL;
 	}
 
@@ -701,7 +728,7 @@ int amdgpu_gem_va_ioctl(struct drm_device *dev, void *data,
 	args->va_address &= AMDGPU_GMC_HOLE_MASK;
 
 	vm_size = adev->vm_manager.max_pfn * AMDGPU_GPU_PAGE_SIZE;
-	vm_size -= AMDGPU_VA_RESERVED_SIZE;
+	vm_size -= AMDGPU_VA_RESERVED_TOP;
 	if (args->va_address + args->map_size > vm_size) {
 		dev_dbg(dev->dev,
 			"va_address 0x%llx is in top reserved area 0x%llx\n",
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
index 82b4b2019fca..78229b61c233 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c
@@ -643,8 +643,8 @@ int amdgpu_gfx_enable_kcq(struct amdgpu_device *adev, int xcc_id)
 	kiq->pmf->kiq_set_resources(kiq_ring, queue_mask);
 	for (i = 0; i < adev->gfx.num_compute_rings; i++) {
 		j = i + xcc_id * adev->gfx.num_compute_rings;
-			kiq->pmf->kiq_map_queues(kiq_ring,
-						 &adev->gfx.compute_ring[j]);
+		kiq->pmf->kiq_map_queues(kiq_ring,
+					 &adev->gfx.compute_ring[j]);
 	}
 
 	r = amdgpu_ring_test_helper(kiq_ring);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
index 55784a9f26c4..d4a848c51a83 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c
@@ -52,7 +52,7 @@ int amdgpu_gmc_pdb0_alloc(struct amdgpu_device *adev)
 	struct amdgpu_bo_param bp;
 	u64 vram_size = adev->gmc.xgmi.node_segment_size * adev->gmc.xgmi.num_physical_nodes;
 	uint32_t pde0_page_shift = adev->gmc.vmid0_page_table_block_size + 21;
-	uint32_t npdes = (vram_size + (1ULL << pde0_page_shift) -1) >> pde0_page_shift;
+	uint32_t npdes = (vram_size + (1ULL << pde0_page_shift) - 1) >> pde0_page_shift;
 
 	memset(&bp, 0, sizeof(bp));
 	bp.size = PAGE_ALIGN((npdes + 1) * 8);
@@ -746,6 +746,59 @@ error_unlock_reset:
 	return r;
 }
 
+void amdgpu_gmc_fw_reg_write_reg_wait(struct amdgpu_device *adev,
+				      uint32_t reg0, uint32_t reg1,
+				      uint32_t ref, uint32_t mask,
+				      uint32_t xcc_inst)
+{
+	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_inst];
+	struct amdgpu_ring *ring = &kiq->ring;
+	signed long r, cnt = 0;
+	unsigned long flags;
+	uint32_t seq;
+
+	if (adev->mes.ring.sched.ready) {
+		amdgpu_mes_reg_write_reg_wait(adev, reg0, reg1,
+					      ref, mask);
+		return;
+	}
+
+	spin_lock_irqsave(&kiq->ring_lock, flags);
+	amdgpu_ring_alloc(ring, 32);
+	amdgpu_ring_emit_reg_write_reg_wait(ring, reg0, reg1,
+					    ref, mask);
+	r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT);
+	if (r)
+		goto failed_undo;
+
+	amdgpu_ring_commit(ring);
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+
+	r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
+
+	/* don't wait anymore for IRQ context */
+	if (r < 1 && in_interrupt())
+		goto failed_kiq;
+
+	might_sleep();
+	while (r < 1 && cnt++ < MAX_KIQ_REG_TRY) {
+
+		msleep(MAX_KIQ_REG_BAILOUT_INTERVAL);
+		r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
+	}
+
+	if (cnt > MAX_KIQ_REG_TRY)
+		goto failed_kiq;
+
+	return;
+
+failed_undo:
+	amdgpu_ring_undo(ring);
+	spin_unlock_irqrestore(&kiq->ring_lock, flags);
+failed_kiq:
+	dev_err(adev->dev, "failed to write reg %x wait reg %x\n", reg0, reg1);
+}
+
 /**
  * amdgpu_gmc_tmz_set -- check and set if a device supports TMZ
  * @adev: amdgpu_device pointer
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h
index e699d1ca8deb..17f40ea1104b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h
@@ -417,6 +417,10 @@ void amdgpu_gmc_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 int amdgpu_gmc_flush_gpu_tlb_pasid(struct amdgpu_device *adev, uint16_t pasid,
 				   uint32_t flush_type, bool all_hub,
 				   uint32_t inst);
+void amdgpu_gmc_fw_reg_write_reg_wait(struct amdgpu_device *adev,
+				      uint32_t reg0, uint32_t reg1,
+				      uint32_t ref, uint32_t mask,
+				      uint32_t xcc_inst);
 
 extern void amdgpu_gmc_tmz_set(struct amdgpu_device *adev);
 extern void amdgpu_gmc_noretry_set(struct amdgpu_device *adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
index ddd0891da116..3d7fcdeaf8cf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ids.c
@@ -62,9 +62,8 @@ int amdgpu_pasid_alloc(unsigned int bits)
 	int pasid = -EINVAL;
 
 	for (bits = min(bits, 31U); bits > 0; bits--) {
-		pasid = ida_simple_get(&amdgpu_pasid_ida,
-				       1U << (bits - 1), 1U << bits,
-				       GFP_KERNEL);
+		pasid = ida_alloc_range(&amdgpu_pasid_ida, 1U << (bits - 1),
+					(1U << bits) - 1, GFP_KERNEL);
 		if (pasid != -ENOSPC)
 			break;
 	}
@@ -82,7 +81,7 @@ int amdgpu_pasid_alloc(unsigned int bits)
 void amdgpu_pasid_free(u32 pasid)
 {
 	trace_amdgpu_pasid_freed(pasid);
-	ida_simple_remove(&amdgpu_pasid_ida, pasid);
+	ida_free(&amdgpu_pasid_ida, pasid);
 }
 
 static void amdgpu_pasid_free_cb(struct dma_fence *fence,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
index bf4f48fe438d..a2df3025a754 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c
@@ -894,14 +894,14 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 			dev_info->ids_flags |= AMDGPU_IDS_FLAGS_CONFORMANT_TRUNC_COORD;
 
 		vm_size = adev->vm_manager.max_pfn * AMDGPU_GPU_PAGE_SIZE;
-		vm_size -= AMDGPU_VA_RESERVED_SIZE;
+		vm_size -= AMDGPU_VA_RESERVED_TOP;
 
 		/* Older VCE FW versions are buggy and can handle only 40bits */
 		if (adev->vce.fw_version &&
 		    adev->vce.fw_version < AMDGPU_VCE_FW_53_45)
 			vm_size = min(vm_size, 1ULL << 40);
 
-		dev_info->virtual_address_offset = AMDGPU_VA_RESERVED_SIZE;
+		dev_info->virtual_address_offset = AMDGPU_VA_RESERVED_BOTTOM;
 		dev_info->virtual_address_max =
 			min(vm_size, AMDGPU_GMC_HOLE_START);
 
@@ -1114,6 +1114,15 @@ int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 			}
 			ui32 >>= 8;
 			break;
+		case AMDGPU_INFO_SENSOR_GPU_INPUT_POWER:
+			/* get input GPU power */
+			if (amdgpu_dpm_read_sensor(adev,
+						   AMDGPU_PP_SENSOR_GPU_INPUT_POWER,
+						   (void *)&ui32, &ui32_size)) {
+				return -EINVAL;
+			}
+			ui32 >>= 8;
+			break;
 		case AMDGPU_INFO_SENSOR_VDDNB:
 			/* get VDDNB in millivolts */
 			if (amdgpu_dpm_read_sensor(adev,
@@ -1370,6 +1379,10 @@ int amdgpu_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 			goto error_vm;
 	}
 
+	r = amdgpu_seq64_map(adev, &fpriv->vm, &fpriv->seq64_va);
+	if (r)
+		goto error_vm;
+
 	mutex_init(&fpriv->bo_list_lock);
 	idr_init_base(&fpriv->bo_list_handles, 1);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c
index 59fafb8392e0..24ad4b97177b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.c
@@ -27,6 +27,16 @@
 #include "umc/umc_6_7_0_offset.h"
 #include "umc/umc_6_7_0_sh_mask.h"
 
+static bool amdgpu_mca_is_deferred_error(struct amdgpu_device *adev,
+					uint64_t mc_status)
+{
+	if (adev->umc.ras->check_ecc_err_status)
+		return adev->umc.ras->check_ecc_err_status(adev,
+				AMDGPU_MCA_ERROR_TYPE_DE, &mc_status);
+
+	return false;
+}
+
 void amdgpu_mca_query_correctable_error_count(struct amdgpu_device *adev,
 					      uint64_t mc_status_addr,
 					      unsigned long *error_count)
@@ -202,16 +212,16 @@ int amdgpu_mca_smu_set_debug_mode(struct amdgpu_device *adev, bool enable)
 
 static void amdgpu_mca_smu_mca_bank_dump(struct amdgpu_device *adev, int idx, struct mca_bank_entry *entry)
 {
-	dev_info(adev->dev, "[Hardware error] Accelerator Check Architecture events logged\n");
-	dev_info(adev->dev, "[Hardware error] aca entry[%02d].STATUS=0x%016llx\n",
+	dev_info(adev->dev, HW_ERR "Accelerator Check Architecture events logged\n");
+	dev_info(adev->dev, HW_ERR "aca entry[%02d].STATUS=0x%016llx\n",
 		 idx, entry->regs[MCA_REG_IDX_STATUS]);
-	dev_info(adev->dev, "[Hardware error] aca entry[%02d].ADDR=0x%016llx\n",
+	dev_info(adev->dev, HW_ERR "aca entry[%02d].ADDR=0x%016llx\n",
 		 idx, entry->regs[MCA_REG_IDX_ADDR]);
-	dev_info(adev->dev, "[Hardware error] aca entry[%02d].MISC0=0x%016llx\n",
+	dev_info(adev->dev, HW_ERR "aca entry[%02d].MISC0=0x%016llx\n",
 		 idx, entry->regs[MCA_REG_IDX_MISC0]);
-	dev_info(adev->dev, "[Hardware error] aca entry[%02d].IPID=0x%016llx\n",
+	dev_info(adev->dev, HW_ERR "aca entry[%02d].IPID=0x%016llx\n",
 		 idx, entry->regs[MCA_REG_IDX_IPID]);
-	dev_info(adev->dev, "[Hardware error] aca entry[%02d].SYND=0x%016llx\n",
+	dev_info(adev->dev, HW_ERR "aca entry[%02d].SYND=0x%016llx\n",
 		 idx, entry->regs[MCA_REG_IDX_SYND]);
 }
 
@@ -256,9 +266,14 @@ int amdgpu_mca_smu_log_ras_error(struct amdgpu_device *adev, enum amdgpu_ras_blo
 		if (type == AMDGPU_MCA_ERROR_TYPE_UE)
 			amdgpu_ras_error_statistic_ue_count(err_data,
 				&mcm_info, &err_addr, (uint64_t)count);
-		else
-			amdgpu_ras_error_statistic_ce_count(err_data,
-				&mcm_info, &err_addr, (uint64_t)count);
+		else {
+			if (amdgpu_mca_is_deferred_error(adev, entry->regs[MCA_REG_IDX_STATUS]))
+				amdgpu_ras_error_statistic_de_count(err_data,
+					&mcm_info, &err_addr, (uint64_t)count);
+			else
+				amdgpu_ras_error_statistic_ce_count(err_data,
+					&mcm_info, &err_addr, (uint64_t)count);
+		}
 	}
 
 out_mca_release:
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h
index b399f1b62887..b964110ed1e0 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h
@@ -65,6 +65,7 @@ enum amdgpu_mca_ip {
 enum amdgpu_mca_error_type {
 	AMDGPU_MCA_ERROR_TYPE_UE = 0,
 	AMDGPU_MCA_ERROR_TYPE_CE,
+	AMDGPU_MCA_ERROR_TYPE_DE,
 };
 
 struct amdgpu_mca_ras_block {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
index da48b6da0107..a98e03e0a51f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mes.c
@@ -1398,7 +1398,7 @@ int amdgpu_mes_self_test(struct amdgpu_device *adev)
 		goto error_fini;
 	}
 
-	ctx_data.meta_data_gpu_addr = AMDGPU_VA_RESERVED_SIZE;
+	ctx_data.meta_data_gpu_addr = AMDGPU_VA_RESERVED_BOTTOM;
 	r = amdgpu_mes_ctx_map_meta_data(adev, vm, &ctx_data);
 	if (r) {
 		DRM_ERROR("failed to map ctx meta data\n");
@@ -1565,9 +1565,9 @@ void amdgpu_debugfs_mes_event_log_init(struct amdgpu_device *adev)
 #if defined(CONFIG_DEBUG_FS)
 	struct drm_minor *minor = adev_to_drm(adev)->primary;
 	struct dentry *root = minor->debugfs_root;
-
-	debugfs_create_file("amdgpu_mes_event_log", 0444, root,
-			    adev, &amdgpu_debugfs_mes_event_log_fops);
+	if (adev->enable_mes)
+		debugfs_create_file("amdgpu_mes_event_log", 0444, root,
+				    adev, &amdgpu_debugfs_mes_event_log_fops);
 
 #endif
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
index 0328616473f8..d9e5eb24341d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c
@@ -291,21 +291,22 @@ static int psp_memory_training_init(struct psp_context *psp)
 	struct psp_memory_training_context *ctx = &psp->mem_train_ctx;
 
 	if (ctx->init != PSP_MEM_TRAIN_RESERVE_SUCCESS) {
-		DRM_DEBUG("memory training is not supported!\n");
+		dev_dbg(psp->adev->dev, "memory training is not supported!\n");
 		return 0;
 	}
 
 	ctx->sys_cache = kzalloc(ctx->train_data_size, GFP_KERNEL);
 	if (ctx->sys_cache == NULL) {
-		DRM_ERROR("alloc mem_train_ctx.sys_cache failed!\n");
+		dev_err(psp->adev->dev, "alloc mem_train_ctx.sys_cache failed!\n");
 		ret = -ENOMEM;
 		goto Err_out;
 	}
 
-	DRM_DEBUG("train_data_size:%llx,p2c_train_data_offset:%llx,c2p_train_data_offset:%llx.\n",
-		  ctx->train_data_size,
-		  ctx->p2c_train_data_offset,
-		  ctx->c2p_train_data_offset);
+	dev_dbg(psp->adev->dev,
+		"train_data_size:%llx,p2c_train_data_offset:%llx,c2p_train_data_offset:%llx.\n",
+		ctx->train_data_size,
+		ctx->p2c_train_data_offset,
+		ctx->c2p_train_data_offset);
 	ctx->init = PSP_MEM_TRAIN_INIT_SUCCESS;
 	return 0;
 
@@ -407,7 +408,7 @@ static int psp_sw_init(void *handle)
 
 	psp->cmd = kzalloc(sizeof(struct psp_gfx_cmd_resp), GFP_KERNEL);
 	if (!psp->cmd) {
-		DRM_ERROR("Failed to allocate memory to command buffer!\n");
+		dev_err(adev->dev, "Failed to allocate memory to command buffer!\n");
 		ret = -ENOMEM;
 	}
 
@@ -454,13 +455,13 @@ static int psp_sw_init(void *handle)
 	if (mem_training_ctx->enable_mem_training) {
 		ret = psp_memory_training_init(psp);
 		if (ret) {
-			DRM_ERROR("Failed to initialize memory training!\n");
+			dev_err(adev->dev, "Failed to initialize memory training!\n");
 			return ret;
 		}
 
 		ret = psp_mem_training(psp, PSP_MEM_TRAIN_COLD_BOOT);
 		if (ret) {
-			DRM_ERROR("Failed to process memory training!\n");
+			dev_err(adev->dev, "Failed to process memory training!\n");
 			return ret;
 		}
 	}
@@ -675,9 +676,11 @@ psp_cmd_submit_buf(struct psp_context *psp,
 	 */
 	if (!skip_unsupport && (psp->cmd_buf_mem->resp.status || !timeout) && !ras_intr) {
 		if (ucode)
-			DRM_WARN("failed to load ucode %s(0x%X) ",
-				  amdgpu_ucode_name(ucode->ucode_id), ucode->ucode_id);
-		DRM_WARN("psp gfx command %s(0x%X) failed and response status is (0x%X)\n",
+			dev_warn(psp->adev->dev,
+				 "failed to load ucode %s(0x%X) ",
+				 amdgpu_ucode_name(ucode->ucode_id), ucode->ucode_id);
+		dev_warn(psp->adev->dev,
+			 "psp gfx command %s(0x%X) failed and response status is (0x%X)\n",
 			 psp_gfx_cmd_name(psp->cmd_buf_mem->cmd_id), psp->cmd_buf_mem->cmd_id,
 			 psp->cmd_buf_mem->resp.status);
 		/* If any firmware (including CAP) load fails under SRIOV, it should
@@ -807,7 +810,7 @@ static int psp_tmr_init(struct psp_context *psp)
 	    psp->fw_pri_buf) {
 		ret = psp_load_toc(psp, &tmr_size);
 		if (ret) {
-			DRM_ERROR("Failed to load toc\n");
+			dev_err(psp->adev->dev, "Failed to load toc\n");
 			return ret;
 		}
 	}
@@ -855,7 +858,7 @@ static int psp_tmr_load(struct psp_context *psp)
 
 	psp_prep_tmr_cmd_buf(psp, cmd, psp->tmr_mc_addr, psp->tmr_bo);
 	if (psp->tmr_bo)
-		DRM_INFO("reserve 0x%lx from 0x%llx for PSP TMR\n",
+		dev_info(psp->adev->dev, "reserve 0x%lx from 0x%llx for PSP TMR\n",
 			 amdgpu_bo_size(psp->tmr_bo), psp->tmr_mc_addr);
 
 	ret = psp_cmd_submit_buf(psp, NULL, cmd,
@@ -1113,7 +1116,7 @@ int psp_reg_program(struct psp_context *psp, enum psp_reg_prog_id reg,
 	psp_prep_reg_prog_cmd_buf(cmd, reg, value);
 	ret = psp_cmd_submit_buf(psp, NULL, cmd, psp->fence_buf_mc_addr);
 	if (ret)
-		DRM_ERROR("PSP failed to program reg id %d", reg);
+		dev_err(psp->adev->dev, "PSP failed to program reg id %d\n", reg);
 
 	release_psp_cmd_buf(psp);
 
@@ -1526,22 +1529,22 @@ static void psp_ras_ta_check_status(struct psp_context *psp)
 	switch (ras_cmd->ras_status) {
 	case TA_RAS_STATUS__ERROR_UNSUPPORTED_IP:
 		dev_warn(psp->adev->dev,
-				"RAS WARNING: cmd failed due to unsupported ip\n");
+			 "RAS WARNING: cmd failed due to unsupported ip\n");
 		break;
 	case TA_RAS_STATUS__ERROR_UNSUPPORTED_ERROR_INJ:
 		dev_warn(psp->adev->dev,
-				"RAS WARNING: cmd failed due to unsupported error injection\n");
+			 "RAS WARNING: cmd failed due to unsupported error injection\n");
 		break;
 	case TA_RAS_STATUS__SUCCESS:
 		break;
 	case TA_RAS_STATUS__TEE_ERROR_ACCESS_DENIED:
 		if (ras_cmd->cmd_id == TA_RAS_COMMAND__TRIGGER_ERROR)
 			dev_warn(psp->adev->dev,
-					"RAS WARNING: Inject error to critical region is not allowed\n");
+				 "RAS WARNING: Inject error to critical region is not allowed\n");
 		break;
 	default:
 		dev_warn(psp->adev->dev,
-				"RAS WARNING: ras status = 0x%X\n", ras_cmd->ras_status);
+			 "RAS WARNING: ras status = 0x%X\n", ras_cmd->ras_status);
 		break;
 	}
 }
@@ -1565,7 +1568,7 @@ int psp_ras_invoke(struct psp_context *psp, uint32_t ta_cmd_id)
 		return ret;
 
 	if (ras_cmd->if_version > RAS_TA_HOST_IF_VER) {
-		DRM_WARN("RAS: Unsupported Interface");
+		dev_warn(psp->adev->dev, "RAS: Unsupported Interface\n");
 		return -EINVAL;
 	}
 
@@ -1715,7 +1718,7 @@ int psp_ras_initialize(struct psp_context *psp)
 		psp->ras_context.context.initialized = true;
 	else {
 		if (ras_cmd->ras_status)
-			dev_warn(psp->adev->dev, "RAS Init Status: 0x%X\n", ras_cmd->ras_status);
+			dev_warn(adev->dev, "RAS Init Status: 0x%X\n", ras_cmd->ras_status);
 
 		/* fail to load RAS TA */
 		psp->ras_context.context.initialized = false;
@@ -1779,6 +1782,31 @@ int psp_ras_trigger_error(struct psp_context *psp,
 
 	return 0;
 }
+
+int psp_ras_query_address(struct psp_context *psp,
+			  struct ta_ras_query_address_input *addr_in,
+			  struct ta_ras_query_address_output *addr_out)
+{
+	struct ta_ras_shared_memory *ras_cmd;
+	int ret;
+
+	if (!psp->ras_context.context.initialized)
+		return -EINVAL;
+
+	ras_cmd = (struct ta_ras_shared_memory *)psp->ras_context.context.mem_context.shared_buf;
+	memset(ras_cmd, 0, sizeof(struct ta_ras_shared_memory));
+
+	ras_cmd->cmd_id = TA_RAS_COMMAND__QUERY_ADDRESS;
+	ras_cmd->ras_in_message.address = *addr_in;
+
+	ret = psp_ras_invoke(psp, ras_cmd->cmd_id);
+	if (ret || ras_cmd->ras_status || psp->cmd_buf_mem->resp.status)
+		return -EINVAL;
+
+	*addr_out = ras_cmd->ras_out_message.address;
+
+	return 0;
+}
 // ras end
 
 // HDCP start
@@ -2125,19 +2153,14 @@ int amdgpu_psp_wait_for_bootloader(struct amdgpu_device *adev)
 	return ret;
 }
 
-int amdgpu_psp_query_boot_status(struct amdgpu_device *adev)
+bool amdgpu_psp_get_ras_capability(struct psp_context *psp)
 {
-	struct psp_context *psp = &adev->psp;
-	int ret = 0;
-
-	if (amdgpu_sriov_vf(adev) || (adev->flags & AMD_IS_APU))
-		return 0;
-
 	if (psp->funcs &&
-	    psp->funcs->query_boot_status)
-		ret = psp->funcs->query_boot_status(psp);
-
-	return ret;
+	    psp->funcs->get_ras_capability) {
+		return psp->funcs->get_ras_capability(psp);
+	} else {
+		return false;
+	}
 }
 
 static int psp_hw_start(struct psp_context *psp)
@@ -2150,7 +2173,7 @@ static int psp_hw_start(struct psp_context *psp)
 		    (psp->funcs->bootloader_load_kdb != NULL)) {
 			ret = psp_bootloader_load_kdb(psp);
 			if (ret) {
-				DRM_ERROR("PSP load kdb failed!\n");
+				dev_err(adev->dev, "PSP load kdb failed!\n");
 				return ret;
 			}
 		}
@@ -2159,7 +2182,7 @@ static int psp_hw_start(struct psp_context *psp)
 		    (psp->funcs->bootloader_load_spl != NULL)) {
 			ret = psp_bootloader_load_spl(psp);
 			if (ret) {
-				DRM_ERROR("PSP load spl failed!\n");
+				dev_err(adev->dev, "PSP load spl failed!\n");
 				return ret;
 			}
 		}
@@ -2168,7 +2191,7 @@ static int psp_hw_start(struct psp_context *psp)
 		    (psp->funcs->bootloader_load_sysdrv != NULL)) {
 			ret = psp_bootloader_load_sysdrv(psp);
 			if (ret) {
-				DRM_ERROR("PSP load sys drv failed!\n");
+				dev_err(adev->dev, "PSP load sys drv failed!\n");
 				return ret;
 			}
 		}
@@ -2177,7 +2200,7 @@ static int psp_hw_start(struct psp_context *psp)
 		    (psp->funcs->bootloader_load_soc_drv != NULL)) {
 			ret = psp_bootloader_load_soc_drv(psp);
 			if (ret) {
-				DRM_ERROR("PSP load soc drv failed!\n");
+				dev_err(adev->dev, "PSP load soc drv failed!\n");
 				return ret;
 			}
 		}
@@ -2186,7 +2209,7 @@ static int psp_hw_start(struct psp_context *psp)
 		    (psp->funcs->bootloader_load_intf_drv != NULL)) {
 			ret = psp_bootloader_load_intf_drv(psp);
 			if (ret) {
-				DRM_ERROR("PSP load intf drv failed!\n");
+				dev_err(adev->dev, "PSP load intf drv failed!\n");
 				return ret;
 			}
 		}
@@ -2195,7 +2218,7 @@ static int psp_hw_start(struct psp_context *psp)
 		    (psp->funcs->bootloader_load_dbg_drv != NULL)) {
 			ret = psp_bootloader_load_dbg_drv(psp);
 			if (ret) {
-				DRM_ERROR("PSP load dbg drv failed!\n");
+				dev_err(adev->dev, "PSP load dbg drv failed!\n");
 				return ret;
 			}
 		}
@@ -2204,7 +2227,7 @@ static int psp_hw_start(struct psp_context *psp)
 		    (psp->funcs->bootloader_load_ras_drv != NULL)) {
 			ret = psp_bootloader_load_ras_drv(psp);
 			if (ret) {
-				DRM_ERROR("PSP load ras_drv failed!\n");
+				dev_err(adev->dev, "PSP load ras_drv failed!\n");
 				return ret;
 			}
 		}
@@ -2213,7 +2236,7 @@ static int psp_hw_start(struct psp_context *psp)
 		    (psp->funcs->bootloader_load_sos != NULL)) {
 			ret = psp_bootloader_load_sos(psp);
 			if (ret) {
-				DRM_ERROR("PSP load sos failed!\n");
+				dev_err(adev->dev, "PSP load sos failed!\n");
 				return ret;
 			}
 		}
@@ -2221,7 +2244,7 @@ static int psp_hw_start(struct psp_context *psp)
 
 	ret = psp_ring_create(psp, PSP_RING_TYPE__KM);
 	if (ret) {
-		DRM_ERROR("PSP create ring failed!\n");
+		dev_err(adev->dev, "PSP create ring failed!\n");
 		return ret;
 	}
 
@@ -2231,7 +2254,7 @@ static int psp_hw_start(struct psp_context *psp)
 	if (!psp_boottime_tmr(psp)) {
 		ret = psp_tmr_init(psp);
 		if (ret) {
-			DRM_ERROR("PSP tmr init failed!\n");
+			dev_err(adev->dev, "PSP tmr init failed!\n");
 			return ret;
 		}
 	}
@@ -2250,7 +2273,7 @@ skip_pin_bo:
 
 	ret = psp_tmr_load(psp);
 	if (ret) {
-		DRM_ERROR("PSP load tmr failed!\n");
+		dev_err(adev->dev, "PSP load tmr failed!\n");
 		return ret;
 	}
 
@@ -2518,7 +2541,8 @@ static void psp_print_fw_hdr(struct psp_context *psp,
 	}
 }
 
-static int psp_prep_load_ip_fw_cmd_buf(struct amdgpu_firmware_info *ucode,
+static int psp_prep_load_ip_fw_cmd_buf(struct psp_context *psp,
+				       struct amdgpu_firmware_info *ucode,
 				       struct psp_gfx_cmd_resp *cmd)
 {
 	int ret;
@@ -2531,7 +2555,7 @@ static int psp_prep_load_ip_fw_cmd_buf(struct amdgpu_firmware_info *ucode,
 
 	ret = psp_get_fw_type(ucode, &cmd->cmd.cmd_load_ip_fw.fw_type);
 	if (ret)
-		DRM_ERROR("Unknown firmware type\n");
+		dev_err(psp->adev->dev, "Unknown firmware type\n");
 
 	return ret;
 }
@@ -2542,7 +2566,7 @@ int psp_execute_ip_fw_load(struct psp_context *psp,
 	int ret = 0;
 	struct psp_gfx_cmd_resp *cmd = acquire_psp_cmd_buf(psp);
 
-	ret = psp_prep_load_ip_fw_cmd_buf(ucode, cmd);
+	ret = psp_prep_load_ip_fw_cmd_buf(psp, ucode, cmd);
 	if (!ret) {
 		ret = psp_cmd_submit_buf(psp, ucode, cmd,
 					 psp->fence_buf_mc_addr);
@@ -2601,13 +2625,13 @@ static int psp_load_smu_fw(struct psp_context *psp)
 	      amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(11, 0, 2)))) {
 		ret = amdgpu_dpm_set_mp1_state(adev, PP_MP1_STATE_UNLOAD);
 		if (ret)
-			DRM_WARN("Failed to set MP1 state prepare for reload\n");
+			dev_err(adev->dev, "Failed to set MP1 state prepare for reload\n");
 	}
 
 	ret = psp_execute_ip_fw_load(psp, ucode);
 
 	if (ret)
-		DRM_ERROR("PSP load smu failed!\n");
+		dev_err(adev->dev, "PSP load smu failed!\n");
 
 	return ret;
 }
@@ -2712,7 +2736,7 @@ static int psp_load_non_psp_fw(struct psp_context *psp)
 		    adev->virt.autoload_ucode_id : AMDGPU_UCODE_ID_RLC_G)) {
 			ret = psp_rlc_autoload_start(psp);
 			if (ret) {
-				DRM_ERROR("Failed to start rlc autoload\n");
+				dev_err(adev->dev, "Failed to start rlc autoload\n");
 				return ret;
 			}
 		}
@@ -2734,7 +2758,7 @@ static int psp_load_fw(struct amdgpu_device *adev)
 
 		ret = psp_ring_init(psp, PSP_RING_TYPE__KM);
 		if (ret) {
-			DRM_ERROR("PSP ring init failed!\n");
+			dev_err(adev->dev, "PSP ring init failed!\n");
 			goto failed;
 		}
 	}
@@ -2749,13 +2773,13 @@ static int psp_load_fw(struct amdgpu_device *adev)
 
 	ret = psp_asd_initialize(psp);
 	if (ret) {
-		DRM_ERROR("PSP load asd failed!\n");
+		dev_err(adev->dev, "PSP load asd failed!\n");
 		goto failed1;
 	}
 
 	ret = psp_rl_load(adev);
 	if (ret) {
-		DRM_ERROR("PSP load RL failed!\n");
+		dev_err(adev->dev, "PSP load RL failed!\n");
 		goto failed1;
 	}
 
@@ -2775,7 +2799,7 @@ static int psp_load_fw(struct amdgpu_device *adev)
 		ret = psp_ras_initialize(psp);
 		if (ret)
 			dev_err(psp->adev->dev,
-					"RAS: Failed to initialize RAS\n");
+				"RAS: Failed to initialize RAS\n");
 
 		ret = psp_hdcp_initialize(psp);
 		if (ret)
@@ -2828,7 +2852,7 @@ static int psp_hw_init(void *handle)
 
 	ret = psp_load_fw(adev);
 	if (ret) {
-		DRM_ERROR("PSP firmware loading failed\n");
+		dev_err(adev->dev, "PSP firmware loading failed\n");
 		goto failed;
 	}
 
@@ -2875,7 +2899,7 @@ static int psp_suspend(void *handle)
 	    psp->xgmi_context.context.initialized) {
 		ret = psp_xgmi_terminate(psp);
 		if (ret) {
-			DRM_ERROR("Failed to terminate xgmi ta\n");
+			dev_err(adev->dev, "Failed to terminate xgmi ta\n");
 			goto out;
 		}
 	}
@@ -2883,46 +2907,46 @@ static int psp_suspend(void *handle)
 	if (psp->ta_fw) {
 		ret = psp_ras_terminate(psp);
 		if (ret) {
-			DRM_ERROR("Failed to terminate ras ta\n");
+			dev_err(adev->dev, "Failed to terminate ras ta\n");
 			goto out;
 		}
 		ret = psp_hdcp_terminate(psp);
 		if (ret) {
-			DRM_ERROR("Failed to terminate hdcp ta\n");
+			dev_err(adev->dev, "Failed to terminate hdcp ta\n");
 			goto out;
 		}
 		ret = psp_dtm_terminate(psp);
 		if (ret) {
-			DRM_ERROR("Failed to terminate dtm ta\n");
+			dev_err(adev->dev, "Failed to terminate dtm ta\n");
 			goto out;
 		}
 		ret = psp_rap_terminate(psp);
 		if (ret) {
-			DRM_ERROR("Failed to terminate rap ta\n");
+			dev_err(adev->dev, "Failed to terminate rap ta\n");
 			goto out;
 		}
 		ret = psp_securedisplay_terminate(psp);
 		if (ret) {
-			DRM_ERROR("Failed to terminate securedisplay ta\n");
+			dev_err(adev->dev, "Failed to terminate securedisplay ta\n");
 			goto out;
 		}
 	}
 
 	ret = psp_asd_terminate(psp);
 	if (ret) {
-		DRM_ERROR("Failed to terminate asd\n");
+		dev_err(adev->dev, "Failed to terminate asd\n");
 		goto out;
 	}
 
 	ret = psp_tmr_terminate(psp);
 	if (ret) {
-		DRM_ERROR("Failed to terminate tmr\n");
+		dev_err(adev->dev, "Failed to terminate tmr\n");
 		goto out;
 	}
 
 	ret = psp_ring_stop(psp, PSP_RING_TYPE__KM);
 	if (ret)
-		DRM_ERROR("PSP ring stop failed\n");
+		dev_err(adev->dev, "PSP ring stop failed\n");
 
 out:
 	return ret;
@@ -2934,12 +2958,12 @@ static int psp_resume(void *handle)
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	struct psp_context *psp = &adev->psp;
 
-	DRM_INFO("PSP is resuming...\n");
+	dev_info(adev->dev, "PSP is resuming...\n");
 
 	if (psp->mem_train_ctx.enable_mem_training) {
 		ret = psp_mem_training(psp, PSP_MEM_TRAIN_RESUME);
 		if (ret) {
-			DRM_ERROR("Failed to process memory training!\n");
+			dev_err(adev->dev, "Failed to process memory training!\n");
 			return ret;
 		}
 	}
@@ -2956,7 +2980,7 @@ static int psp_resume(void *handle)
 
 	ret = psp_asd_initialize(psp);
 	if (ret) {
-		DRM_ERROR("PSP load asd failed!\n");
+		dev_err(adev->dev, "PSP load asd failed!\n");
 		goto failed;
 	}
 
@@ -2980,7 +3004,7 @@ static int psp_resume(void *handle)
 		ret = psp_ras_initialize(psp);
 		if (ret)
 			dev_err(psp->adev->dev,
-					"RAS: Failed to initialize RAS\n");
+				"RAS: Failed to initialize RAS\n");
 
 		ret = psp_hdcp_initialize(psp);
 		if (ret)
@@ -3008,7 +3032,7 @@ static int psp_resume(void *handle)
 	return 0;
 
 failed:
-	DRM_ERROR("PSP resume failed\n");
+	dev_err(adev->dev, "PSP resume failed\n");
 	mutex_unlock(&adev->firmware.mutex);
 	return ret;
 }
@@ -3069,9 +3093,11 @@ int psp_ring_cmd_submit(struct psp_context *psp,
 		write_frame = ring_buffer_start + (psp_write_ptr_reg / rb_frame_size_dw);
 	/* Check invalid write_frame ptr address */
 	if ((write_frame < ring_buffer_start) || (ring_buffer_end < write_frame)) {
-		DRM_ERROR("ring_buffer_start = %p; ring_buffer_end = %p; write_frame = %p\n",
-			  ring_buffer_start, ring_buffer_end, write_frame);
-		DRM_ERROR("write_frame is pointing to address out of bounds\n");
+		dev_err(adev->dev,
+			"ring_buffer_start = %p; ring_buffer_end = %p; write_frame = %p\n",
+			ring_buffer_start, ring_buffer_end, write_frame);
+		dev_err(adev->dev,
+			"write_frame is pointing to address out of bounds\n");
 		return -EINVAL;
 	}
 
@@ -3597,7 +3623,7 @@ static ssize_t psp_usbc_pd_fw_sysfs_read(struct device *dev,
 	int ret;
 
 	if (!adev->ip_blocks[AMD_IP_BLOCK_TYPE_PSP].status.late_initialized) {
-		DRM_INFO("PSP block is not ready yet.");
+		dev_info(adev->dev, "PSP block is not ready yet\n.");
 		return -EBUSY;
 	}
 
@@ -3606,7 +3632,7 @@ static ssize_t psp_usbc_pd_fw_sysfs_read(struct device *dev,
 	mutex_unlock(&adev->psp.mutex);
 
 	if (ret) {
-		DRM_ERROR("Failed to read USBC PD FW, err = %d", ret);
+		dev_err(adev->dev, "Failed to read USBC PD FW, err = %d\n", ret);
 		return ret;
 	}
 
@@ -3628,7 +3654,7 @@ static ssize_t psp_usbc_pd_fw_sysfs_write(struct device *dev,
 	void *fw_pri_cpu_addr;
 
 	if (!adev->ip_blocks[AMD_IP_BLOCK_TYPE_PSP].status.late_initialized) {
-		DRM_INFO("PSP block is not ready yet.");
+		dev_err(adev->dev, "PSP block is not ready yet.");
 		return -EBUSY;
 	}
 
@@ -3661,7 +3687,7 @@ rel_buf:
 	release_firmware(usbc_pd_fw);
 fail:
 	if (ret) {
-		DRM_ERROR("Failed to load USBC PD FW, err = %d", ret);
+		dev_err(adev->dev, "Failed to load USBC PD FW, err = %d", ret);
 		count = ret;
 	}
 
@@ -3708,7 +3734,7 @@ static ssize_t amdgpu_psp_vbflash_write(struct file *filp, struct kobject *kobj,
 
 	/* Safeguard against memory drain */
 	if (adev->psp.vbflash_image_size > AMD_VBIOS_FILE_MAX_SIZE_B) {
-		dev_err(adev->dev, "File size cannot exceed %u", AMD_VBIOS_FILE_MAX_SIZE_B);
+		dev_err(adev->dev, "File size cannot exceed %u\n", AMD_VBIOS_FILE_MAX_SIZE_B);
 		kvfree(adev->psp.vbflash_tmp_buf);
 		adev->psp.vbflash_tmp_buf = NULL;
 		adev->psp.vbflash_image_size = 0;
@@ -3727,7 +3753,7 @@ static ssize_t amdgpu_psp_vbflash_write(struct file *filp, struct kobject *kobj,
 	adev->psp.vbflash_image_size += count;
 	mutex_unlock(&adev->psp.mutex);
 
-	dev_dbg(adev->dev, "IFWI staged for update");
+	dev_dbg(adev->dev, "IFWI staged for update\n");
 
 	return count;
 }
@@ -3747,7 +3773,7 @@ static ssize_t amdgpu_psp_vbflash_read(struct file *filp, struct kobject *kobj,
 	if (adev->psp.vbflash_image_size == 0)
 		return -EINVAL;
 
-	dev_dbg(adev->dev, "PSP IFWI flash process initiated");
+	dev_dbg(adev->dev, "PSP IFWI flash process initiated\n");
 
 	ret = amdgpu_bo_create_kernel(adev, adev->psp.vbflash_image_size,
 					AMDGPU_GPU_PAGE_SIZE,
@@ -3772,11 +3798,11 @@ rel_buf:
 	adev->psp.vbflash_image_size = 0;
 
 	if (ret) {
-		dev_err(adev->dev, "Failed to load IFWI, err = %d", ret);
+		dev_err(adev->dev, "Failed to load IFWI, err = %d\n", ret);
 		return ret;
 	}
 
-	dev_dbg(adev->dev, "PSP IFWI flash process done");
+	dev_dbg(adev->dev, "PSP IFWI flash process done\n");
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
index c4d9cbde55b9..9951bdd022de 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h
@@ -134,7 +134,7 @@ struct psp_funcs {
 	int (*update_spirom)(struct psp_context *psp, uint64_t fw_pri_mc_addr);
 	int (*vbflash_stat)(struct psp_context *psp);
 	int (*fatal_error_recovery_quirk)(struct psp_context *psp);
-	int (*query_boot_status)(struct psp_context *psp);
+	bool (*get_ras_capability)(struct psp_context *psp);
 };
 
 struct ta_funcs {
@@ -502,6 +502,9 @@ int psp_ras_enable_features(struct psp_context *psp,
 int psp_ras_trigger_error(struct psp_context *psp,
 			  struct ta_ras_trigger_error_input *info, uint32_t instance_mask);
 int psp_ras_terminate(struct psp_context *psp);
+int psp_ras_query_address(struct psp_context *psp,
+			  struct ta_ras_query_address_input *addr_in,
+			  struct ta_ras_query_address_output *addr_out);
 
 int psp_hdcp_invoke(struct psp_context *psp, uint32_t ta_cmd_id);
 int psp_dtm_invoke(struct psp_context *psp, uint32_t ta_cmd_id);
@@ -538,7 +541,5 @@ int psp_spatial_partition(struct psp_context *psp, int mode);
 int is_psp_fw_valid(struct psp_bin_desc bin);
 
 int amdgpu_psp_wait_for_bootloader(struct amdgpu_device *adev);
-
-int amdgpu_psp_query_boot_status(struct amdgpu_device *adev);
-
+bool amdgpu_psp_get_ras_capability(struct psp_context *psp);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp_ta.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp_ta.c
index 468a67b302d4..ca5c86e5f7cd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp_ta.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp_ta.c
@@ -362,7 +362,7 @@ static ssize_t ta_if_invoke_debugfs_write(struct file *fp, const char *buf, size
 		}
 	}
 
-	if (copy_to_user((char *)buf, context->mem_context.shared_buf, shared_buf_len))
+	if (copy_to_user((char *)&buf[copy_pos], context->mem_context.shared_buf, shared_buf_len))
 		ret = -EFAULT;
 
 err_free_shared_buf:
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index 31823a30dea2..46f3d1013e8c 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -39,6 +39,7 @@
 #include "nbio_v7_9.h"
 #include "atom.h"
 #include "amdgpu_reset.h"
+#include "amdgpu_psp.h"
 
 #ifdef CONFIG_X86_MCE_AMD
 #include <asm/mce.h>
@@ -73,6 +74,8 @@ const char *ras_block_string[] = {
 	"mca",
 	"vcn",
 	"jpeg",
+	"ih",
+	"mpio",
 };
 
 const char *ras_mca_block_string[] = {
@@ -94,7 +97,8 @@ const char *get_ras_block_str(struct ras_common_if *ras_block)
 	if (!ras_block)
 		return "NULL";
 
-	if (ras_block->block >= AMDGPU_RAS_BLOCK_COUNT)
+	if (ras_block->block >= AMDGPU_RAS_BLOCK_COUNT ||
+	    ras_block->block >= ARRAY_SIZE(ras_block_string))
 		return "OUT OF RANGE";
 
 	if (ras_block->block == AMDGPU_RAS_BLOCK__MCA)
@@ -116,6 +120,8 @@ const char *get_ras_block_str(struct ras_common_if *ras_block)
 /* typical ECC bad page rate is 1 bad page per 100MB VRAM */
 #define RAS_BAD_PAGE_COVER              (100 * 1024 * 1024ULL)
 
+#define MAX_UMC_POISON_POLLING_TIME_ASYNC  100  //ms
+
 enum amdgpu_ras_retire_page_reservation {
 	AMDGPU_RAS_RETIRE_PAGE_RESERVED,
 	AMDGPU_RAS_RETIRE_PAGE_PENDING,
@@ -628,8 +634,12 @@ static ssize_t amdgpu_ras_sysfs_read(struct device *dev,
 			dev_warn(obj->adev->dev, "Failed to reset error counter and error status");
 	}
 
-	return sysfs_emit(buf, "%s: %lu\n%s: %lu\n", "ue", info.ue_count,
-			  "ce", info.ce_count);
+	if (info.head.block == AMDGPU_RAS_BLOCK__UMC)
+		return sysfs_emit(buf, "%s: %lu\n%s: %lu\n%s: %lu\n", "ue", info.ue_count,
+				"ce", info.ce_count, "de", info.de_count);
+	else
+		return sysfs_emit(buf, "%s: %lu\n%s: %lu\n", "ue", info.ue_count,
+				"ce", info.ce_count);
 }
 
 /* obj begin */
@@ -1036,7 +1046,8 @@ static void amdgpu_ras_error_print_error_data(struct amdgpu_device *adev,
 					      struct ras_manager *ras_mgr,
 					      struct ras_err_data *err_data,
 					      const char *blk_name,
-					      bool is_ue)
+					      bool is_ue,
+					      bool is_de)
 {
 	struct amdgpu_smuio_mcm_config_info *mcm_info;
 	struct ras_err_node *err_node;
@@ -1065,25 +1076,50 @@ static void amdgpu_ras_error_print_error_data(struct amdgpu_device *adev,
 		}
 
 	} else {
-		for_each_ras_error(err_node, err_data) {
-			err_info = &err_node->err_info;
-			mcm_info = &err_info->mcm_info;
-			if (err_info->ce_count) {
+		if (is_de) {
+			for_each_ras_error(err_node, err_data) {
+				err_info = &err_node->err_info;
+				mcm_info = &err_info->mcm_info;
+				if (err_info->de_count) {
+					dev_info(adev->dev, "socket: %d, die: %d, "
+						"%lld new deferred hardware errors detected in %s block\n",
+						mcm_info->socket_id,
+						mcm_info->die_id,
+						err_info->de_count,
+						blk_name);
+				}
+			}
+
+			for_each_ras_error(err_node, &ras_mgr->err_data) {
+				err_info = &err_node->err_info;
+				mcm_info = &err_info->mcm_info;
 				dev_info(adev->dev, "socket: %d, die: %d, "
-					 "%lld new correctable hardware errors detected in %s block\n",
-					 mcm_info->socket_id,
-					 mcm_info->die_id,
-					 err_info->ce_count,
-					 blk_name);
+					"%lld deferred hardware errors detected in total in %s block\n",
+					mcm_info->socket_id, mcm_info->die_id,
+					err_info->de_count, blk_name);
+			}
+		} else {
+			for_each_ras_error(err_node, err_data) {
+				err_info = &err_node->err_info;
+				mcm_info = &err_info->mcm_info;
+				if (err_info->ce_count) {
+					dev_info(adev->dev, "socket: %d, die: %d, "
+						"%lld new correctable hardware errors detected in %s block\n",
+						mcm_info->socket_id,
+						mcm_info->die_id,
+						err_info->ce_count,
+						blk_name);
+				}
 			}
-		}
 
-		for_each_ras_error(err_node, &ras_mgr->err_data) {
-			err_info = &err_node->err_info;
-			mcm_info = &err_info->mcm_info;
-			dev_info(adev->dev, "socket: %d, die: %d, "
-				 "%lld correctable hardware errors detected in total in %s block\n",
-				 mcm_info->socket_id, mcm_info->die_id, err_info->ce_count, blk_name);
+			for_each_ras_error(err_node, &ras_mgr->err_data) {
+				err_info = &err_node->err_info;
+				mcm_info = &err_info->mcm_info;
+				dev_info(adev->dev, "socket: %d, die: %d, "
+					"%lld correctable hardware errors detected in total in %s block\n",
+					mcm_info->socket_id, mcm_info->die_id,
+					err_info->ce_count, blk_name);
+			}
 		}
 	}
 }
@@ -1102,7 +1138,8 @@ static void amdgpu_ras_error_generate_report(struct amdgpu_device *adev,
 
 	if (err_data->ce_count) {
 		if (err_data_has_source_info(err_data)) {
-			amdgpu_ras_error_print_error_data(adev, ras_mgr, err_data, blk_name, false);
+			amdgpu_ras_error_print_error_data(adev, ras_mgr, err_data,
+							  blk_name, false, false);
 		} else if (!adev->aid_mask &&
 			   adev->smuio.funcs &&
 			   adev->smuio.funcs->get_socket_id &&
@@ -1124,7 +1161,8 @@ static void amdgpu_ras_error_generate_report(struct amdgpu_device *adev,
 
 	if (err_data->ue_count) {
 		if (err_data_has_source_info(err_data)) {
-			amdgpu_ras_error_print_error_data(adev, ras_mgr, err_data, blk_name, true);
+			amdgpu_ras_error_print_error_data(adev, ras_mgr, err_data,
+							  blk_name, true, false);
 		} else if (!adev->aid_mask &&
 			   adev->smuio.funcs &&
 			   adev->smuio.funcs->get_socket_id &&
@@ -1144,6 +1182,28 @@ static void amdgpu_ras_error_generate_report(struct amdgpu_device *adev,
 		}
 	}
 
+	if (err_data->de_count) {
+		if (err_data_has_source_info(err_data)) {
+			amdgpu_ras_error_print_error_data(adev, ras_mgr, err_data,
+							  blk_name, false, true);
+		} else if (!adev->aid_mask &&
+			   adev->smuio.funcs &&
+			   adev->smuio.funcs->get_socket_id &&
+			   adev->smuio.funcs->get_die_id) {
+			dev_info(adev->dev, "socket: %d, die: %d "
+				 "%ld deferred hardware errors "
+				 "detected in %s block\n",
+				 adev->smuio.funcs->get_socket_id(adev),
+				 adev->smuio.funcs->get_die_id(adev),
+				 ras_mgr->err_data.de_count,
+				 blk_name);
+		} else {
+			dev_info(adev->dev, "%ld deferred hardware errors "
+				 "detected in %s block\n",
+				 ras_mgr->err_data.de_count,
+				 blk_name);
+		}
+	}
 }
 
 static void amdgpu_rasmgr_error_data_statistic_update(struct ras_manager *obj, struct ras_err_data *err_data)
@@ -1154,7 +1214,8 @@ static void amdgpu_rasmgr_error_data_statistic_update(struct ras_manager *obj, s
 	if (err_data_has_source_info(err_data)) {
 		for_each_ras_error(err_node, err_data) {
 			err_info = &err_node->err_info;
-
+			amdgpu_ras_error_statistic_de_count(&obj->err_data,
+					&err_info->mcm_info, NULL, err_info->de_count);
 			amdgpu_ras_error_statistic_ce_count(&obj->err_data,
 					&err_info->mcm_info, NULL, err_info->ce_count);
 			amdgpu_ras_error_statistic_ue_count(&obj->err_data,
@@ -1164,9 +1225,72 @@ static void amdgpu_rasmgr_error_data_statistic_update(struct ras_manager *obj, s
 		/* for legacy asic path which doesn't has error source info */
 		obj->err_data.ue_count += err_data->ue_count;
 		obj->err_data.ce_count += err_data->ce_count;
+		obj->err_data.de_count += err_data->de_count;
 	}
 }
 
+static struct ras_manager *get_ras_manager(struct amdgpu_device *adev, enum amdgpu_ras_block blk)
+{
+	struct ras_common_if head;
+
+	memset(&head, 0, sizeof(head));
+	head.block = blk;
+
+	return amdgpu_ras_find_obj(adev, &head);
+}
+
+int amdgpu_ras_bind_aca(struct amdgpu_device *adev, enum amdgpu_ras_block blk,
+			const struct aca_info *aca_info, void *data)
+{
+	struct ras_manager *obj;
+
+	obj = get_ras_manager(adev, blk);
+	if (!obj)
+		return -EINVAL;
+
+	return amdgpu_aca_add_handle(adev, &obj->aca_handle, ras_block_str(blk), aca_info, data);
+}
+
+int amdgpu_ras_unbind_aca(struct amdgpu_device *adev, enum amdgpu_ras_block blk)
+{
+	struct ras_manager *obj;
+
+	obj = get_ras_manager(adev, blk);
+	if (!obj)
+		return -EINVAL;
+
+	amdgpu_aca_remove_handle(&obj->aca_handle);
+
+	return 0;
+}
+
+static int amdgpu_aca_log_ras_error_data(struct amdgpu_device *adev, enum amdgpu_ras_block blk,
+					 enum aca_error_type type, struct ras_err_data *err_data)
+{
+	struct ras_manager *obj;
+
+	obj = get_ras_manager(adev, blk);
+	if (!obj)
+		return -EINVAL;
+
+	return amdgpu_aca_get_error_data(adev, &obj->aca_handle, type, err_data);
+}
+
+ssize_t amdgpu_ras_aca_sysfs_read(struct device *dev, struct device_attribute *attr,
+				  struct aca_handle *handle, char *buf, void *data)
+{
+	struct ras_manager *obj = container_of(handle, struct ras_manager, aca_handle);
+	struct ras_query_if info = {
+		.head = obj->head,
+	};
+
+	if (amdgpu_ras_query_error_status(obj->adev, &info))
+		return -EINVAL;
+
+	return sysfs_emit(buf, "%s: %lu\n%s: %lu\n", "ue", info.ue_count,
+			  "ce", info.ce_count);
+}
+
 static int amdgpu_ras_query_error_status_helper(struct amdgpu_device *adev,
 						struct ras_query_if *info,
 						struct ras_err_data *err_data,
@@ -1174,6 +1298,7 @@ static int amdgpu_ras_query_error_status_helper(struct amdgpu_device *adev,
 {
 	enum amdgpu_ras_block blk = info ? info->head.block : AMDGPU_RAS_BLOCK_COUNT;
 	struct amdgpu_ras_block_object *block_obj = NULL;
+	int ret;
 
 	if (blk == AMDGPU_RAS_BLOCK_COUNT)
 		return -EINVAL;
@@ -1203,9 +1328,19 @@ static int amdgpu_ras_query_error_status_helper(struct amdgpu_device *adev,
 			}
 		}
 	} else {
-		/* FIXME: add code to check return value later */
-		amdgpu_mca_smu_log_ras_error(adev, blk, AMDGPU_MCA_ERROR_TYPE_UE, err_data);
-		amdgpu_mca_smu_log_ras_error(adev, blk, AMDGPU_MCA_ERROR_TYPE_CE, err_data);
+		if (amdgpu_aca_is_enabled(adev)) {
+			ret = amdgpu_aca_log_ras_error_data(adev, blk, ACA_ERROR_TYPE_UE, err_data);
+			if (ret)
+				return ret;
+
+			ret = amdgpu_aca_log_ras_error_data(adev, blk, ACA_ERROR_TYPE_CE, err_data);
+			if (ret)
+				return ret;
+		} else {
+			/* FIXME: add code to check return value later */
+			amdgpu_mca_smu_log_ras_error(adev, blk, AMDGPU_MCA_ERROR_TYPE_UE, err_data);
+			amdgpu_mca_smu_log_ras_error(adev, blk, AMDGPU_MCA_ERROR_TYPE_CE, err_data);
+		}
 	}
 
 	return 0;
@@ -1239,6 +1374,7 @@ int amdgpu_ras_query_error_status(struct amdgpu_device *adev, struct ras_query_i
 
 	info->ue_count = obj->err_data.ue_count;
 	info->ce_count = obj->err_data.ce_count;
+	info->de_count = obj->err_data.de_count;
 
 	amdgpu_ras_error_generate_report(adev, info, &err_data);
 
@@ -1254,6 +1390,7 @@ int amdgpu_ras_reset_error_count(struct amdgpu_device *adev,
 	struct amdgpu_ras_block_object *block_obj = amdgpu_ras_get_ras_block(adev, block, 0);
 	struct amdgpu_ras *ras = amdgpu_ras_get_context(adev);
 	const struct amdgpu_mca_smu_funcs *mca_funcs = adev->mca.mca_funcs;
+	const struct aca_smu_funcs *smu_funcs = adev->aca.smu_funcs;
 	struct amdgpu_hive_info *hive;
 	int hive_ras_recovery = 0;
 
@@ -1264,7 +1401,7 @@ int amdgpu_ras_reset_error_count(struct amdgpu_device *adev,
 	}
 
 	if (!amdgpu_ras_is_supported(adev, block) ||
-	    !amdgpu_ras_get_mca_debug_mode(adev))
+	    !amdgpu_ras_get_aca_debug_mode(adev))
 		return -EOPNOTSUPP;
 
 	hive = amdgpu_get_xgmi_hive(adev);
@@ -1276,7 +1413,8 @@ int amdgpu_ras_reset_error_count(struct amdgpu_device *adev,
 	/* skip ras error reset in gpu reset */
 	if ((amdgpu_in_reset(adev) || atomic_read(&ras->in_recovery) ||
 	    hive_ras_recovery) &&
-	    mca_funcs && mca_funcs->mca_set_debug_mode)
+	    ((smu_funcs && smu_funcs->set_debug_mode) ||
+	     (mca_funcs && mca_funcs->mca_set_debug_mode)))
 		return -EOPNOTSUPP;
 
 	if (block_obj->hw_ops->reset_ras_error_count)
@@ -1772,7 +1910,10 @@ void amdgpu_ras_debugfs_create_all(struct amdgpu_device *adev)
 		}
 	}
 
-	amdgpu_mca_smu_debugfs_init(adev, dir);
+	if (amdgpu_aca_is_enabled(adev))
+		amdgpu_aca_smu_debugfs_init(adev, dir);
+	else
+		amdgpu_mca_smu_debugfs_init(adev, dir);
 }
 
 /* debugfs end */
@@ -1900,7 +2041,7 @@ static void amdgpu_ras_interrupt_poison_consumption_handler(struct ras_manager *
 		}
 	}
 
-	amdgpu_umc_poison_handler(adev, false);
+	amdgpu_umc_poison_handler(adev, obj->head.block, false);
 
 	if (block_obj->hw_ops && block_obj->hw_ops->handle_poison_consumption)
 		poison_stat = block_obj->hw_ops->handle_poison_consumption(adev);
@@ -1951,6 +2092,7 @@ static void amdgpu_ras_interrupt_umc_handler(struct ras_manager *obj,
 		 */
 		obj->err_data.ue_count += err_data.ue_count;
 		obj->err_data.ce_count += err_data.ce_count;
+		obj->err_data.de_count += err_data.de_count;
 	}
 
 	amdgpu_ras_error_data_fini(&err_data);
@@ -2520,6 +2662,32 @@ static void amdgpu_ras_validate_threshold(struct amdgpu_device *adev,
 	}
 }
 
+static int amdgpu_ras_page_retirement_thread(void *param)
+{
+	struct amdgpu_device *adev = (struct amdgpu_device *)param;
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+
+	while (!kthread_should_stop()) {
+
+		wait_event_interruptible(con->page_retirement_wq,
+				kthread_should_stop() ||
+				atomic_read(&con->page_retirement_req_cnt));
+
+		if (kthread_should_stop())
+			break;
+
+		dev_info(adev->dev, "Start processing page retirement. request:%d\n",
+			atomic_read(&con->page_retirement_req_cnt));
+
+		atomic_dec(&con->page_retirement_req_cnt);
+
+		amdgpu_umc_bad_page_polling_timeout(adev,
+				false, MAX_UMC_POISON_POLLING_TIME_ASYNC);
+	}
+
+	return 0;
+}
+
 int amdgpu_ras_recovery_init(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
@@ -2583,6 +2751,16 @@ int amdgpu_ras_recovery_init(struct amdgpu_device *adev)
 		}
 	}
 
+	mutex_init(&con->page_retirement_lock);
+	init_waitqueue_head(&con->page_retirement_wq);
+	atomic_set(&con->page_retirement_req_cnt, 0);
+	con->page_retirement_thread =
+		kthread_run(amdgpu_ras_page_retirement_thread, adev, "umc_page_retirement");
+	if (IS_ERR(con->page_retirement_thread)) {
+		con->page_retirement_thread = NULL;
+		dev_warn(adev->dev, "Failed to create umc_page_retirement thread!!!\n");
+	}
+
 #ifdef CONFIG_X86_MCE_AMD
 	if ((adev->asic_type == CHIP_ALDEBARAN) &&
 	    (adev->gmc.xgmi.connected_to_cpu))
@@ -2618,6 +2796,11 @@ static int amdgpu_ras_recovery_fini(struct amdgpu_device *adev)
 	if (!data)
 		return 0;
 
+	if (con->page_retirement_thread)
+		kthread_stop(con->page_retirement_thread);
+
+	atomic_set(&con->page_retirement_req_cnt, 0);
+
 	cancel_work_sync(&con->recovery_work);
 
 	mutex_lock(&con->recovery_lock);
@@ -2679,6 +2862,87 @@ static void amdgpu_ras_get_quirks(struct amdgpu_device *adev)
 		adev->ras_hw_enabled |= (1 << AMDGPU_RAS_BLOCK__GFX);
 }
 
+/* Query ras capablity via atomfirmware interface */
+static void amdgpu_ras_query_ras_capablity_from_vbios(struct amdgpu_device *adev)
+{
+	/* mem_ecc cap */
+	if (amdgpu_atomfirmware_mem_ecc_supported(adev)) {
+		dev_info(adev->dev, "MEM ECC is active.\n");
+		adev->ras_hw_enabled |= (1 << AMDGPU_RAS_BLOCK__UMC |
+					 1 << AMDGPU_RAS_BLOCK__DF);
+	} else {
+		dev_info(adev->dev, "MEM ECC is not presented.\n");
+	}
+
+	/* sram_ecc cap */
+	if (amdgpu_atomfirmware_sram_ecc_supported(adev)) {
+		dev_info(adev->dev, "SRAM ECC is active.\n");
+		if (!amdgpu_sriov_vf(adev))
+			adev->ras_hw_enabled |= ~(1 << AMDGPU_RAS_BLOCK__UMC |
+						  1 << AMDGPU_RAS_BLOCK__DF);
+		else
+			adev->ras_hw_enabled |= (1 << AMDGPU_RAS_BLOCK__PCIE_BIF |
+						 1 << AMDGPU_RAS_BLOCK__SDMA |
+						 1 << AMDGPU_RAS_BLOCK__GFX);
+
+		/*
+		 * VCN/JPEG RAS can be supported on both bare metal and
+		 * SRIOV environment
+		 */
+		if (amdgpu_ip_version(adev, VCN_HWIP, 0) == IP_VERSION(2, 6, 0) ||
+		    amdgpu_ip_version(adev, VCN_HWIP, 0) == IP_VERSION(4, 0, 0) ||
+		    amdgpu_ip_version(adev, VCN_HWIP, 0) == IP_VERSION(4, 0, 3))
+			adev->ras_hw_enabled |= (1 << AMDGPU_RAS_BLOCK__VCN |
+						 1 << AMDGPU_RAS_BLOCK__JPEG);
+		else
+			adev->ras_hw_enabled &= ~(1 << AMDGPU_RAS_BLOCK__VCN |
+						  1 << AMDGPU_RAS_BLOCK__JPEG);
+
+		/*
+		 * XGMI RAS is not supported if xgmi num physical nodes
+		 * is zero
+		 */
+		if (!adev->gmc.xgmi.num_physical_nodes)
+			adev->ras_hw_enabled &= ~(1 << AMDGPU_RAS_BLOCK__XGMI_WAFL);
+	} else {
+		dev_info(adev->dev, "SRAM ECC is not presented.\n");
+	}
+}
+
+/* Query poison mode from umc/df IP callbacks */
+static void amdgpu_ras_query_poison_mode(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	bool df_poison, umc_poison;
+
+	/* poison setting is useless on SRIOV guest */
+	if (amdgpu_sriov_vf(adev) || !con)
+		return;
+
+	/* Init poison supported flag, the default value is false */
+	if (adev->gmc.xgmi.connected_to_cpu ||
+	    adev->gmc.is_app_apu) {
+		/* enabled by default when GPU is connected to CPU */
+		con->poison_supported = true;
+	} else if (adev->df.funcs &&
+	    adev->df.funcs->query_ras_poison_mode &&
+	    adev->umc.ras &&
+	    adev->umc.ras->query_ras_poison_mode) {
+		df_poison =
+			adev->df.funcs->query_ras_poison_mode(adev);
+		umc_poison =
+			adev->umc.ras->query_ras_poison_mode(adev);
+
+		/* Only poison is set in both DF and UMC, we can support it */
+		if (df_poison && umc_poison)
+			con->poison_supported = true;
+		else if (df_poison != umc_poison)
+			dev_warn(adev->dev,
+				"Poison setting is inconsistent in DF/UMC(%d:%d)!\n",
+				df_poison, umc_poison);
+	}
+}
+
 /*
  * check hardware's ras ability which will be saved in hw_supported.
  * if hardware does not support ras, we can skip some ras initializtion and
@@ -2695,49 +2959,13 @@ static void amdgpu_ras_check_supported(struct amdgpu_device *adev)
 	if (!amdgpu_ras_asic_supported(adev))
 		return;
 
-	if (!adev->gmc.xgmi.connected_to_cpu &&	!adev->gmc.is_app_apu) {
-		if (amdgpu_atomfirmware_mem_ecc_supported(adev)) {
-			dev_info(adev->dev, "MEM ECC is active.\n");
-			adev->ras_hw_enabled |= (1 << AMDGPU_RAS_BLOCK__UMC |
-						   1 << AMDGPU_RAS_BLOCK__DF);
-		} else {
-			dev_info(adev->dev, "MEM ECC is not presented.\n");
-		}
-
-		if (amdgpu_atomfirmware_sram_ecc_supported(adev)) {
-			dev_info(adev->dev, "SRAM ECC is active.\n");
-			if (!amdgpu_sriov_vf(adev))
-				adev->ras_hw_enabled |= ~(1 << AMDGPU_RAS_BLOCK__UMC |
-							    1 << AMDGPU_RAS_BLOCK__DF);
-			else
-				adev->ras_hw_enabled |= (1 << AMDGPU_RAS_BLOCK__PCIE_BIF |
-								1 << AMDGPU_RAS_BLOCK__SDMA |
-								1 << AMDGPU_RAS_BLOCK__GFX);
-
-			/* VCN/JPEG RAS can be supported on both bare metal and
-			 * SRIOV environment
-			 */
-			if (amdgpu_ip_version(adev, VCN_HWIP, 0) ==
-				    IP_VERSION(2, 6, 0) ||
-			    amdgpu_ip_version(adev, VCN_HWIP, 0) ==
-				    IP_VERSION(4, 0, 0) ||
-			    amdgpu_ip_version(adev, VCN_HWIP, 0) ==
-				    IP_VERSION(4, 0, 3))
-				adev->ras_hw_enabled |= (1 << AMDGPU_RAS_BLOCK__VCN |
-							1 << AMDGPU_RAS_BLOCK__JPEG);
-			else
-				adev->ras_hw_enabled &= ~(1 << AMDGPU_RAS_BLOCK__VCN |
-							1 << AMDGPU_RAS_BLOCK__JPEG);
+	/* query ras capability from psp */
+	if (amdgpu_psp_get_ras_capability(&adev->psp))
+		goto init_ras_enabled_flag;
 
-			/*
-			 * XGMI RAS is not supported if xgmi num physical nodes
-			 * is zero
-			 */
-			if (!adev->gmc.xgmi.num_physical_nodes)
-				adev->ras_hw_enabled &= ~(1 << AMDGPU_RAS_BLOCK__XGMI_WAFL);
-		} else {
-			dev_info(adev->dev, "SRAM ECC is not presented.\n");
-		}
+	/* query ras capablity from bios */
+	if (!adev->gmc.xgmi.connected_to_cpu && !adev->gmc.is_app_apu) {
+		amdgpu_ras_query_ras_capablity_from_vbios(adev);
 	} else {
 		/* driver only manages a few IP blocks RAS feature
 		 * when GPU is connected cpu through XGMI */
@@ -2746,13 +2974,21 @@ static void amdgpu_ras_check_supported(struct amdgpu_device *adev)
 					   1 << AMDGPU_RAS_BLOCK__MMHUB);
 	}
 
+	/* apply asic specific settings (vega20 only for now) */
 	amdgpu_ras_get_quirks(adev);
 
+	/* query poison mode from umc/df ip callback */
+	amdgpu_ras_query_poison_mode(adev);
+
+init_ras_enabled_flag:
 	/* hw_supported needs to be aligned with RAS block mask. */
 	adev->ras_hw_enabled &= AMDGPU_RAS_BLOCK_MASK;
 
 	adev->ras_enabled = amdgpu_ras_enable == 0 ? 0 :
 		adev->ras_hw_enabled & amdgpu_ras_mask;
+
+	/* aca is disabled by default */
+	adev->aca.is_enabled = false;
 }
 
 static void amdgpu_ras_counte_dw(struct work_struct *work)
@@ -2780,39 +3016,6 @@ Out:
 	pm_runtime_put_autosuspend(dev->dev);
 }
 
-static void amdgpu_ras_query_poison_mode(struct amdgpu_device *adev)
-{
-	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
-	bool df_poison, umc_poison;
-
-	/* poison setting is useless on SRIOV guest */
-	if (amdgpu_sriov_vf(adev) || !con)
-		return;
-
-	/* Init poison supported flag, the default value is false */
-	if (adev->gmc.xgmi.connected_to_cpu ||
-	    adev->gmc.is_app_apu) {
-		/* enabled by default when GPU is connected to CPU */
-		con->poison_supported = true;
-	} else if (adev->df.funcs &&
-	    adev->df.funcs->query_ras_poison_mode &&
-	    adev->umc.ras &&
-	    adev->umc.ras->query_ras_poison_mode) {
-		df_poison =
-			adev->df.funcs->query_ras_poison_mode(adev);
-		umc_poison =
-			adev->umc.ras->query_ras_poison_mode(adev);
-
-		/* Only poison is set in both DF and UMC, we can support it */
-		if (df_poison && umc_poison)
-			con->poison_supported = true;
-		else if (df_poison != umc_poison)
-			dev_warn(adev->dev,
-				"Poison setting is inconsistent in DF/UMC(%d:%d)!\n",
-				df_poison, umc_poison);
-	}
-}
-
 static int amdgpu_get_ras_schema(struct amdgpu_device *adev)
 {
 	return  amdgpu_ras_is_poison_mode_supported(adev) ? AMDGPU_RAS_ERROR__POISON : 0 |
@@ -2917,12 +3120,11 @@ int amdgpu_ras_init(struct amdgpu_device *adev)
 			goto release_con;
 	}
 
-	amdgpu_ras_query_poison_mode(adev);
-
 	/* Packed socket_id to ras feature mask bits[31:29] */
 	if (adev->smuio.funcs &&
 	    adev->smuio.funcs->get_socket_id)
-		con->features |= ((adev->smuio.funcs->get_socket_id(adev)) << 29);
+		con->features |= ((adev->smuio.funcs->get_socket_id(adev)) <<
+					AMDGPU_RAS_FEATURES_SOCKETID_SHIFT);
 
 	/* Get RAS schema for particular SOC */
 	con->schema = amdgpu_get_ras_schema(adev);
@@ -3128,7 +3330,7 @@ void amdgpu_ras_suspend(struct amdgpu_device *adev)
 
 	amdgpu_ras_disable_all_features(adev, 0);
 	/* Make sure all ras objects are disabled. */
-	if (con->features)
+	if (AMDGPU_RAS_GET_FEATURES(con->features))
 		amdgpu_ras_disable_all_features(adev, 1);
 }
 
@@ -3142,15 +3344,29 @@ int amdgpu_ras_late_init(struct amdgpu_device *adev)
 	if (amdgpu_sriov_vf(adev))
 		return 0;
 
-	amdgpu_ras_set_mca_debug_mode(adev, false);
+	if (amdgpu_aca_is_enabled(adev)) {
+		if (amdgpu_in_reset(adev))
+			r = amdgpu_aca_reset(adev);
+		 else
+			r = amdgpu_aca_init(adev);
+		if (r)
+			return r;
+
+		amdgpu_ras_set_aca_debug_mode(adev, false);
+	} else {
+		amdgpu_ras_set_mca_debug_mode(adev, false);
+	}
 
 	list_for_each_entry_safe(node, tmp, &adev->ras_list, node) {
-		if (!node->ras_obj) {
+		obj = node->ras_obj;
+		if (!obj) {
 			dev_warn(adev->dev, "Warning: abnormal ras list node.\n");
 			continue;
 		}
 
-		obj = node->ras_obj;
+		if (!amdgpu_ras_is_supported(adev, obj->ras_comm.block))
+			continue;
+
 		if (obj->ras_late_init) {
 			r = obj->ras_late_init(adev, &obj->ras_comm);
 			if (r) {
@@ -3175,7 +3391,7 @@ int amdgpu_ras_pre_fini(struct amdgpu_device *adev)
 
 
 	/* Need disable ras on all IPs here before ip [hw/sw]fini */
-	if (con->features)
+	if (AMDGPU_RAS_GET_FEATURES(con->features))
 		amdgpu_ras_disable_all_features(adev, 0);
 	amdgpu_ras_recovery_fini(adev);
 	return 0;
@@ -3208,10 +3424,13 @@ int amdgpu_ras_fini(struct amdgpu_device *adev)
 	amdgpu_ras_fs_fini(adev);
 	amdgpu_ras_interrupt_remove_all(adev);
 
-	WARN(con->features, "Feature mask is not cleared");
+	if (amdgpu_aca_is_enabled(adev))
+		amdgpu_aca_fini(adev);
 
-	if (con->features)
-		amdgpu_ras_disable_all_features(adev, 1);
+	WARN(AMDGPU_RAS_GET_FEATURES(con->features), "Feature mask is not cleared");
+
+	if (AMDGPU_RAS_GET_FEATURES(con->features))
+		amdgpu_ras_disable_all_features(adev, 0);
 
 	cancel_delayed_work_sync(&con->ras_counte_delay_work);
 
@@ -3425,22 +3644,41 @@ int amdgpu_ras_set_mca_debug_mode(struct amdgpu_device *adev, bool enable)
 	if (con) {
 		ret = amdgpu_mca_smu_set_debug_mode(adev, enable);
 		if (!ret)
-			con->is_mca_debug_mode = enable;
+			con->is_aca_debug_mode = enable;
+	}
+
+	return ret;
+}
+
+int amdgpu_ras_set_aca_debug_mode(struct amdgpu_device *adev, bool enable)
+{
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	int ret = 0;
+
+	if (con) {
+		if (amdgpu_aca_is_enabled(adev))
+			ret = amdgpu_aca_smu_set_debug_mode(adev, enable);
+		else
+			ret = amdgpu_mca_smu_set_debug_mode(adev, enable);
+		if (!ret)
+			con->is_aca_debug_mode = enable;
 	}
 
 	return ret;
 }
 
-bool amdgpu_ras_get_mca_debug_mode(struct amdgpu_device *adev)
+bool amdgpu_ras_get_aca_debug_mode(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	const struct aca_smu_funcs *smu_funcs = adev->aca.smu_funcs;
 	const struct amdgpu_mca_smu_funcs *mca_funcs = adev->mca.mca_funcs;
 
 	if (!con)
 		return false;
 
-	if (mca_funcs && mca_funcs->mca_set_debug_mode)
-		return con->is_mca_debug_mode;
+	if ((amdgpu_aca_is_enabled(adev) && smu_funcs && smu_funcs->set_debug_mode) ||
+	    (!amdgpu_aca_is_enabled(adev) && mca_funcs && mca_funcs->mca_set_debug_mode))
+		return con->is_aca_debug_mode;
 	else
 		return true;
 }
@@ -3450,15 +3688,16 @@ bool amdgpu_ras_get_error_query_mode(struct amdgpu_device *adev,
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	const struct amdgpu_mca_smu_funcs *mca_funcs = adev->mca.mca_funcs;
+	const struct aca_smu_funcs *smu_funcs = adev->aca.smu_funcs;
 
 	if (!con) {
 		*error_query_mode = AMDGPU_RAS_INVALID_ERROR_QUERY;
 		return false;
 	}
 
-	if (mca_funcs && mca_funcs->mca_set_debug_mode)
+	if ((smu_funcs && smu_funcs->set_debug_mode) || (mca_funcs && mca_funcs->mca_set_debug_mode))
 		*error_query_mode =
-			(con->is_mca_debug_mode) ? AMDGPU_RAS_DIRECT_ERROR_QUERY : AMDGPU_RAS_FIRMWARE_ERROR_QUERY;
+			(con->is_aca_debug_mode) ? AMDGPU_RAS_DIRECT_ERROR_QUERY : AMDGPU_RAS_FIRMWARE_ERROR_QUERY;
 	else
 		*error_query_mode = AMDGPU_RAS_DIRECT_ERROR_QUERY;
 
@@ -3699,8 +3938,7 @@ static int ras_err_info_cmp(void *priv, const struct list_head *a, const struct
 }
 
 static struct ras_err_info *amdgpu_ras_error_get_info(struct ras_err_data *err_data,
-				struct amdgpu_smuio_mcm_config_info *mcm_info,
-				struct ras_err_addr *err_addr)
+				struct amdgpu_smuio_mcm_config_info *mcm_info)
 {
 	struct ras_err_node *err_node;
 
@@ -3712,10 +3950,9 @@ static struct ras_err_info *amdgpu_ras_error_get_info(struct ras_err_data *err_d
 	if (!err_node)
 		return NULL;
 
-	memcpy(&err_node->err_info.mcm_info, mcm_info, sizeof(*mcm_info));
+	INIT_LIST_HEAD(&err_node->err_info.err_addr_list);
 
-	if (err_addr)
-		memcpy(&err_node->err_info.err_addr, err_addr, sizeof(*err_addr));
+	memcpy(&err_node->err_info.mcm_info, mcm_info, sizeof(*mcm_info));
 
 	err_data->err_list_count++;
 	list_add_tail(&err_node->node, &err_data->err_node_list);
@@ -3724,6 +3961,29 @@ static struct ras_err_info *amdgpu_ras_error_get_info(struct ras_err_data *err_d
 	return &err_node->err_info;
 }
 
+void amdgpu_ras_add_mca_err_addr(struct ras_err_info *err_info, struct ras_err_addr *err_addr)
+{
+	struct ras_err_addr *mca_err_addr;
+
+	mca_err_addr = kzalloc(sizeof(*mca_err_addr), GFP_KERNEL);
+	if (!mca_err_addr)
+		return;
+
+	INIT_LIST_HEAD(&mca_err_addr->node);
+
+	mca_err_addr->err_status = err_addr->err_status;
+	mca_err_addr->err_ipid = err_addr->err_ipid;
+	mca_err_addr->err_addr = err_addr->err_addr;
+
+	list_add_tail(&mca_err_addr->node, &err_info->err_addr_list);
+}
+
+void amdgpu_ras_del_mca_err_addr(struct ras_err_info *err_info, struct ras_err_addr *mca_err_addr)
+{
+	list_del(&mca_err_addr->node);
+	kfree(mca_err_addr);
+}
+
 int amdgpu_ras_error_statistic_ue_count(struct ras_err_data *err_data,
 		struct amdgpu_smuio_mcm_config_info *mcm_info,
 		struct ras_err_addr *err_addr, u64 count)
@@ -3736,10 +3996,13 @@ int amdgpu_ras_error_statistic_ue_count(struct ras_err_data *err_data,
 	if (!count)
 		return 0;
 
-	err_info = amdgpu_ras_error_get_info(err_data, mcm_info, err_addr);
+	err_info = amdgpu_ras_error_get_info(err_data, mcm_info);
 	if (!err_info)
 		return -EINVAL;
 
+	if (err_addr && err_addr->err_status)
+		amdgpu_ras_add_mca_err_addr(err_info, err_addr);
+
 	err_info->ue_count += count;
 	err_data->ue_count += count;
 
@@ -3758,7 +4021,7 @@ int amdgpu_ras_error_statistic_ce_count(struct ras_err_data *err_data,
 	if (!count)
 		return 0;
 
-	err_info = amdgpu_ras_error_get_info(err_data, mcm_info, err_addr);
+	err_info = amdgpu_ras_error_get_info(err_data, mcm_info);
 	if (!err_info)
 		return -EINVAL;
 
@@ -3767,3 +4030,135 @@ int amdgpu_ras_error_statistic_ce_count(struct ras_err_data *err_data,
 
 	return 0;
 }
+
+int amdgpu_ras_error_statistic_de_count(struct ras_err_data *err_data,
+		struct amdgpu_smuio_mcm_config_info *mcm_info,
+		struct ras_err_addr *err_addr, u64 count)
+{
+	struct ras_err_info *err_info;
+
+	if (!err_data || !mcm_info)
+		return -EINVAL;
+
+	if (!count)
+		return 0;
+
+	err_info = amdgpu_ras_error_get_info(err_data, mcm_info);
+	if (!err_info)
+		return -EINVAL;
+
+	if (err_addr && err_addr->err_status)
+		amdgpu_ras_add_mca_err_addr(err_info, err_addr);
+
+	err_info->de_count += count;
+	err_data->de_count += count;
+
+	return 0;
+}
+
+#define mmMP0_SMN_C2PMSG_92	0x1609C
+#define mmMP0_SMN_C2PMSG_126	0x160BE
+static void amdgpu_ras_boot_time_error_reporting(struct amdgpu_device *adev,
+						 u32 instance, u32 boot_error)
+{
+	u32 socket_id, aid_id, hbm_id;
+	u32 reg_data;
+	u64 reg_addr;
+
+	socket_id = AMDGPU_RAS_GPU_ERR_SOCKET_ID(boot_error);
+	aid_id = AMDGPU_RAS_GPU_ERR_AID_ID(boot_error);
+	hbm_id = AMDGPU_RAS_GPU_ERR_HBM_ID(boot_error);
+
+	/* The pattern for smn addressing in other SOC could be different from
+	 * the one for aqua_vanjaram. We should revisit the code if the pattern
+	 * is changed. In such case, replace the aqua_vanjaram implementation
+	 * with more common helper */
+	reg_addr = (mmMP0_SMN_C2PMSG_92 << 2) +
+		   aqua_vanjaram_encode_ext_smn_addressing(instance);
+
+	reg_data = amdgpu_device_indirect_rreg_ext(adev, reg_addr);
+	dev_err(adev->dev, "socket: %d, aid: %d, firmware boot failed, fw status is 0x%x\n",
+		socket_id, aid_id, reg_data);
+
+	if (AMDGPU_RAS_GPU_ERR_MEM_TRAINING(boot_error))
+		dev_info(adev->dev, "socket: %d, aid: %d, hbm: %d, memory training failed\n",
+			 socket_id, aid_id, hbm_id);
+
+	if (AMDGPU_RAS_GPU_ERR_FW_LOAD(boot_error))
+		dev_info(adev->dev, "socket: %d, aid: %d, firmware load failed at boot time\n",
+			 socket_id, aid_id);
+
+	if (AMDGPU_RAS_GPU_ERR_WAFL_LINK_TRAINING(boot_error))
+		dev_info(adev->dev, "socket: %d, aid: %d, wafl link training failed\n",
+			 socket_id, aid_id);
+
+	if (AMDGPU_RAS_GPU_ERR_XGMI_LINK_TRAINING(boot_error))
+		dev_info(adev->dev, "socket: %d, aid: %d, xgmi link training failed\n",
+			 socket_id, aid_id);
+
+	if (AMDGPU_RAS_GPU_ERR_USR_CP_LINK_TRAINING(boot_error))
+		dev_info(adev->dev, "socket: %d, aid: %d, usr cp link training failed\n",
+			 socket_id, aid_id);
+
+	if (AMDGPU_RAS_GPU_ERR_USR_DP_LINK_TRAINING(boot_error))
+		dev_info(adev->dev, "socket: %d, aid: %d, usr dp link training failed\n",
+			 socket_id, aid_id);
+
+	if (AMDGPU_RAS_GPU_ERR_HBM_MEM_TEST(boot_error))
+		dev_info(adev->dev, "socket: %d, aid: %d, hbm: %d, hbm memory test failed\n",
+			 socket_id, aid_id, hbm_id);
+
+	if (AMDGPU_RAS_GPU_ERR_HBM_BIST_TEST(boot_error))
+		dev_info(adev->dev, "socket: %d, aid: %d, hbm: %d, hbm bist test failed\n",
+			 socket_id, aid_id, hbm_id);
+}
+
+static int amdgpu_ras_wait_for_boot_complete(struct amdgpu_device *adev,
+					     u32 instance, u32 *boot_error)
+{
+	u32 reg_addr;
+	u32 reg_data;
+	int retry_loop;
+
+	reg_addr = (mmMP0_SMN_C2PMSG_92 << 2) +
+		   aqua_vanjaram_encode_ext_smn_addressing(instance);
+
+	for (retry_loop = 0; retry_loop < AMDGPU_RAS_BOOT_STATUS_POLLING_LIMIT; retry_loop++) {
+		reg_data = amdgpu_device_indirect_rreg_ext(adev, reg_addr);
+		if ((reg_data & AMDGPU_RAS_BOOT_STATUS_MASK) == AMDGPU_RAS_BOOT_STEADY_STATUS) {
+			*boot_error = AMDGPU_RAS_BOOT_SUCEESS;
+			return 0;
+		}
+		msleep(1);
+	}
+
+	/* The pattern for smn addressing in other SOC could be different from
+	 * the one for aqua_vanjaram. We should revisit the code if the pattern
+	 * is changed. In such case, replace the aqua_vanjaram implementation
+	 * with more common helper */
+	reg_addr = (mmMP0_SMN_C2PMSG_126 << 2) +
+		   aqua_vanjaram_encode_ext_smn_addressing(instance);
+
+	for (retry_loop = 0; retry_loop < AMDGPU_RAS_BOOT_STATUS_POLLING_LIMIT; retry_loop++) {
+		reg_data = amdgpu_device_indirect_rreg_ext(adev, reg_addr);
+		if (AMDGPU_RAS_GPU_ERR_BOOT_STATUS(reg_data)) {
+			*boot_error = reg_data;
+			return 0;
+		}
+		msleep(1);
+	}
+
+	*boot_error = reg_data;
+	return -ETIME;
+}
+
+void amdgpu_ras_query_boot_status(struct amdgpu_device *adev, u32 num_instances)
+{
+	u32 boot_error = 0;
+	u32 i;
+
+	for (i = 0; i < num_instances; i++) {
+		if (amdgpu_ras_wait_for_boot_complete(adev, i, &boot_error))
+			amdgpu_ras_boot_time_error_reporting(adev, i, boot_error);
+	}
+}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
index 76fb85628716..d10e5bb0e52f 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
@@ -29,9 +29,28 @@
 #include "ta_ras_if.h"
 #include "amdgpu_ras_eeprom.h"
 #include "amdgpu_smuio.h"
+#include "amdgpu_aca.h"
 
 struct amdgpu_iv_entry;
 
+#define AMDGPU_RAS_GPU_ERR_MEM_TRAINING(x)		AMDGPU_GET_REG_FIELD(x, 0, 0)
+#define AMDGPU_RAS_GPU_ERR_FW_LOAD(x)			AMDGPU_GET_REG_FIELD(x, 1, 1)
+#define AMDGPU_RAS_GPU_ERR_WAFL_LINK_TRAINING(x)	AMDGPU_GET_REG_FIELD(x, 2, 2)
+#define AMDGPU_RAS_GPU_ERR_XGMI_LINK_TRAINING(x)	AMDGPU_GET_REG_FIELD(x, 3, 3)
+#define AMDGPU_RAS_GPU_ERR_USR_CP_LINK_TRAINING(x)	AMDGPU_GET_REG_FIELD(x, 4, 4)
+#define AMDGPU_RAS_GPU_ERR_USR_DP_LINK_TRAINING(x)	AMDGPU_GET_REG_FIELD(x, 5, 5)
+#define AMDGPU_RAS_GPU_ERR_HBM_MEM_TEST(x)		AMDGPU_GET_REG_FIELD(x, 6, 6)
+#define AMDGPU_RAS_GPU_ERR_HBM_BIST_TEST(x)		AMDGPU_GET_REG_FIELD(x, 7, 7)
+#define AMDGPU_RAS_GPU_ERR_SOCKET_ID(x)			AMDGPU_GET_REG_FIELD(x, 10, 8)
+#define AMDGPU_RAS_GPU_ERR_AID_ID(x)			AMDGPU_GET_REG_FIELD(x, 12, 11)
+#define AMDGPU_RAS_GPU_ERR_HBM_ID(x)			AMDGPU_GET_REG_FIELD(x, 13, 13)
+#define AMDGPU_RAS_GPU_ERR_BOOT_STATUS(x)		AMDGPU_GET_REG_FIELD(x, 31, 31)
+
+#define AMDGPU_RAS_BOOT_STATUS_POLLING_LIMIT	1000
+#define AMDGPU_RAS_BOOT_STEADY_STATUS		0xBA
+#define AMDGPU_RAS_BOOT_STATUS_MASK		0xFF
+#define AMDGPU_RAS_BOOT_SUCEESS			0x80000000
+
 #define AMDGPU_RAS_FLAG_INIT_BY_VBIOS		(0x1 << 0)
 /* position of instance value in sub_block_index of
  * ta_ras_trigger_error_input, the sub block uses lower 12 bits
@@ -39,6 +58,12 @@ struct amdgpu_iv_entry;
 #define AMDGPU_RAS_INST_MASK 0xfffff000
 #define AMDGPU_RAS_INST_SHIFT 0xc
 
+#define AMDGPU_RAS_FEATURES_SOCKETID_SHIFT 29
+#define AMDGPU_RAS_FEATURES_SOCKETID_MASK 0xe0000000
+
+/* The high three bits indicates socketid */
+#define AMDGPU_RAS_GET_FEATURES(val)  ((val) & ~AMDGPU_RAS_FEATURES_SOCKETID_MASK)
+
 enum amdgpu_ras_block {
 	AMDGPU_RAS_BLOCK__UMC = 0,
 	AMDGPU_RAS_BLOCK__SDMA,
@@ -57,6 +82,8 @@ enum amdgpu_ras_block {
 	AMDGPU_RAS_BLOCK__MCA,
 	AMDGPU_RAS_BLOCK__VCN,
 	AMDGPU_RAS_BLOCK__JPEG,
+	AMDGPU_RAS_BLOCK__IH,
+	AMDGPU_RAS_BLOCK__MPIO,
 
 	AMDGPU_RAS_BLOCK__LAST
 };
@@ -441,10 +468,15 @@ struct amdgpu_ras {
 	/* Indicates smu whether need update bad channel info */
 	bool update_channel_flag;
 	/* Record status of smu mca debug mode */
-	bool is_mca_debug_mode;
+	bool is_aca_debug_mode;
 
 	/* Record special requirements of gpu reset caller */
 	uint32_t  gpu_reset_flags;
+
+	struct task_struct *page_retirement_thread;
+	wait_queue_head_t page_retirement_wq;
+	struct mutex page_retirement_lock;
+	atomic_t page_retirement_req_cnt;
 };
 
 struct ras_fs_data {
@@ -453,6 +485,7 @@ struct ras_fs_data {
 };
 
 struct ras_err_addr {
+	struct list_head node;
 	uint64_t err_status;
 	uint64_t err_ipid;
 	uint64_t err_addr;
@@ -462,7 +495,8 @@ struct ras_err_info {
 	struct amdgpu_smuio_mcm_config_info mcm_info;
 	u64 ce_count;
 	u64 ue_count;
-	struct ras_err_addr err_addr;
+	u64 de_count;
+	struct list_head err_addr_list;
 };
 
 struct ras_err_node {
@@ -473,6 +507,7 @@ struct ras_err_node {
 struct ras_err_data {
 	unsigned long ue_count;
 	unsigned long ce_count;
+	unsigned long de_count;
 	unsigned long err_addr_cnt;
 	struct eeprom_table_record *err_addr;
 	u32 err_list_count;
@@ -529,6 +564,8 @@ struct ras_manager {
 	struct ras_ih_data ih_data;
 
 	struct ras_err_data err_data;
+
+	struct aca_handle aca_handle;
 };
 
 struct ras_badpage {
@@ -548,6 +585,7 @@ struct ras_query_if {
 	struct ras_common_if head;
 	unsigned long ue_count;
 	unsigned long ce_count;
+	unsigned long de_count;
 };
 
 struct ras_inject_if {
@@ -781,7 +819,8 @@ struct amdgpu_ras* amdgpu_ras_get_context(struct amdgpu_device *adev);
 int amdgpu_ras_set_context(struct amdgpu_device *adev, struct amdgpu_ras *ras_con);
 
 int amdgpu_ras_set_mca_debug_mode(struct amdgpu_device *adev, bool enable);
-bool amdgpu_ras_get_mca_debug_mode(struct amdgpu_device *adev);
+int amdgpu_ras_set_aca_debug_mode(struct amdgpu_device *adev, bool enable);
+bool amdgpu_ras_get_aca_debug_mode(struct amdgpu_device *adev);
 bool amdgpu_ras_get_error_query_mode(struct amdgpu_device *adev,
 				     unsigned int *mode);
 
@@ -818,5 +857,20 @@ int amdgpu_ras_error_statistic_ce_count(struct ras_err_data *err_data,
 int amdgpu_ras_error_statistic_ue_count(struct ras_err_data *err_data,
 		struct amdgpu_smuio_mcm_config_info *mcm_info,
 		struct ras_err_addr *err_addr, u64 count);
+int amdgpu_ras_error_statistic_de_count(struct ras_err_data *err_data,
+		struct amdgpu_smuio_mcm_config_info *mcm_info,
+		struct ras_err_addr *err_addr, u64 count);
+void amdgpu_ras_query_boot_status(struct amdgpu_device *adev, u32 num_instances);
+int amdgpu_ras_bind_aca(struct amdgpu_device *adev, enum amdgpu_ras_block blk,
+			       const struct aca_info *aca_info, void *data);
+int amdgpu_ras_unbind_aca(struct amdgpu_device *adev, enum amdgpu_ras_block blk);
+
+ssize_t amdgpu_ras_aca_sysfs_read(struct device *dev, struct device_attribute *attr,
+				  struct aca_handle *handle, char *buf, void *data);
+
+void amdgpu_ras_add_mca_err_addr(struct ras_err_info *err_info,
+			struct ras_err_addr *err_addr);
 
+void amdgpu_ras_del_mca_err_addr(struct ras_err_info *err_info,
+		struct ras_err_addr *mca_err_addr);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c
index 2c3675d91614..db5791e1a7ce 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.c
@@ -241,7 +241,7 @@ void amdgpu_gfx_rlc_setup_cp_table(struct amdgpu_device *adev)
 			table_size = le32_to_cpu(hdr->jt_size);
 		}
 
-		for (i = 0; i < table_size; i ++) {
+		for (i = 0; i < table_size; i++) {
 			dst_ptr[bo_offset + i] =
 				cpu_to_le32(le32_to_cpu(fw_data[table_offset + i]));
 		}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h
index b591d33af264..5a17e0ff2ab8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h
@@ -169,7 +169,7 @@ struct amdgpu_rlc_funcs {
 	void (*stop)(struct amdgpu_device *adev);
 	void (*reset)(struct amdgpu_device *adev);
 	void (*start)(struct amdgpu_device *adev);
-	void (*update_spm_vmid)(struct amdgpu_device *adev, unsigned vmid);
+	void (*update_spm_vmid)(struct amdgpu_device *adev, struct amdgpu_ring *ring, unsigned vmid);
 	bool (*is_rlcg_access_range)(struct amdgpu_device *adev, uint32_t reg);
 };
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.c
index 7a6a67275404..e9081a98cf81 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.c
@@ -36,13 +36,28 @@
  */
 
 /**
+ * amdgpu_seq64_get_va_base - Get the seq64 va base address
+ *
+ * @adev: amdgpu_device pointer
+ *
+ * Returns:
+ * va base address on success
+ */
+static inline u64 amdgpu_seq64_get_va_base(struct amdgpu_device *adev)
+{
+	u64 addr = adev->vm_manager.max_pfn << AMDGPU_GPU_PAGE_SHIFT;
+
+	addr -= AMDGPU_VA_RESERVED_TOP;
+
+	return addr;
+}
+
+/**
  * amdgpu_seq64_map - Map the seq64 memory to VM
  *
  * @adev: amdgpu_device pointer
  * @vm: vm pointer
  * @bo_va: bo_va pointer
- * @seq64_addr: seq64 vaddr start address
- * @size: seq64 pool size
  *
  * Map the seq64 memory to the given VM.
  *
@@ -50,11 +65,11 @@
  * 0 on success or a negative error code on failure
  */
 int amdgpu_seq64_map(struct amdgpu_device *adev, struct amdgpu_vm *vm,
-		     struct amdgpu_bo_va **bo_va, u64 seq64_addr,
-		     uint32_t size)
+		     struct amdgpu_bo_va **bo_va)
 {
 	struct amdgpu_bo *bo;
 	struct drm_exec exec;
+	u64 seq64_addr;
 	int r;
 
 	bo = adev->seq64.sbo;
@@ -77,9 +92,9 @@ int amdgpu_seq64_map(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 		goto error;
 	}
 
-	r = amdgpu_vm_bo_map(adev, *bo_va, seq64_addr, 0, size,
-			     AMDGPU_PTE_READABLE | AMDGPU_PTE_WRITEABLE |
-			     AMDGPU_PTE_EXECUTABLE);
+	seq64_addr = amdgpu_seq64_get_va_base(adev);
+	r = amdgpu_vm_bo_map(adev, *bo_va, seq64_addr, 0, AMDGPU_VA_RESERVED_SEQ64_SIZE,
+			     AMDGPU_PTE_READABLE);
 	if (r) {
 		DRM_ERROR("failed to do bo_map on userq sem, err=%d\n", r);
 		amdgpu_vm_bo_del(adev, *bo_va);
@@ -144,31 +159,25 @@ error:
  * amdgpu_seq64_alloc - Allocate a 64 bit memory
  *
  * @adev: amdgpu_device pointer
- * @gpu_addr: allocated gpu VA start address
- * @cpu_addr: allocated cpu VA start address
+ * @va: VA to access the seq in process address space
+ * @cpu_addr: CPU address to access the seq
  *
  * Alloc a 64 bit memory from seq64 pool.
  *
  * Returns:
  * 0 on success or a negative error code on failure
  */
-int amdgpu_seq64_alloc(struct amdgpu_device *adev, u64 *gpu_addr,
-		       u64 **cpu_addr)
+int amdgpu_seq64_alloc(struct amdgpu_device *adev, u64 *va, u64 **cpu_addr)
 {
 	unsigned long bit_pos;
-	u32 offset;
 
 	bit_pos = find_first_zero_bit(adev->seq64.used, adev->seq64.num_sem);
+	if (bit_pos >= adev->seq64.num_sem)
+		return -ENOSPC;
 
-	if (bit_pos < adev->seq64.num_sem) {
-		__set_bit(bit_pos, adev->seq64.used);
-		offset = bit_pos << 6; /* convert to qw offset */
-	} else {
-		return -EINVAL;
-	}
-
-	*gpu_addr = offset + AMDGPU_SEQ64_VADDR_START;
-	*cpu_addr = offset + adev->seq64.cpu_base_addr;
+	__set_bit(bit_pos, adev->seq64.used);
+	*va = bit_pos * sizeof(u64) + amdgpu_seq64_get_va_base(adev);
+	*cpu_addr = bit_pos + adev->seq64.cpu_base_addr;
 
 	return 0;
 }
@@ -177,20 +186,17 @@ int amdgpu_seq64_alloc(struct amdgpu_device *adev, u64 *gpu_addr,
  * amdgpu_seq64_free - Free the given 64 bit memory
  *
  * @adev: amdgpu_device pointer
- * @gpu_addr: gpu start address to be freed
+ * @va: gpu start address to be freed
  *
  * Free the given 64 bit memory from seq64 pool.
- *
  */
-void amdgpu_seq64_free(struct amdgpu_device *adev, u64 gpu_addr)
+void amdgpu_seq64_free(struct amdgpu_device *adev, u64 va)
 {
-	u32 offset;
-
-	offset = gpu_addr - AMDGPU_SEQ64_VADDR_START;
+	unsigned long bit_pos;
 
-	offset >>= 6;
-	if (offset < adev->seq64.num_sem)
-		__clear_bit(offset, adev->seq64.used);
+	bit_pos = (va - amdgpu_seq64_get_va_base(adev)) / sizeof(u64);
+	if (bit_pos < adev->seq64.num_sem)
+		__clear_bit(bit_pos, adev->seq64.used);
 }
 
 /**
@@ -229,7 +235,7 @@ int amdgpu_seq64_init(struct amdgpu_device *adev)
 	 * AMDGPU_MAX_SEQ64_SLOTS * sizeof(u64) * 8 = AMDGPU_MAX_SEQ64_SLOTS
 	 * 64bit slots
 	 */
-	r = amdgpu_bo_create_kernel(adev, AMDGPU_SEQ64_SIZE,
+	r = amdgpu_bo_create_kernel(adev, AMDGPU_VA_RESERVED_SEQ64_SIZE,
 				    PAGE_SIZE, AMDGPU_GEM_DOMAIN_GTT,
 				    &adev->seq64.sbo, NULL,
 				    (void **)&adev->seq64.cpu_base_addr);
@@ -238,7 +244,7 @@ int amdgpu_seq64_init(struct amdgpu_device *adev)
 		return r;
 	}
 
-	memset(adev->seq64.cpu_base_addr, 0, AMDGPU_SEQ64_SIZE);
+	memset(adev->seq64.cpu_base_addr, 0, AMDGPU_VA_RESERVED_SEQ64_SIZE);
 
 	adev->seq64.num_sem = AMDGPU_MAX_SEQ64_SLOTS;
 	memset(&adev->seq64.used, 0, sizeof(adev->seq64.used));
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.h
index 2196e72be508..4203b2ab318d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_seq64.h
@@ -25,10 +25,9 @@
 #ifndef __AMDGPU_SEQ64_H__
 #define __AMDGPU_SEQ64_H__
 
-#define AMDGPU_SEQ64_SIZE		(2ULL << 20)
-#define AMDGPU_MAX_SEQ64_SLOTS		(AMDGPU_SEQ64_SIZE / (sizeof(u64) * 8))
-#define AMDGPU_SEQ64_VADDR_OFFSET	0x50000
-#define AMDGPU_SEQ64_VADDR_START	(AMDGPU_VA_RESERVED_SIZE + AMDGPU_SEQ64_VADDR_OFFSET)
+#include "amdgpu_vm.h"
+
+#define AMDGPU_MAX_SEQ64_SLOTS         (AMDGPU_VA_RESERVED_SEQ64_SIZE / sizeof(u64))
 
 struct amdgpu_seq64 {
 	struct amdgpu_bo *sbo;
@@ -42,7 +41,7 @@ int amdgpu_seq64_init(struct amdgpu_device *adev);
 int amdgpu_seq64_alloc(struct amdgpu_device *adev, u64 *gpu_addr, u64 **cpu_addr);
 void amdgpu_seq64_free(struct amdgpu_device *adev, u64 gpu_addr);
 int amdgpu_seq64_map(struct amdgpu_device *adev, struct amdgpu_vm *vm,
-		     struct amdgpu_bo_va **bo_va, u64 seq64_addr, uint32_t size);
+		     struct amdgpu_bo_va **bo_va);
 void amdgpu_seq64_unmap(struct amdgpu_device *adev, struct amdgpu_fpriv *fpriv);
 
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
index d65e21914d8c..20436f81856a 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
@@ -23,6 +23,7 @@
 
 #include "amdgpu.h"
 #include "umc_v6_7.h"
+#define MAX_UMC_POISON_POLLING_TIME_SYNC   20  //ms
 
 static int amdgpu_umc_convert_error_address(struct amdgpu_device *adev,
 				    struct ras_err_data *err_data, uint64_t err_addr,
@@ -85,18 +86,21 @@ out_fini_err_data:
 	return ret;
 }
 
-static int amdgpu_umc_do_page_retirement(struct amdgpu_device *adev,
-		void *ras_error_status,
-		struct amdgpu_iv_entry *entry,
-		bool reset)
+static void amdgpu_umc_handle_bad_pages(struct amdgpu_device *adev,
+			void *ras_error_status)
 {
 	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	unsigned int error_query_mode;
 	int ret = 0;
+	unsigned long err_count;
 
-	kgd2kfd_set_sram_ecc_flag(adev->kfd.dev);
+	amdgpu_ras_get_error_query_mode(adev, &error_query_mode);
+
+	mutex_lock(&con->page_retirement_lock);
 	ret = amdgpu_dpm_get_ecc_info(adev, (void *)&(con->umc_ecc));
-	if (ret == -EOPNOTSUPP) {
+	if (ret == -EOPNOTSUPP &&
+	    error_query_mode == AMDGPU_RAS_DIRECT_ERROR_QUERY) {
 		if (adev->umc.ras && adev->umc.ras->ras_block.hw_ops &&
 		    adev->umc.ras->ras_block.hw_ops->query_ras_error_count)
 		    adev->umc.ras->ras_block.hw_ops->query_ras_error_count(adev, ras_error_status);
@@ -120,7 +124,8 @@ static int amdgpu_umc_do_page_retirement(struct amdgpu_device *adev,
 			 */
 			adev->umc.ras->ras_block.hw_ops->query_ras_error_address(adev, ras_error_status);
 		}
-	} else if (!ret) {
+	} else if (error_query_mode == AMDGPU_RAS_FIRMWARE_ERROR_QUERY ||
+	    (!ret && error_query_mode == AMDGPU_RAS_DIRECT_ERROR_QUERY)) {
 		if (adev->umc.ras &&
 		    adev->umc.ras->ecc_info_query_ras_error_count)
 		    adev->umc.ras->ecc_info_query_ras_error_count(adev, ras_error_status);
@@ -147,16 +152,13 @@ static int amdgpu_umc_do_page_retirement(struct amdgpu_device *adev,
 	}
 
 	/* only uncorrectable error needs gpu reset */
-	if (err_data->ue_count) {
-		dev_info(adev->dev, "%ld uncorrectable hardware errors "
-				"detected in UMC block\n",
-				err_data->ue_count);
-
+	if (err_data->ue_count || err_data->de_count) {
+		err_count = err_data->ue_count + err_data->de_count;
 		if ((amdgpu_bad_page_threshold != 0) &&
 			err_data->err_addr_cnt) {
 			amdgpu_ras_add_bad_pages(adev, err_data->err_addr,
 						err_data->err_addr_cnt);
-			amdgpu_ras_save_bad_pages(adev, &(err_data->ue_count));
+			amdgpu_ras_save_bad_pages(adev, &err_count);
 
 			amdgpu_dpm_send_hbm_bad_pages_num(adev, con->eeprom_control.ras_num_recs);
 
@@ -165,20 +167,87 @@ static int amdgpu_umc_do_page_retirement(struct amdgpu_device *adev,
 				con->update_channel_flag = false;
 			}
 		}
-
-		if (reset) {
-			/* use mode-2 reset for poison consumption */
-			if (!entry)
-				con->gpu_reset_flags |= AMDGPU_RAS_GPU_RESET_MODE2_RESET;
-			amdgpu_ras_reset_gpu(adev);
-		}
 	}
 
 	kfree(err_data->err_addr);
+
+	mutex_unlock(&con->page_retirement_lock);
+}
+
+static int amdgpu_umc_do_page_retirement(struct amdgpu_device *adev,
+		void *ras_error_status,
+		struct amdgpu_iv_entry *entry,
+		bool reset)
+{
+	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+
+	kgd2kfd_set_sram_ecc_flag(adev->kfd.dev);
+	amdgpu_umc_handle_bad_pages(adev, ras_error_status);
+
+	if (err_data->ue_count && reset) {
+		/* use mode-2 reset for poison consumption */
+		if (!entry)
+			con->gpu_reset_flags |= AMDGPU_RAS_GPU_RESET_MODE2_RESET;
+		amdgpu_ras_reset_gpu(adev);
+	}
+
 	return AMDGPU_RAS_SUCCESS;
 }
 
-int amdgpu_umc_poison_handler(struct amdgpu_device *adev, bool reset)
+int amdgpu_umc_bad_page_polling_timeout(struct amdgpu_device *adev,
+			bool reset, uint32_t timeout_ms)
+{
+	struct ras_err_data err_data;
+	struct ras_common_if head = {
+		.block = AMDGPU_RAS_BLOCK__UMC,
+	};
+	struct ras_manager *obj = amdgpu_ras_find_obj(adev, &head);
+	uint32_t timeout = timeout_ms;
+
+	memset(&err_data, 0, sizeof(err_data));
+	amdgpu_ras_error_data_init(&err_data);
+
+	do {
+
+		amdgpu_umc_handle_bad_pages(adev, &err_data);
+
+		if (timeout && !err_data.de_count) {
+			msleep(1);
+			timeout--;
+		}
+
+	} while (timeout && !err_data.de_count);
+
+	if (!timeout)
+		dev_warn(adev->dev, "Can't find bad pages\n");
+
+	if (err_data.de_count)
+		dev_info(adev->dev, "%ld new deferred hardware errors detected\n", err_data.de_count);
+
+	if (obj) {
+		obj->err_data.ue_count += err_data.ue_count;
+		obj->err_data.ce_count += err_data.ce_count;
+		obj->err_data.de_count += err_data.de_count;
+	}
+
+	amdgpu_ras_error_data_fini(&err_data);
+
+	kgd2kfd_set_sram_ecc_flag(adev->kfd.dev);
+
+	if (reset) {
+		struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+
+		/* use mode-2 reset for poison consumption */
+		con->gpu_reset_flags |= AMDGPU_RAS_GPU_RESET_MODE2_RESET;
+		amdgpu_ras_reset_gpu(adev);
+	}
+
+	return 0;
+}
+
+int amdgpu_umc_poison_handler(struct amdgpu_device *adev,
+			enum amdgpu_ras_block block, bool reset)
 {
 	int ret = AMDGPU_RAS_SUCCESS;
 
@@ -195,27 +264,41 @@ int amdgpu_umc_poison_handler(struct amdgpu_device *adev, bool reset)
 	}
 
 	if (!amdgpu_sriov_vf(adev)) {
-		struct ras_err_data err_data;
-		struct ras_common_if head = {
-			.block = AMDGPU_RAS_BLOCK__UMC,
-		};
-		struct ras_manager *obj = amdgpu_ras_find_obj(adev, &head);
+		if (amdgpu_ip_version(adev, UMC_HWIP, 0) < IP_VERSION(12, 0, 0)) {
+			struct ras_err_data err_data;
+			struct ras_common_if head = {
+				.block = AMDGPU_RAS_BLOCK__UMC,
+			};
+			struct ras_manager *obj = amdgpu_ras_find_obj(adev, &head);
+
+			ret = amdgpu_ras_error_data_init(&err_data);
+			if (ret)
+				return ret;
 
-		ret = amdgpu_ras_error_data_init(&err_data);
-		if (ret)
-			return ret;
+			ret = amdgpu_umc_do_page_retirement(adev, &err_data, NULL, reset);
 
-		ret = amdgpu_umc_do_page_retirement(adev, &err_data, NULL, reset);
+			if (ret == AMDGPU_RAS_SUCCESS && obj) {
+				obj->err_data.ue_count += err_data.ue_count;
+				obj->err_data.ce_count += err_data.ce_count;
+				obj->err_data.de_count += err_data.de_count;
+			}
 
-		if (ret == AMDGPU_RAS_SUCCESS && obj) {
-			obj->err_data.ue_count += err_data.ue_count;
-			obj->err_data.ce_count += err_data.ce_count;
-		}
+			amdgpu_ras_error_data_fini(&err_data);
+		} else {
+			if (reset) {
+				amdgpu_umc_bad_page_polling_timeout(adev,
+							reset, MAX_UMC_POISON_POLLING_TIME_SYNC);
+			} else {
+				struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 
-		amdgpu_ras_error_data_fini(&err_data);
+				atomic_inc(&con->page_retirement_req_cnt);
+
+				wake_up(&con->page_retirement_wq);
+			}
+		}
 	} else {
 		if (adev->virt.ops && adev->virt.ops->ras_poison_handler)
-			adev->virt.ops->ras_poison_handler(adev);
+			adev->virt.ops->ras_poison_handler(adev, block);
 		else
 			dev_warn(adev->dev,
 				"No ras_poison_handler interface in SRIOV!\n");
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h
index 417a6726c71b..26d2ae498daf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.h
@@ -21,7 +21,7 @@
 #ifndef __AMDGPU_UMC_H__
 #define __AMDGPU_UMC_H__
 #include "amdgpu_ras.h"
-
+#include "amdgpu_mca.h"
 /*
  * (addr / 256) * 4096, the higher 26 bits in ErrorAddr
  * is the index of 4KB block
@@ -64,6 +64,8 @@ struct amdgpu_umc_ras {
 				      void *ras_error_status);
 	void (*ecc_info_query_ras_error_address)(struct amdgpu_device *adev,
 					void *ras_error_status);
+	bool (*check_ecc_err_status)(struct amdgpu_device *adev,
+			enum amdgpu_mca_error_type type, void *ras_error_status);
 	/* support different eeprom table version for different asic */
 	void (*set_eeprom_table_version)(struct amdgpu_ras_eeprom_table_header *hdr);
 };
@@ -100,7 +102,8 @@ struct amdgpu_umc {
 
 int amdgpu_umc_ras_sw_init(struct amdgpu_device *adev);
 int amdgpu_umc_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block);
-int amdgpu_umc_poison_handler(struct amdgpu_device *adev, bool reset);
+int amdgpu_umc_poison_handler(struct amdgpu_device *adev,
+			enum amdgpu_ras_block block, bool reset);
 int amdgpu_umc_process_ecc_irq(struct amdgpu_device *adev,
 		struct amdgpu_irq_src *source,
 		struct amdgpu_iv_entry *entry);
@@ -118,4 +121,7 @@ int amdgpu_umc_page_retirement_mca(struct amdgpu_device *adev,
 
 int amdgpu_umc_loop_channels(struct amdgpu_device *adev,
 			umc_func func, void *data);
+
+int amdgpu_umc_bad_page_polling_timeout(struct amdgpu_device *adev,
+			bool reset, uint32_t timeout_ms);
 #endif
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umr.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_umr.h
index 107f9bb0e24f..5b27fc41ffbf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umr.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umr.h
@@ -69,12 +69,12 @@ struct amdgpu_debugfs_gprwave_data {
 };
 
 enum AMDGPU_DEBUGFS_REGS2_CMDS {
-	AMDGPU_DEBUGFS_REGS2_CMD_SET_STATE=0,
+	AMDGPU_DEBUGFS_REGS2_CMD_SET_STATE = 0,
 	AMDGPU_DEBUGFS_REGS2_CMD_SET_STATE_V2,
 };
 
 enum AMDGPU_DEBUGFS_GPRWAVE_CMDS {
-	AMDGPU_DEBUGFS_GPRWAVE_CMD_SET_STATE=0,
+	AMDGPU_DEBUGFS_GPRWAVE_CMD_SET_STATE = 0,
 };
 
 //reg2 interface
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c
index bfbf59326ee1..ab820cf52668 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umsch_mm.c
@@ -358,7 +358,7 @@ static int setup_umsch_mm_test(struct amdgpu_device *adev,
 
 	memset(test->ring_data_cpu_addr, 0, sizeof(struct umsch_mm_test_ring_data));
 
-	test->ring_data_gpu_addr = AMDGPU_VA_RESERVED_SIZE;
+	test->ring_data_gpu_addr = AMDGPU_VA_RESERVED_BOTTOM;
 	r = map_ring_data(adev, test->vm, test->ring_data_obj, &test->bo_va,
 			  test->ring_data_gpu_addr, sizeof(struct umsch_mm_test_ring_data));
 	if (r)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
index f4963330c772..f300d4a4457d 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c
@@ -1189,7 +1189,7 @@ int amdgpu_vcn_process_poison_irq(struct amdgpu_device *adev,
 		amdgpu_ras_interrupt_dispatch(adev, &ih_data);
 	} else {
 		if (adev->virt.ops && adev->virt.ops->ras_poison_handler)
-			adev->virt.ops->ras_poison_handler(adev);
+			adev->virt.ops->ras_poison_handler(adev, ras_if->block);
 		else
 			dev_warn(adev->dev,
 				"No ras_poison_handler interface in SRIOV for VCN!\n");
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
index 0dcff2889e25..6ff7d3fb2008 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c
@@ -71,59 +71,6 @@ void amdgpu_virt_init_setting(struct amdgpu_device *adev)
 		amdgpu_num_kcq = 2;
 }
 
-void amdgpu_virt_kiq_reg_write_reg_wait(struct amdgpu_device *adev,
-					uint32_t reg0, uint32_t reg1,
-					uint32_t ref, uint32_t mask,
-					uint32_t xcc_inst)
-{
-	struct amdgpu_kiq *kiq = &adev->gfx.kiq[xcc_inst];
-	struct amdgpu_ring *ring = &kiq->ring;
-	signed long r, cnt = 0;
-	unsigned long flags;
-	uint32_t seq;
-
-	if (adev->mes.ring.sched.ready) {
-		amdgpu_mes_reg_write_reg_wait(adev, reg0, reg1,
-					      ref, mask);
-		return;
-	}
-
-	spin_lock_irqsave(&kiq->ring_lock, flags);
-	amdgpu_ring_alloc(ring, 32);
-	amdgpu_ring_emit_reg_write_reg_wait(ring, reg0, reg1,
-					    ref, mask);
-	r = amdgpu_fence_emit_polling(ring, &seq, MAX_KIQ_REG_WAIT);
-	if (r)
-		goto failed_undo;
-
-	amdgpu_ring_commit(ring);
-	spin_unlock_irqrestore(&kiq->ring_lock, flags);
-
-	r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
-
-	/* don't wait anymore for IRQ context */
-	if (r < 1 && in_interrupt())
-		goto failed_kiq;
-
-	might_sleep();
-	while (r < 1 && cnt++ < MAX_KIQ_REG_TRY) {
-
-		msleep(MAX_KIQ_REG_BAILOUT_INTERVAL);
-		r = amdgpu_fence_wait_polling(ring, seq, MAX_KIQ_REG_WAIT);
-	}
-
-	if (cnt > MAX_KIQ_REG_TRY)
-		goto failed_kiq;
-
-	return;
-
-failed_undo:
-	amdgpu_ring_undo(ring);
-	spin_unlock_irqrestore(&kiq->ring_lock, flags);
-failed_kiq:
-	dev_err(adev->dev, "failed to write reg %x wait reg %x\n", reg0, reg1);
-}
-
 /**
  * amdgpu_virt_request_full_gpu() - request full gpu access
  * @adev:	amdgpu device.
@@ -303,11 +250,11 @@ static int amdgpu_virt_init_ras_err_handler_data(struct amdgpu_device *adev)
 	if (!*data)
 		goto data_failure;
 
-	bps = kmalloc_array(align_space, sizeof((*data)->bps), GFP_KERNEL);
+	bps = kmalloc_array(align_space, sizeof(*(*data)->bps), GFP_KERNEL);
 	if (!bps)
 		goto bps_failure;
 
-	bps_bo = kmalloc_array(align_space, sizeof((*data)->bps_bo), GFP_KERNEL);
+	bps_bo = kmalloc_array(align_space, sizeof(*(*data)->bps_bo), GFP_KERNEL);
 	if (!bps_bo)
 		goto bps_bo_failure;
 
@@ -340,8 +287,10 @@ static void amdgpu_virt_ras_release_bp(struct amdgpu_device *adev)
 
 	for (i = data->last_reserved - 1; i >= 0; i--) {
 		bo = data->bps_bo[i];
-		amdgpu_bo_free_kernel(&bo, NULL, NULL);
-		data->bps_bo[i] = bo;
+		if (bo) {
+			amdgpu_bo_free_kernel(&bo, NULL, NULL);
+			data->bps_bo[i] = bo;
+		}
 		data->last_reserved = i;
 	}
 }
@@ -381,6 +330,8 @@ static void amdgpu_virt_ras_reserve_bps(struct amdgpu_device *adev)
 {
 	struct amdgpu_virt *virt = &adev->virt;
 	struct amdgpu_virt_ras_err_handler_data *data = virt->virt_eh_data;
+	struct amdgpu_vram_mgr *mgr = &adev->mman.vram_mgr;
+	struct ttm_resource_manager *man = &mgr->manager;
 	struct amdgpu_bo *bo = NULL;
 	uint64_t bp;
 	int i;
@@ -396,12 +347,18 @@ static void amdgpu_virt_ras_reserve_bps(struct amdgpu_device *adev)
 		 * 2) a ras bad page has been reserved (duplicate error injection
 		 *    for one page);
 		 */
-		if (amdgpu_bo_create_kernel_at(adev, bp << AMDGPU_GPU_PAGE_SHIFT,
-					       AMDGPU_GPU_PAGE_SIZE,
-					       &bo, NULL))
-			DRM_DEBUG("RAS WARN: reserve vram for retired page %llx fail\n", bp);
-
-		data->bps_bo[i] = bo;
+		if  (ttm_resource_manager_used(man)) {
+			amdgpu_vram_mgr_reserve_range(&adev->mman.vram_mgr,
+				bp << AMDGPU_GPU_PAGE_SHIFT,
+				AMDGPU_GPU_PAGE_SIZE);
+			data->bps_bo[i] = NULL;
+		} else {
+			if (amdgpu_bo_create_kernel_at(adev, bp << AMDGPU_GPU_PAGE_SHIFT,
+							AMDGPU_GPU_PAGE_SIZE,
+							&bo, NULL))
+				DRM_DEBUG("RAS WARN: reserve vram for retired page %llx fail\n", bp);
+			data->bps_bo[i] = bo;
+		}
 		data->last_reserved = i + 1;
 		bo = NULL;
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
index d4207e44141f..fa7be5f277b9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_virt.h
@@ -88,7 +88,8 @@ struct amdgpu_virt_ops {
 	int (*wait_reset)(struct amdgpu_device *adev);
 	void (*trans_msg)(struct amdgpu_device *adev, enum idh_request req,
 			  u32 data1, u32 data2, u32 data3);
-	void (*ras_poison_handler)(struct amdgpu_device *adev);
+	void (*ras_poison_handler)(struct amdgpu_device *adev,
+					enum amdgpu_ras_block block);
 };
 
 /*
@@ -332,10 +333,6 @@ static inline bool is_virtual_machine(void)
 	((adev)->virt.gim_feature & AMDGIM_FEATURE_VCN_RB_DECOUPLE)
 bool amdgpu_virt_mmio_blocked(struct amdgpu_device *adev);
 void amdgpu_virt_init_setting(struct amdgpu_device *adev);
-void amdgpu_virt_kiq_reg_write_reg_wait(struct amdgpu_device *adev,
-					uint32_t reg0, uint32_t rreg1,
-					uint32_t ref, uint32_t mask,
-					uint32_t xcc_inst);
 int amdgpu_virt_request_full_gpu(struct amdgpu_device *adev, bool init);
 int amdgpu_virt_release_full_gpu(struct amdgpu_device *adev, bool init);
 int amdgpu_virt_reset_gpu(struct amdgpu_device *adev);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
index 453a4b786cfc..8baa2e0935cc 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vkms.c
@@ -660,8 +660,7 @@ static const struct amd_ip_funcs amdgpu_vkms_ip_funcs = {
 	.set_powergating_state = amdgpu_vkms_set_powergating_state,
 };
 
-const struct amdgpu_ip_block_version amdgpu_vkms_ip_block =
-{
+const struct amdgpu_ip_block_version amdgpu_vkms_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_DCE,
 	.major = 1,
 	.minor = 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index b8fcb6c55698..ed4a8c5d26d7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
@@ -234,6 +234,22 @@ static void amdgpu_vm_bo_invalidated(struct amdgpu_vm_bo_base *vm_bo)
 }
 
 /**
+ * amdgpu_vm_bo_evicted_user - vm_bo is evicted
+ *
+ * @vm_bo: vm_bo which is evicted
+ *
+ * State for BOs used by user mode queues which are not at the location they
+ * should be.
+ */
+static void amdgpu_vm_bo_evicted_user(struct amdgpu_vm_bo_base *vm_bo)
+{
+	vm_bo->moved = true;
+	spin_lock(&vm_bo->vm->status_lock);
+	list_move(&vm_bo->vm_status, &vm_bo->vm->evicted_user);
+	spin_unlock(&vm_bo->vm->status_lock);
+}
+
+/**
  * amdgpu_vm_bo_relocated - vm_bo is reloacted
  *
  * @vm_bo: vm_bo which is relocated
@@ -427,21 +443,25 @@ uint64_t amdgpu_vm_generation(struct amdgpu_device *adev, struct amdgpu_vm *vm)
 }
 
 /**
- * amdgpu_vm_validate_pt_bos - validate the page table BOs
+ * amdgpu_vm_validate - validate evicted BOs tracked in the VM
  *
  * @adev: amdgpu device pointer
  * @vm: vm providing the BOs
+ * @ticket: optional reservation ticket used to reserve the VM
  * @validate: callback to do the validation
  * @param: parameter for the validation callback
  *
- * Validate the page table BOs on command submission if neccessary.
+ * Validate the page table BOs and per-VM BOs on command submission if
+ * necessary. If a ticket is given, also try to validate evicted user queue
+ * BOs. They must already be reserved with the given ticket.
  *
  * Returns:
  * Validation result.
  */
-int amdgpu_vm_validate_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm,
-			      int (*validate)(void *p, struct amdgpu_bo *bo),
-			      void *param)
+int amdgpu_vm_validate(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+		       struct ww_acquire_ctx *ticket,
+		       int (*validate)(void *p, struct amdgpu_bo *bo),
+		       void *param)
 {
 	struct amdgpu_vm_bo_base *bo_base;
 	struct amdgpu_bo *shadow;
@@ -484,6 +504,28 @@ int amdgpu_vm_validate_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 		}
 		spin_lock(&vm->status_lock);
 	}
+	while (ticket && !list_empty(&vm->evicted_user)) {
+		bo_base = list_first_entry(&vm->evicted_user,
+					   struct amdgpu_vm_bo_base,
+					   vm_status);
+		spin_unlock(&vm->status_lock);
+
+		bo = bo_base->bo;
+
+		if (dma_resv_locking_ctx(bo->tbo.base.resv) != ticket) {
+			pr_warn_ratelimited("Evicted user BO is not reserved in pid %d\n",
+					    vm->task_info.pid);
+			return -EINVAL;
+		}
+
+		r = validate(param, bo);
+		if (r)
+			return r;
+
+		amdgpu_vm_bo_invalidated(bo_base);
+
+		spin_lock(&vm->status_lock);
+	}
 	spin_unlock(&vm->status_lock);
 
 	amdgpu_vm_eviction_lock(vm);
@@ -651,7 +693,7 @@ int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job,
 		amdgpu_gmc_emit_pasid_mapping(ring, job->vmid, job->pasid);
 
 	if (spm_update_needed && adev->gfx.rlc.funcs->update_spm_vmid)
-		adev->gfx.rlc.funcs->update_spm_vmid(adev, job->vmid);
+		adev->gfx.rlc.funcs->update_spm_vmid(adev, ring, job->vmid);
 
 	if (!ring->is_mes_queue && ring->funcs->emit_gds_switch &&
 	    gds_switch_needed) {
@@ -1426,11 +1468,21 @@ int amdgpu_vm_handle_moved(struct amdgpu_device *adev,
 		}
 
 		r = amdgpu_vm_bo_update(adev, bo_va, clear);
-		if (r)
-			return r;
 
 		if (unlock)
 			dma_resv_unlock(resv);
+		if (r)
+			return r;
+
+		/* Remember evicted DMABuf imports in compute VMs for later
+		 * validation
+		 */
+		if (vm->is_compute_context &&
+		    bo_va->base.bo->tbo.base.import_attach &&
+		    (!bo_va->base.bo->tbo.resource ||
+		     bo_va->base.bo->tbo.resource->mem_type == TTM_PL_SYSTEM))
+			amdgpu_vm_bo_evicted_user(&bo_va->base);
+
 		spin_lock(&vm->status_lock);
 	}
 	spin_unlock(&vm->status_lock);
@@ -2196,6 +2248,7 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm,
 	for (i = 0; i < AMDGPU_MAX_VMHUBS; i++)
 		vm->reserved_vmid[i] = NULL;
 	INIT_LIST_HEAD(&vm->evicted);
+	INIT_LIST_HEAD(&vm->evicted_user);
 	INIT_LIST_HEAD(&vm->relocated);
 	INIT_LIST_HEAD(&vm->moved);
 	INIT_LIST_HEAD(&vm->idle);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
index 4740dd65b99d..666698a57192 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h
@@ -136,7 +136,11 @@ struct amdgpu_mem_stats;
 #define AMDGPU_IS_MMHUB1(x) ((x) >= AMDGPU_MMHUB1_START && (x) < AMDGPU_MAX_VMHUBS)
 
 /* Reserve 2MB at top/bottom of address space for kernel use */
-#define AMDGPU_VA_RESERVED_SIZE			(2ULL << 20)
+#define AMDGPU_VA_RESERVED_CSA_SIZE		(2ULL << 20)
+#define AMDGPU_VA_RESERVED_SEQ64_SIZE		(2ULL << 20)
+#define AMDGPU_VA_RESERVED_BOTTOM		(2ULL << 20)
+#define AMDGPU_VA_RESERVED_TOP			(AMDGPU_VA_RESERVED_SEQ64_SIZE + \
+						 AMDGPU_VA_RESERVED_CSA_SIZE)
 
 /* See vm_update_mode */
 #define AMDGPU_VM_USE_CPU_FOR_GFX (1 << 0)
@@ -288,9 +292,12 @@ struct amdgpu_vm {
 	/* Lock to protect vm_bo add/del/move on all lists of vm */
 	spinlock_t		status_lock;
 
-	/* BOs who needs a validation */
+	/* Per-VM and PT BOs who needs a validation */
 	struct list_head	evicted;
 
+	/* BOs for user mode queues that need a validation */
+	struct list_head	evicted_user;
+
 	/* PT BOs which relocated and their parent need an update */
 	struct list_head	relocated;
 
@@ -434,9 +441,10 @@ int amdgpu_vm_lock_pd(struct amdgpu_vm *vm, struct drm_exec *exec,
 		      unsigned int num_fences);
 bool amdgpu_vm_ready(struct amdgpu_vm *vm);
 uint64_t amdgpu_vm_generation(struct amdgpu_device *adev, struct amdgpu_vm *vm);
-int amdgpu_vm_validate_pt_bos(struct amdgpu_device *adev, struct amdgpu_vm *vm,
-			      int (*callback)(void *p, struct amdgpu_bo *bo),
-			      void *param);
+int amdgpu_vm_validate(struct amdgpu_device *adev, struct amdgpu_vm *vm,
+		       struct ww_acquire_ctx *ticket,
+		       int (*callback)(void *p, struct amdgpu_bo *bo),
+		       void *param);
 int amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_job *job, bool need_pipe_sync);
 int amdgpu_vm_update_pdes(struct amdgpu_device *adev,
 			  struct amdgpu_vm *vm, bool immediate);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
index a6c88f2fe6e5..20d51f6c9bb8 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_xgmi.c
@@ -1035,15 +1035,74 @@ int amdgpu_xgmi_remove_device(struct amdgpu_device *adev)
 	return 0;
 }
 
+static int xgmi_v6_4_0_aca_bank_generate_report(struct aca_handle *handle, struct aca_bank *bank, enum aca_error_type type,
+						struct aca_bank_report *report, void *data)
+{
+	struct amdgpu_device *adev = handle->adev;
+	const char *error_str;
+	u64 status;
+	int ret, ext_error_code;
+
+	ret = aca_bank_info_decode(bank, &report->info);
+	if (ret)
+		return ret;
+
+	status = bank->regs[ACA_REG_IDX_STATUS];
+	ext_error_code = ACA_REG__STATUS__ERRORCODEEXT(status);
+
+	error_str = ext_error_code < ARRAY_SIZE(xgmi_v6_4_0_ras_error_code_ext) ?
+		xgmi_v6_4_0_ras_error_code_ext[ext_error_code] : NULL;
+	if (error_str)
+		dev_info(adev->dev, "%s detected\n", error_str);
+
+	if ((type == ACA_ERROR_TYPE_UE && ext_error_code == 0) ||
+	    (type == ACA_ERROR_TYPE_CE && ext_error_code == 6))
+		report->count[type] = ACA_REG__MISC0__ERRCNT(bank->regs[ACA_REG_IDX_MISC0]);
+
+	return 0;
+}
+
+static const struct aca_bank_ops xgmi_v6_4_0_aca_bank_ops = {
+	.aca_bank_generate_report = xgmi_v6_4_0_aca_bank_generate_report,
+};
+
+static const struct aca_info xgmi_v6_4_0_aca_info = {
+	.hwip = ACA_HWIP_TYPE_PCS_XGMI,
+	.mask = ACA_ERROR_UE_MASK | ACA_ERROR_CE_MASK,
+	.bank_ops = &xgmi_v6_4_0_aca_bank_ops,
+};
+
 static int amdgpu_xgmi_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
 {
+	int r;
+
 	if (!adev->gmc.xgmi.supported ||
 	    adev->gmc.xgmi.num_physical_nodes == 0)
 		return 0;
 
 	amdgpu_ras_reset_error_count(adev, AMDGPU_RAS_BLOCK__XGMI_WAFL);
 
-	return amdgpu_ras_block_late_init(adev, ras_block);
+	r = amdgpu_ras_block_late_init(adev, ras_block);
+	if (r)
+		return r;
+
+	switch (amdgpu_ip_version(adev, XGMI_HWIP, 0)) {
+	case IP_VERSION(6, 4, 0):
+		r = amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__XGMI_WAFL,
+					&xgmi_v6_4_0_aca_info, NULL);
+		if (r)
+			goto late_fini;
+		break;
+	default:
+		break;
+	}
+
+	return 0;
+
+late_fini:
+	amdgpu_ras_block_late_fini(adev, ras_block);
+
+	return r;
 }
 
 uint64_t amdgpu_xgmi_get_relative_phy_addr(struct amdgpu_device *adev,
@@ -1099,7 +1158,7 @@ static void amdgpu_xgmi_legacy_reset_ras_error_count(struct amdgpu_device *adev)
 
 static void __xgmi_v6_4_0_reset_error_count(struct amdgpu_device *adev, int xgmi_inst, u64 mca_base)
 {
-	WREG64_MCA(xgmi_inst, mca_base, MCA_REG_IDX_STATUS, 0ULL);
+	WREG64_MCA(xgmi_inst, mca_base, ACA_REG_IDX_STATUS, 0ULL);
 }
 
 static void xgmi_v6_4_0_reset_error_count(struct amdgpu_device *adev, int xgmi_inst)
@@ -1277,12 +1336,12 @@ static void amdgpu_xgmi_legacy_query_ras_error_count(struct amdgpu_device *adev,
 	err_data->ce_count += ce_cnt;
 }
 
-static enum amdgpu_mca_error_type xgmi_v6_4_0_pcs_mca_get_error_type(struct amdgpu_device *adev, u64 status)
+static enum aca_error_type xgmi_v6_4_0_pcs_mca_get_error_type(struct amdgpu_device *adev, u64 status)
 {
 	const char *error_str;
 	int ext_error_code;
 
-	ext_error_code = MCA_REG__STATUS__ERRORCODEEXT(status);
+	ext_error_code = ACA_REG__STATUS__ERRORCODEEXT(status);
 
 	error_str = ext_error_code < ARRAY_SIZE(xgmi_v6_4_0_ras_error_code_ext) ?
 		xgmi_v6_4_0_ras_error_code_ext[ext_error_code] : NULL;
@@ -1291,9 +1350,9 @@ static enum amdgpu_mca_error_type xgmi_v6_4_0_pcs_mca_get_error_type(struct amdg
 
 	switch (ext_error_code) {
 	case 0:
-		return AMDGPU_MCA_ERROR_TYPE_UE;
+		return ACA_ERROR_TYPE_UE;
 	case 6:
-		return AMDGPU_MCA_ERROR_TYPE_CE;
+		return ACA_ERROR_TYPE_CE;
 	default:
 		return -EINVAL;
 	}
@@ -1307,22 +1366,22 @@ static void __xgmi_v6_4_0_query_error_count(struct amdgpu_device *adev, struct a
 	int xgmi_inst = mcm_info->die_id;
 	u64 status = 0;
 
-	status = RREG64_MCA(xgmi_inst, mca_base, MCA_REG_IDX_STATUS);
-	if (!MCA_REG__STATUS__VAL(status))
+	status = RREG64_MCA(xgmi_inst, mca_base, ACA_REG_IDX_STATUS);
+	if (!ACA_REG__STATUS__VAL(status))
 		return;
 
 	switch (xgmi_v6_4_0_pcs_mca_get_error_type(adev, status)) {
-	case AMDGPU_MCA_ERROR_TYPE_UE:
+	case ACA_ERROR_TYPE_UE:
 		amdgpu_ras_error_statistic_ue_count(err_data, mcm_info, NULL, 1ULL);
 		break;
-	case AMDGPU_MCA_ERROR_TYPE_CE:
+	case ACA_ERROR_TYPE_CE:
 		amdgpu_ras_error_statistic_ce_count(err_data, mcm_info, NULL, 1ULL);
 		break;
 	default:
 		break;
 	}
 
-	WREG64_MCA(xgmi_inst, mca_base, MCA_REG_IDX_STATUS, 0ULL);
+	WREG64_MCA(xgmi_inst, mca_base, ACA_REG_IDX_STATUS, 0ULL);
 }
 
 static void xgmi_v6_4_0_query_error_count(struct amdgpu_device *adev, int xgmi_inst, struct ras_err_data *err_data)
diff --git a/drivers/gpu/drm/amd/amdgpu/atom.c b/drivers/gpu/drm/amd/amdgpu/atom.c
index a33e890c70d9..b888613f653f 100644
--- a/drivers/gpu/drm/amd/amdgpu/atom.c
+++ b/drivers/gpu/drm/amd/amdgpu/atom.c
@@ -62,6 +62,7 @@
 typedef struct {
 	struct atom_context *ctx;
 	uint32_t *ps, *ws;
+	int ps_size, ws_size;
 	int ps_shift;
 	uint16_t start;
 	unsigned last_jump;
@@ -70,8 +71,8 @@ typedef struct {
 } atom_exec_context;
 
 int amdgpu_atom_debug;
-static int amdgpu_atom_execute_table_locked(struct atom_context *ctx, int index, uint32_t *params);
-int amdgpu_atom_execute_table(struct atom_context *ctx, int index, uint32_t *params);
+static int amdgpu_atom_execute_table_locked(struct atom_context *ctx, int index, uint32_t *params, int params_size);
+int amdgpu_atom_execute_table(struct atom_context *ctx, int index, uint32_t *params, int params_size);
 
 static uint32_t atom_arg_mask[8] =
 	{ 0xFFFFFFFF, 0xFFFF, 0xFFFF00, 0xFFFF0000, 0xFF, 0xFF00, 0xFF0000,
@@ -223,7 +224,10 @@ static uint32_t atom_get_src_int(atom_exec_context *ctx, uint8_t attr,
 		(*ptr)++;
 		/* get_unaligned_le32 avoids unaligned accesses from atombios
 		 * tables, noticed on a DEC Alpha. */
-		val = get_unaligned_le32((u32 *)&ctx->ps[idx]);
+		if (idx < ctx->ps_size)
+			val = get_unaligned_le32((u32 *)&ctx->ps[idx]);
+		else
+			pr_info("PS index out of range: %i > %i\n", idx, ctx->ps_size);
 		if (print)
 			DEBUG("PS[0x%02X,0x%04X]", idx, val);
 		break;
@@ -261,7 +265,10 @@ static uint32_t atom_get_src_int(atom_exec_context *ctx, uint8_t attr,
 			val = gctx->reg_block;
 			break;
 		default:
-			val = ctx->ws[idx];
+			if (idx < ctx->ws_size)
+				val = ctx->ws[idx];
+			else
+				pr_info("WS index out of range: %i > %i\n", idx, ctx->ws_size);
 		}
 		break;
 	case ATOM_ARG_ID:
@@ -495,6 +502,10 @@ static void atom_put_dst(atom_exec_context *ctx, int arg, uint8_t attr,
 		idx = U8(*ptr);
 		(*ptr)++;
 		DEBUG("PS[0x%02X]", idx);
+		if (idx >= ctx->ps_size) {
+			pr_info("PS index out of range: %i > %i\n", idx, ctx->ps_size);
+			return;
+		}
 		ctx->ps[idx] = cpu_to_le32(val);
 		break;
 	case ATOM_ARG_WS:
@@ -527,6 +538,10 @@ static void atom_put_dst(atom_exec_context *ctx, int arg, uint8_t attr,
 			gctx->reg_block = val;
 			break;
 		default:
+			if (idx >= ctx->ws_size) {
+				pr_info("WS index out of range: %i > %i\n", idx, ctx->ws_size);
+				return;
+			}
 			ctx->ws[idx] = val;
 		}
 		break;
@@ -624,7 +639,7 @@ static void atom_op_calltable(atom_exec_context *ctx, int *ptr, int arg)
 	else
 		SDEBUG("   table: %d\n", idx);
 	if (U16(ctx->ctx->cmd_table + 4 + 2 * idx))
-		r = amdgpu_atom_execute_table_locked(ctx->ctx, idx, ctx->ps + ctx->ps_shift);
+		r = amdgpu_atom_execute_table_locked(ctx->ctx, idx, ctx->ps + ctx->ps_shift, ctx->ps_size - ctx->ps_shift);
 	if (r) {
 		ctx->abort = true;
 	}
@@ -1203,7 +1218,7 @@ static struct {
 	atom_op_div32, ATOM_ARG_WS},
 };
 
-static int amdgpu_atom_execute_table_locked(struct atom_context *ctx, int index, uint32_t *params)
+static int amdgpu_atom_execute_table_locked(struct atom_context *ctx, int index, uint32_t *params, int params_size)
 {
 	int base = CU16(ctx->cmd_table + 4 + 2 * index);
 	int len, ws, ps, ptr;
@@ -1225,12 +1240,16 @@ static int amdgpu_atom_execute_table_locked(struct atom_context *ctx, int index,
 	ectx.ps_shift = ps / 4;
 	ectx.start = base;
 	ectx.ps = params;
+	ectx.ps_size = params_size;
 	ectx.abort = false;
 	ectx.last_jump = 0;
-	if (ws)
+	if (ws) {
 		ectx.ws = kcalloc(4, ws, GFP_KERNEL);
-	else
+		ectx.ws_size = ws;
+	} else {
 		ectx.ws = NULL;
+		ectx.ws_size = 0;
+	}
 
 	debug_depth++;
 	while (1) {
@@ -1264,7 +1283,7 @@ free:
 	return ret;
 }
 
-int amdgpu_atom_execute_table(struct atom_context *ctx, int index, uint32_t *params)
+int amdgpu_atom_execute_table(struct atom_context *ctx, int index, uint32_t *params, int params_size)
 {
 	int r;
 
@@ -1280,7 +1299,7 @@ int amdgpu_atom_execute_table(struct atom_context *ctx, int index, uint32_t *par
 	/* reset divmul */
 	ctx->divmul[0] = 0;
 	ctx->divmul[1] = 0;
-	r = amdgpu_atom_execute_table_locked(ctx, index, params);
+	r = amdgpu_atom_execute_table_locked(ctx, index, params, params_size);
 	mutex_unlock(&ctx->mutex);
 	return r;
 }
@@ -1552,7 +1571,7 @@ int amdgpu_atom_asic_init(struct atom_context *ctx)
 
 	if (!CU16(ctx->cmd_table + 4 + 2 * ATOM_CMD_INIT))
 		return 1;
-	ret = amdgpu_atom_execute_table(ctx, ATOM_CMD_INIT, ps);
+	ret = amdgpu_atom_execute_table(ctx, ATOM_CMD_INIT, ps, 16);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/atom.h b/drivers/gpu/drm/amd/amdgpu/atom.h
index c11cf18a0f18..b807f6639a4c 100644
--- a/drivers/gpu/drm/amd/amdgpu/atom.h
+++ b/drivers/gpu/drm/amd/amdgpu/atom.h
@@ -156,7 +156,7 @@ struct atom_context {
 extern int amdgpu_atom_debug;
 
 struct atom_context *amdgpu_atom_parse(struct card_info *card, void *bios);
-int amdgpu_atom_execute_table(struct atom_context *ctx, int index, uint32_t *params);
+int amdgpu_atom_execute_table(struct atom_context *ctx, int index, uint32_t *params, int params_size);
 int amdgpu_atom_asic_init(struct atom_context *ctx);
 void amdgpu_atom_destroy(struct atom_context *ctx);
 bool amdgpu_atom_parse_data_header(struct atom_context *ctx, int index, uint16_t *size,
diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_crtc.c b/drivers/gpu/drm/amd/amdgpu/atombios_crtc.c
index 10098fdd33fc..3dfc28840a7d 100644
--- a/drivers/gpu/drm/amd/amdgpu/atombios_crtc.c
+++ b/drivers/gpu/drm/amd/amdgpu/atombios_crtc.c
@@ -77,7 +77,7 @@ void amdgpu_atombios_crtc_overscan_setup(struct drm_crtc *crtc,
 		args.usOverscanTop = cpu_to_le16(amdgpu_crtc->v_border);
 		break;
 	}
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 void amdgpu_atombios_crtc_scaler_setup(struct drm_crtc *crtc)
@@ -106,7 +106,7 @@ void amdgpu_atombios_crtc_scaler_setup(struct drm_crtc *crtc)
 		args.ucEnable = ATOM_SCALER_DISABLE;
 		break;
 	}
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 void amdgpu_atombios_crtc_lock(struct drm_crtc *crtc, int lock)
@@ -123,7 +123,7 @@ void amdgpu_atombios_crtc_lock(struct drm_crtc *crtc, int lock)
 	args.ucCRTC = amdgpu_crtc->crtc_id;
 	args.ucEnable = lock;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 void amdgpu_atombios_crtc_enable(struct drm_crtc *crtc, int state)
@@ -139,7 +139,7 @@ void amdgpu_atombios_crtc_enable(struct drm_crtc *crtc, int state)
 	args.ucCRTC = amdgpu_crtc->crtc_id;
 	args.ucEnable = state;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 void amdgpu_atombios_crtc_blank(struct drm_crtc *crtc, int state)
@@ -155,7 +155,7 @@ void amdgpu_atombios_crtc_blank(struct drm_crtc *crtc, int state)
 	args.ucCRTC = amdgpu_crtc->crtc_id;
 	args.ucBlanking = state;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 void amdgpu_atombios_crtc_powergate(struct drm_crtc *crtc, int state)
@@ -171,7 +171,7 @@ void amdgpu_atombios_crtc_powergate(struct drm_crtc *crtc, int state)
 	args.ucDispPipeId = amdgpu_crtc->crtc_id;
 	args.ucEnable = state;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 void amdgpu_atombios_crtc_powergate_init(struct amdgpu_device *adev)
@@ -183,7 +183,7 @@ void amdgpu_atombios_crtc_powergate_init(struct amdgpu_device *adev)
 
 	args.ucEnable = ATOM_INIT;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 void amdgpu_atombios_crtc_set_dtd_timing(struct drm_crtc *crtc,
@@ -228,7 +228,7 @@ void amdgpu_atombios_crtc_set_dtd_timing(struct drm_crtc *crtc,
 	args.susModeMiscInfo.usAccess = cpu_to_le16(misc);
 	args.ucCRTC = amdgpu_crtc->crtc_id;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 union atom_enable_ss {
@@ -293,7 +293,7 @@ static void amdgpu_atombios_crtc_program_ss(struct amdgpu_device *adev,
 	args.v3.usSpreadSpectrumStep = cpu_to_le16(ss->step);
 	args.v3.ucEnable = enable;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 union adjust_pixel_clock {
@@ -395,7 +395,7 @@ static u32 amdgpu_atombios_crtc_adjust_pll(struct drm_crtc *crtc,
 					ADJUST_DISPLAY_CONFIG_SS_ENABLE;
 
 			amdgpu_atom_execute_table(adev->mode_info.atom_context,
-					   index, (uint32_t *)&args);
+					   index, (uint32_t *)&args, sizeof(args));
 			adjusted_clock = le16_to_cpu(args.v1.usPixelClock) * 10;
 			break;
 		case 3:
@@ -428,7 +428,7 @@ static u32 amdgpu_atombios_crtc_adjust_pll(struct drm_crtc *crtc,
 				args.v3.sInput.ucExtTransmitterID = 0;
 
 			amdgpu_atom_execute_table(adev->mode_info.atom_context,
-					   index, (uint32_t *)&args);
+					   index, (uint32_t *)&args, sizeof(args));
 			adjusted_clock = le32_to_cpu(args.v3.sOutput.ulDispPllFreq) * 10;
 			if (args.v3.sOutput.ucRefDiv) {
 				amdgpu_crtc->pll_flags |= AMDGPU_PLL_USE_FRAC_FB_DIV;
@@ -514,7 +514,7 @@ void amdgpu_atombios_crtc_set_disp_eng_pll(struct amdgpu_device *adev,
 		DRM_ERROR("Unknown table version %d %d\n", frev, crev);
 		return;
 	}
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 union set_dce_clock {
@@ -544,7 +544,7 @@ u32 amdgpu_atombios_crtc_set_dce_clock(struct amdgpu_device *adev,
 			args.v2_1.asParam.ulDCEClkFreq = cpu_to_le32(freq); /* 10kHz units */
 			args.v2_1.asParam.ucDCEClkType = clk_type;
 			args.v2_1.asParam.ucDCEClkSrc = clk_src;
-			amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+			amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 			ret_freq = le32_to_cpu(args.v2_1.asParam.ulDCEClkFreq) * 10;
 			break;
 		default:
@@ -740,7 +740,7 @@ void amdgpu_atombios_crtc_program_pll(struct drm_crtc *crtc,
 		return;
 	}
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 int amdgpu_atombios_crtc_prepare_pll(struct drm_crtc *crtc,
diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_dp.c b/drivers/gpu/drm/amd/amdgpu/atombios_dp.c
index 87c41e0e9b7c..622634c08c7b 100644
--- a/drivers/gpu/drm/amd/amdgpu/atombios_dp.c
+++ b/drivers/gpu/drm/amd/amdgpu/atombios_dp.c
@@ -83,7 +83,7 @@ static int amdgpu_atombios_dp_process_aux_ch(struct amdgpu_i2c_chan *chan,
 	args.v2.ucDelay = delay / 10;
 	args.v2.ucHPD_ID = chan->rec.hpd;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 
 	*ack = args.v2.ucReplyStatus;
 
@@ -301,7 +301,7 @@ static u8 amdgpu_atombios_dp_encoder_service(struct amdgpu_device *adev,
 	args.ucLaneNum = lane_num;
 	args.ucStatus = 0;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 	return args.ucStatus;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c b/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c
index 7672abe6c140..25feab188dfe 100644
--- a/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c
+++ b/drivers/gpu/drm/amd/amdgpu/atombios_encoders.c
@@ -335,7 +335,7 @@ amdgpu_atombios_encoder_setup_dac(struct drm_encoder *encoder, int action)
 	args.ucDacStandard = ATOM_DAC1_PS2;
 	args.usPixelClock = cpu_to_le16(amdgpu_encoder->pixel_clock / 10);
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 
 }
 
@@ -432,7 +432,7 @@ amdgpu_atombios_encoder_setup_dvo(struct drm_encoder *encoder, int action)
 		break;
 	}
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 int amdgpu_atombios_encoder_get_encoder_mode(struct drm_encoder *encoder)
@@ -732,7 +732,7 @@ amdgpu_atombios_encoder_setup_dig_encoder(struct drm_encoder *encoder,
 		break;
 	}
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 
 }
 
@@ -1136,7 +1136,7 @@ amdgpu_atombios_encoder_setup_dig_transmitter(struct drm_encoder *encoder, int a
 		break;
 	}
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 bool
@@ -1164,7 +1164,7 @@ amdgpu_atombios_encoder_set_edp_panel_power(struct drm_connector *connector,
 
 	args.v1.ucAction = action;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 
 	/* wait for the panel to power up */
 	if (action == ATOM_TRANSMITTER_ACTION_POWER_ON) {
@@ -1288,7 +1288,7 @@ amdgpu_atombios_encoder_setup_external_encoder(struct drm_encoder *encoder,
 		DRM_ERROR("Unknown table version: %d, %d\n", frev, crev);
 		return;
 	}
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 static void
@@ -1633,7 +1633,7 @@ amdgpu_atombios_encoder_set_crtc_source(struct drm_encoder *encoder)
 		return;
 	}
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
 
 /* This only needs to be called once at startup */
@@ -1706,7 +1706,7 @@ amdgpu_atombios_encoder_dac_load_detect(struct drm_encoder *encoder,
 				args.sDacload.ucMisc = DAC_LOAD_MISC_YPrPb;
 		}
 
-		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+		amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 
 		return true;
 	} else
diff --git a/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c b/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c
index af0335535f82..a6501114322f 100644
--- a/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c
+++ b/drivers/gpu/drm/amd/amdgpu/atombios_i2c.c
@@ -86,7 +86,7 @@ static int amdgpu_atombios_i2c_process_i2c_ch(struct amdgpu_i2c_chan *chan,
 	args.ucSlaveAddr = slave_addr << 1;
 	args.ucLineNumber = chan->rec.i2c_id;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 
 	/* error */
 	if (args.ucStatus != HW_ASSISTED_I2C_STATUS_SUCCESS) {
@@ -172,5 +172,5 @@ void amdgpu_atombios_i2c_channel_trans(struct amdgpu_device *adev, u8 slave_addr
 	args.ucSlaveAddr = slave_addr;
 	args.ucLineNumber = line_number;
 
-	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args);
+	amdgpu_atom_execute_table(adev->mode_info.atom_context, index, (uint32_t *)&args, sizeof(args));
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/clearstate_gfx9.h b/drivers/gpu/drm/amd/amdgpu/clearstate_gfx9.h
index 567a904804bc..9c85ca6358c1 100644
--- a/drivers/gpu/drm/amd/amdgpu/clearstate_gfx9.h
+++ b/drivers/gpu/drm/amd/amdgpu/clearstate_gfx9.h
@@ -21,8 +21,7 @@
  *
  */
 
-static const unsigned int gfx9_SECT_CONTEXT_def_1[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_1[] = {
     0x00000000, // DB_RENDER_CONTROL
     0x00000000, // DB_COUNT_CONTROL
     0x00000000, // DB_DEPTH_VIEW
@@ -236,8 +235,7 @@ static const unsigned int gfx9_SECT_CONTEXT_def_1[] =
     0x00000000, // PA_SC_VPORT_ZMIN_15
     0x3f800000, // PA_SC_VPORT_ZMAX_15
 };
-static const unsigned int gfx9_SECT_CONTEXT_def_2[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_2[] = {
     0x00000000, // PA_SC_SCREEN_EXTENT_CONTROL
     0x00000000, // PA_SC_TILE_STEERING_OVERRIDE
     0x00000000, // CP_PERFMON_CNTX_CNTL
@@ -521,15 +519,13 @@ static const unsigned int gfx9_SECT_CONTEXT_def_2[] =
     0x00000000, // CB_MRT6_EPITCH
     0x00000000, // CB_MRT7_EPITCH
 };
-static const unsigned int gfx9_SECT_CONTEXT_def_3[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_3[] = {
     0x00000000, // PA_CL_POINT_X_RAD
     0x00000000, // PA_CL_POINT_Y_RAD
     0x00000000, // PA_CL_POINT_SIZE
     0x00000000, // PA_CL_POINT_CULL_RAD
 };
-static const unsigned int gfx9_SECT_CONTEXT_def_4[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_4[] = {
     0x00000000, // DB_DEPTH_CONTROL
     0x00000000, // DB_EQAA
     0x00000000, // CB_COLOR_CONTROL
@@ -688,17 +684,14 @@ static const unsigned int gfx9_SECT_CONTEXT_def_4[] =
     0x00000000, // VGT_GS_OUT_PRIM_TYPE
     0x00000000, // IA_ENHANCE
 };
-static const unsigned int gfx9_SECT_CONTEXT_def_5[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_5[] = {
     0x00000000, // WD_ENHANCE
     0x00000000, // VGT_PRIMITIVEID_EN
 };
-static const unsigned int gfx9_SECT_CONTEXT_def_6[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_6[] = {
     0x00000000, // VGT_PRIMITIVEID_RESET
 };
-static const unsigned int gfx9_SECT_CONTEXT_def_7[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_7[] = {
     0x00000000, // VGT_GS_MAX_PRIMS_PER_SUBGROUP
     0x00000000, // VGT_DRAW_PAYLOAD_CNTL
     0, // HOLE
@@ -766,8 +759,7 @@ static const unsigned int gfx9_SECT_CONTEXT_def_7[] =
     0x00000000, // VGT_STRMOUT_CONFIG
     0x00000000, // VGT_STRMOUT_BUFFER_CONFIG
 };
-static const unsigned int gfx9_SECT_CONTEXT_def_8[] =
-{
+static const unsigned int gfx9_SECT_CONTEXT_def_8[] = {
     0x00000000, // PA_SC_CENTROID_PRIORITY_0
     0x00000000, // PA_SC_CENTROID_PRIORITY_1
     0x00001000, // PA_SC_LINE_CNTL
@@ -924,8 +916,7 @@ static const unsigned int gfx9_SECT_CONTEXT_def_8[] =
     0x00000000, // CB_COLOR7_DCC_BASE
     0x00000000, // CB_COLOR7_DCC_BASE_EXT
 };
-static const struct cs_extent_def gfx9_SECT_CONTEXT_defs[] =
-{
+static const struct cs_extent_def gfx9_SECT_CONTEXT_defs[] = {
     {gfx9_SECT_CONTEXT_def_1, 0x0000a000, 212 },
     {gfx9_SECT_CONTEXT_def_2, 0x0000a0d6, 282 },
     {gfx9_SECT_CONTEXT_def_3, 0x0000a1f5, 4 },
diff --git a/drivers/gpu/drm/amd/amdgpu/clearstate_si.h b/drivers/gpu/drm/amd/amdgpu/clearstate_si.h
index 66e39cdb5cb0..5fd96ddd7f0f 100644
--- a/drivers/gpu/drm/amd/amdgpu/clearstate_si.h
+++ b/drivers/gpu/drm/amd/amdgpu/clearstate_si.h
@@ -21,8 +21,7 @@
  *
  */
 
-static const u32 si_SECT_CONTEXT_def_1[] =
-{
+static const u32 si_SECT_CONTEXT_def_1[] = {
     0x00000000, // DB_RENDER_CONTROL
     0x00000000, // DB_COUNT_CONTROL
     0x00000000, // DB_DEPTH_VIEW
@@ -236,8 +235,7 @@ static const u32 si_SECT_CONTEXT_def_1[] =
     0x00000000, // PA_SC_VPORT_ZMIN_15
     0x3f800000, // PA_SC_VPORT_ZMAX_15
 };
-static const u32 si_SECT_CONTEXT_def_2[] =
-{
+static const u32 si_SECT_CONTEXT_def_2[] = {
     0x00000000, // CP_PERFMON_CNTX_CNTL
     0x00000000, // CP_RINGID
     0x00000000, // CP_VMID
@@ -511,8 +509,7 @@ static const u32 si_SECT_CONTEXT_def_2[] =
     0x00000000, // CB_BLEND6_CONTROL
     0x00000000, // CB_BLEND7_CONTROL
 };
-static const u32 si_SECT_CONTEXT_def_3[] =
-{
+static const u32 si_SECT_CONTEXT_def_3[] = {
     0x00000000, // PA_CL_POINT_X_RAD
     0x00000000, // PA_CL_POINT_Y_RAD
     0x00000000, // PA_CL_POINT_SIZE
@@ -520,8 +517,7 @@ static const u32 si_SECT_CONTEXT_def_3[] =
     0x00000000, // VGT_DMA_BASE_HI
     0x00000000, // VGT_DMA_BASE
 };
-static const u32 si_SECT_CONTEXT_def_4[] =
-{
+static const u32 si_SECT_CONTEXT_def_4[] = {
     0x00000000, // DB_DEPTH_CONTROL
     0x00000000, // DB_EQAA
     0x00000000, // CB_COLOR_CONTROL
@@ -680,16 +676,13 @@ static const u32 si_SECT_CONTEXT_def_4[] =
     0x00000000, // VGT_GS_OUT_PRIM_TYPE
     0x00000000, // IA_ENHANCE
 };
-static const u32 si_SECT_CONTEXT_def_5[] =
-{
+static const u32 si_SECT_CONTEXT_def_5[] = {
     0x00000000, // VGT_PRIMITIVEID_EN
 };
-static const u32 si_SECT_CONTEXT_def_6[] =
-{
+static const u32 si_SECT_CONTEXT_def_6[] = {
     0x00000000, // VGT_PRIMITIVEID_RESET
 };
-static const u32 si_SECT_CONTEXT_def_7[] =
-{
+static const u32 si_SECT_CONTEXT_def_7[] = {
     0x00000000, // VGT_MULTI_PRIM_IB_RESET_EN
     0, // HOLE
     0, // HOLE
@@ -924,8 +917,7 @@ static const u32 si_SECT_CONTEXT_def_7[] =
     0x00000000, // CB_COLOR7_CLEAR_WORD0
     0x00000000, // CB_COLOR7_CLEAR_WORD1
 };
-static const struct cs_extent_def si_SECT_CONTEXT_defs[] =
-{
+static const struct cs_extent_def si_SECT_CONTEXT_defs[] = {
     {si_SECT_CONTEXT_def_1, 0x0000a000, 212 },
     {si_SECT_CONTEXT_def_2, 0x0000a0d8, 272 },
     {si_SECT_CONTEXT_def_3, 0x0000a1f5, 6 },
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c
index 587ee632a3b8..221af054d874 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c
@@ -52,6 +52,7 @@
 
 static void dce_v10_0_set_display_funcs(struct amdgpu_device *adev);
 static void dce_v10_0_set_irq_funcs(struct amdgpu_device *adev);
+static void dce_v10_0_hpd_int_ack(struct amdgpu_device *adev, int hpd);
 
 static const u32 crtc_offsets[] = {
 	CRTC0_REGISTER_OFFSET,
@@ -364,6 +365,7 @@ static void dce_v10_0_hpd_init(struct amdgpu_device *adev)
 				    AMDGPU_HPD_DISCONNECT_INT_DELAY_IN_MS);
 		WREG32(mmDC_HPD_TOGGLE_FILT_CNTL + hpd_offsets[amdgpu_connector->hpd.hpd], tmp);
 
+		dce_v10_0_hpd_int_ack(adev, amdgpu_connector->hpd.hpd);
 		dce_v10_0_hpd_set_polarity(adev, amdgpu_connector->hpd.hpd);
 		amdgpu_irq_get(adev, &adev->hpd_irq,
 			       amdgpu_connector->hpd.hpd);
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c
index f22ec27365bd..69e8b0db6cf7 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c
@@ -52,6 +52,7 @@
 
 static void dce_v11_0_set_display_funcs(struct amdgpu_device *adev);
 static void dce_v11_0_set_irq_funcs(struct amdgpu_device *adev);
+static void dce_v11_0_hpd_int_ack(struct amdgpu_device *adev, int hpd);
 
 static const u32 crtc_offsets[] =
 {
@@ -388,6 +389,7 @@ static void dce_v11_0_hpd_init(struct amdgpu_device *adev)
 				    AMDGPU_HPD_DISCONNECT_INT_DELAY_IN_MS);
 		WREG32(mmDC_HPD_TOGGLE_FILT_CNTL + hpd_offsets[amdgpu_connector->hpd.hpd], tmp);
 
+		dce_v11_0_hpd_int_ack(adev, amdgpu_connector->hpd.hpd);
 		dce_v11_0_hpd_set_polarity(adev, amdgpu_connector->hpd.hpd);
 		amdgpu_irq_get(adev, &adev->hpd_irq, amdgpu_connector->hpd.hpd);
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
index 4dbe9b3259b5..60d40201fdd1 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v6_0.c
@@ -273,6 +273,21 @@ static void dce_v6_0_hpd_set_polarity(struct amdgpu_device *adev,
 	WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], tmp);
 }
 
+static void dce_v6_0_hpd_int_ack(struct amdgpu_device *adev,
+				 int hpd)
+{
+	u32 tmp;
+
+	if (hpd >= adev->mode_info.num_hpd) {
+		DRM_DEBUG("invalid hdp %d\n", hpd);
+		return;
+	}
+
+	tmp = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd]);
+	tmp |= DC_HPD1_INT_CONTROL__DC_HPD1_INT_ACK_MASK;
+	WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], tmp);
+}
+
 /**
  * dce_v6_0_hpd_init - hpd setup callback.
  *
@@ -312,6 +327,7 @@ static void dce_v6_0_hpd_init(struct amdgpu_device *adev)
 			continue;
 		}
 
+		dce_v6_0_hpd_int_ack(adev, amdgpu_connector->hpd.hpd);
 		dce_v6_0_hpd_set_polarity(adev, amdgpu_connector->hpd.hpd);
 		amdgpu_irq_get(adev, &adev->hpd_irq, amdgpu_connector->hpd.hpd);
 	}
@@ -3089,7 +3105,7 @@ static int dce_v6_0_hpd_irq(struct amdgpu_device *adev,
 			    struct amdgpu_irq_src *source,
 			    struct amdgpu_iv_entry *entry)
 {
-	uint32_t disp_int, mask, tmp;
+	uint32_t disp_int, mask;
 	unsigned hpd;
 
 	if (entry->src_data[0] >= adev->mode_info.num_hpd) {
@@ -3102,9 +3118,7 @@ static int dce_v6_0_hpd_irq(struct amdgpu_device *adev,
 	mask = interrupt_status_offsets[hpd].hpd;
 
 	if (disp_int & mask) {
-		tmp = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd]);
-		tmp |= DC_HPD1_INT_CONTROL__DC_HPD1_INT_ACK_MASK;
-		WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], tmp);
+		dce_v6_0_hpd_int_ack(adev, hpd);
 		schedule_delayed_work(&adev->hotplug_work, 0);
 		DRM_DEBUG("IH: HPD%d\n", hpd + 1);
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
index 05bcce23385e..5a5fcc45e452 100644
--- a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c
@@ -265,6 +265,21 @@ static void dce_v8_0_hpd_set_polarity(struct amdgpu_device *adev,
 	WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], tmp);
 }
 
+static void dce_v8_0_hpd_int_ack(struct amdgpu_device *adev,
+				 int hpd)
+{
+	u32 tmp;
+
+	if (hpd >= adev->mode_info.num_hpd) {
+		DRM_DEBUG("invalid hdp %d\n", hpd);
+		return;
+	}
+
+	tmp = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd]);
+	tmp |= DC_HPD1_INT_CONTROL__DC_HPD1_INT_ACK_MASK;
+	WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], tmp);
+}
+
 /**
  * dce_v8_0_hpd_init - hpd setup callback.
  *
@@ -304,6 +319,7 @@ static void dce_v8_0_hpd_init(struct amdgpu_device *adev)
 			continue;
 		}
 
+		dce_v8_0_hpd_int_ack(adev, amdgpu_connector->hpd.hpd);
 		dce_v8_0_hpd_set_polarity(adev, amdgpu_connector->hpd.hpd);
 		amdgpu_irq_get(adev, &adev->hpd_irq, amdgpu_connector->hpd.hpd);
 	}
@@ -3177,7 +3193,7 @@ static int dce_v8_0_hpd_irq(struct amdgpu_device *adev,
 			    struct amdgpu_irq_src *source,
 			    struct amdgpu_iv_entry *entry)
 {
-	uint32_t disp_int, mask, tmp;
+	uint32_t disp_int, mask;
 	unsigned hpd;
 
 	if (entry->src_data[0] >= adev->mode_info.num_hpd) {
@@ -3190,9 +3206,7 @@ static int dce_v8_0_hpd_irq(struct amdgpu_device *adev,
 	mask = interrupt_status_offsets[hpd].hpd;
 
 	if (disp_int & mask) {
-		tmp = RREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd]);
-		tmp |= DC_HPD1_INT_CONTROL__DC_HPD1_INT_ACK_MASK;
-		WREG32(mmDC_HPD1_INT_CONTROL + hpd_offsets[hpd], tmp);
+		dce_v8_0_hpd_int_ack(adev, hpd);
 		schedule_delayed_work(&adev->hotplug_work, 0);
 		DRM_DEBUG("IH: HPD%d\n", hpd + 1);
 	}
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index dcdecb18b230..b02d63328f1c 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
@@ -7947,7 +7947,7 @@ static void gfx_v10_0_update_spm_vmid_internal(struct amdgpu_device *adev,
 	WREG32_SOC15_NO_KIQ(GC, 0, mmRLC_SPM_MC_CNTL, data);
 }
 
-static void gfx_v10_0_update_spm_vmid(struct amdgpu_device *adev, unsigned int vmid)
+static void gfx_v10_0_update_spm_vmid(struct amdgpu_device *adev, struct amdgpu_ring *ring, unsigned int vmid)
 {
 	amdgpu_gfx_off_ctrl(adev, false);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
index 4f3bfdc75b37..2fb1342d5bd9 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c
@@ -727,7 +727,7 @@ static int gfx_v11_0_rlc_init(struct amdgpu_device *adev)
 
 	/* init spm vmid with 0xf */
 	if (adev->gfx.rlc.funcs->update_spm_vmid)
-		adev->gfx.rlc.funcs->update_spm_vmid(adev, 0xf);
+		adev->gfx.rlc.funcs->update_spm_vmid(adev, NULL, 0xf);
 
 	return 0;
 }
@@ -5027,7 +5027,7 @@ static int gfx_v11_0_update_gfx_clock_gating(struct amdgpu_device *adev,
 	return 0;
 }
 
-static void gfx_v11_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
+static void gfx_v11_0_update_spm_vmid(struct amdgpu_device *adev, struct amdgpu_ring *ring, unsigned vmid)
 {
 	u32 data;
 
@@ -5041,6 +5041,14 @@ static void gfx_v11_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
 	WREG32_SOC15_NO_KIQ(GC, 0, regRLC_SPM_MC_CNTL, data);
 
 	amdgpu_gfx_off_ctrl(adev, true);
+
+	if (ring
+		&& amdgpu_sriov_is_pp_one_vf(adev)
+		&& ((ring->funcs->type == AMDGPU_RING_TYPE_GFX)
+			|| (ring->funcs->type == AMDGPU_RING_TYPE_COMPUTE))) {
+		uint32_t reg = SOC15_REG_OFFSET(GC, 0, regRLC_SPM_MC_CNTL);
+		amdgpu_ring_emit_wreg(ring, reg, data);
+	}
 }
 
 static const struct amdgpu_rlc_funcs gfx_v11_0_rlc_funcs = {
@@ -6104,7 +6112,8 @@ static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_gfx = {
 	.get_rptr = gfx_v11_0_ring_get_rptr_gfx,
 	.get_wptr = gfx_v11_0_ring_get_wptr_gfx,
 	.set_wptr = gfx_v11_0_ring_set_wptr_gfx,
-	.emit_frame_size = /* totally 242 maximum if 16 IBs */
+	.emit_frame_size = /* totally 247 maximum if 16 IBs */
+		5 + /* update_spm_vmid */
 		5 + /* COND_EXEC */
 		9 + /* SET_Q_PREEMPTION_MODE */
 		7 + /* PIPELINE_SYNC */
@@ -6154,6 +6163,7 @@ static const struct amdgpu_ring_funcs gfx_v11_0_ring_funcs_compute = {
 	.get_wptr = gfx_v11_0_ring_get_wptr_compute,
 	.set_wptr = gfx_v11_0_ring_set_wptr_compute,
 	.emit_frame_size =
+		5 + /* update_spm_vmid */
 		20 + /* gfx_v11_0_ring_emit_gds_switch */
 		7 + /* gfx_v11_0_ring_emit_hdp_flush */
 		5 + /* hdp invalidate */
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3.c
index 26d6286d86c9..9e7ce1e6bc06 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0_3.c
@@ -69,7 +69,7 @@ static int gfx_v11_0_3_rlc_gc_fed_irq(struct amdgpu_device *adev,
 		amdgpu_ras_interrupt_dispatch(adev, &ih_data);
 	} else {
 		if (adev->virt.ops && adev->virt.ops->ras_poison_handler)
-			adev->virt.ops->ras_poison_handler(adev);
+			adev->virt.ops->ras_poison_handler(adev, ras_if->block);
 		else
 			dev_warn(adev->dev,
 				"No ras_poison_handler interface in SRIOV for %s!\n", ras_if->name);
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
index c2faf6b4c2fc..86a4865b1ae5 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c
@@ -3274,7 +3274,7 @@ static int gfx_v7_0_rlc_init(struct amdgpu_device *adev)
 
 	/* init spm vmid with 0xf */
 	if (adev->gfx.rlc.funcs->update_spm_vmid)
-		adev->gfx.rlc.funcs->update_spm_vmid(adev, 0xf);
+		adev->gfx.rlc.funcs->update_spm_vmid(adev, NULL, 0xf);
 
 	return 0;
 }
@@ -3500,7 +3500,7 @@ static int gfx_v7_0_rlc_resume(struct amdgpu_device *adev)
 	return 0;
 }
 
-static void gfx_v7_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
+static void gfx_v7_0_update_spm_vmid(struct amdgpu_device *adev, struct amdgpu_ring *ring, unsigned vmid)
 {
 	u32 data;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
index 1943beb135c4..ea174b76ee70 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c
@@ -1288,7 +1288,7 @@ static int gfx_v8_0_rlc_init(struct amdgpu_device *adev)
 
 	/* init spm vmid with 0xf */
 	if (adev->gfx.rlc.funcs->update_spm_vmid)
-		adev->gfx.rlc.funcs->update_spm_vmid(adev, 0xf);
+		adev->gfx.rlc.funcs->update_spm_vmid(adev, NULL, 0xf);
 
 	return 0;
 }
@@ -5579,7 +5579,7 @@ static void gfx_v8_0_unset_safe_mode(struct amdgpu_device *adev, int xcc_id)
 	}
 }
 
-static void gfx_v8_0_update_spm_vmid(struct amdgpu_device *adev, unsigned vmid)
+static void gfx_v8_0_update_spm_vmid(struct amdgpu_device *adev, struct amdgpu_ring *ring, unsigned vmid)
 {
 	u32 data;
 
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
index 69c500910746..169d45268ef6 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c
@@ -3034,6 +3034,14 @@ static int gfx_v9_0_cp_gfx_start(struct amdgpu_device *adev)
 
 	gfx_v9_0_cp_gfx_enable(adev, true);
 
+	/* Now only limit the quirk on the APU gfx9 series and already
+	 * confirmed that the APU gfx10/gfx11 needn't such update.
+	 */
+	if (adev->flags & AMD_IS_APU &&
+			adev->in_s3 && !adev->suspend_complete) {
+		DRM_INFO(" Will skip the CSB packet resubmit\n");
+		return 0;
+	}
 	r = amdgpu_ring_alloc(ring, gfx_v9_0_get_csb_size(adev) + 4 + 3);
 	if (r) {
 		DRM_ERROR("amdgpu: cp failed to lock ring (%d).\n", r);
@@ -4894,7 +4902,7 @@ static void gfx_v9_0_update_spm_vmid_internal(struct amdgpu_device *adev,
 		WREG32_SOC15(GC, 0, mmRLC_SPM_MC_CNTL, data);
 }
 
-static void gfx_v9_0_update_spm_vmid(struct amdgpu_device *adev, unsigned int vmid)
+static void gfx_v9_0_update_spm_vmid(struct amdgpu_device *adev, struct amdgpu_ring *ring, unsigned int vmid)
 {
 	amdgpu_gfx_off_ctrl(adev, false);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c
index bc8416afb62c..f53b379d8971 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4.c
@@ -970,8 +970,9 @@ static void gfx_v9_4_reset_ras_error_count(struct amdgpu_device *adev)
 	WREG32_SOC15(GC, 0, mmATC_L2_CACHE_4K_DSM_INDEX, 255);
 }
 
-static const struct soc15_reg_entry gfx_v9_4_ea_err_status_regs =
-	{ SOC15_REG_ENTRY(GC, 0, mmGCEA_ERR_STATUS), 0, 1, 32 };
+static const struct soc15_reg_entry gfx_v9_4_ea_err_status_regs = {
+	SOC15_REG_ENTRY(GC, 0, mmGCEA_ERR_STATUS), 0, 1, 32
+};
 
 static void gfx_v9_4_query_ras_error_status(struct amdgpu_device *adev)
 {
diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
index 131cddbdda0d..aace4594a603 100644
--- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
+++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c
@@ -38,6 +38,7 @@
 
 #include "gfx_v9_4_3.h"
 #include "amdgpu_xcp.h"
+#include "amdgpu_aca.h"
 
 MODULE_FIRMWARE("amdgpu/gc_9_4_3_mec.bin");
 MODULE_FIRMWARE("amdgpu/gc_9_4_3_rlc.bin");
@@ -48,6 +49,10 @@ MODULE_FIRMWARE("amdgpu/gc_9_4_3_rlc.bin");
 #define GOLDEN_GB_ADDR_CONFIG 0x2a114042
 #define CP_HQD_PERSISTENT_STATE_DEFAULT 0xbe05301
 
+#define mmSMNAID_XCD0_MCA_SMU 0x36430400	/* SMN AID XCD0 */
+#define mmSMNAID_XCD1_MCA_SMU 0x38430400	/* SMN AID XCD1 */
+#define mmSMNXCD_XCD0_MCA_SMU 0x40430400	/* SMN XCD XCD0 */
+
 struct amdgpu_gfx_ras gfx_v9_4_3_ras;
 
 static void gfx_v9_4_3_set_ring_funcs(struct amdgpu_device *adev);
@@ -675,6 +680,66 @@ static const struct amdgpu_gfx_funcs gfx_v9_4_3_gfx_funcs = {
 	.ih_node_to_logical_xcc = &gfx_v9_4_3_ih_to_xcc_inst,
 };
 
+static int gfx_v9_4_3_aca_bank_generate_report(struct aca_handle *handle,
+					       struct aca_bank *bank, enum aca_error_type type,
+					       struct aca_bank_report *report, void *data)
+{
+	u64 status, misc0;
+	u32 instlo;
+	int ret;
+
+	status = bank->regs[ACA_REG_IDX_STATUS];
+	if ((type == ACA_ERROR_TYPE_UE &&
+	     ACA_REG__STATUS__ERRORCODEEXT(status) == ACA_EXTERROR_CODE_FAULT) ||
+	    (type == ACA_ERROR_TYPE_CE &&
+	     ACA_REG__STATUS__ERRORCODEEXT(status) == ACA_EXTERROR_CODE_CE)) {
+
+		ret = aca_bank_info_decode(bank, &report->info);
+		if (ret)
+			return ret;
+
+		/* NOTE: overwrite info.die_id with xcd id for gfx */
+		instlo = ACA_REG__IPID__INSTANCEIDLO(bank->regs[ACA_REG_IDX_IPID]);
+		instlo &= GENMASK(31, 1);
+		report->info.die_id = instlo == mmSMNAID_XCD0_MCA_SMU ? 0 : 1;
+
+		misc0 = bank->regs[ACA_REG_IDX_MISC0];
+		report->count[type] = ACA_REG__MISC0__ERRCNT(misc0);
+	}
+
+	return 0;
+}
+
+static bool gfx_v9_4_3_aca_bank_is_valid(struct aca_handle *handle, struct aca_bank *bank,
+					 enum aca_error_type type, void *data)
+{
+	u32 instlo;
+
+	instlo = ACA_REG__IPID__INSTANCEIDLO(bank->regs[ACA_REG_IDX_IPID]);
+	instlo &= GENMASK(31, 1);
+	switch (instlo) {
+	case mmSMNAID_XCD0_MCA_SMU:
+	case mmSMNAID_XCD1_MCA_SMU:
+	case mmSMNXCD_XCD0_MCA_SMU:
+		return true;
+	default:
+		break;
+	}
+
+	return false;
+}
+
+static const struct aca_bank_ops gfx_v9_4_3_aca_bank_ops = {
+	.aca_bank_generate_report = gfx_v9_4_3_aca_bank_generate_report,
+	.aca_bank_is_valid = gfx_v9_4_3_aca_bank_is_valid,
+};
+
+static const struct aca_info gfx_v9_4_3_aca_info = {
+	.hwip = ACA_HWIP_TYPE_SMU,
+	.mask = ACA_ERROR_UE_MASK | ACA_ERROR_CE_MASK,
+	.bank_ops = &gfx_v9_4_3_aca_bank_ops,
+};
+
 static int gfx_v9_4_3_gpu_early_init(struct amdgpu_device *adev)
 {
 	u32 gb_addr_config;
@@ -1109,7 +1174,7 @@ static int gfx_v9_4_3_rlc_init(struct amdgpu_device *adev)
 {
 	/* init spm vmid with 0xf */
 	if (adev->gfx.rlc.funcs->update_spm_vmid)
-		adev->gfx.rlc.funcs->update_spm_vmid(adev, 0xf);
+		adev->gfx.rlc.funcs->update_spm_vmid(adev, NULL, 0xf);
 
 	return 0;
 }
@@ -1320,7 +1385,7 @@ static int gfx_v9_4_3_rlc_resume(struct amdgpu_device *adev)
 	return 0;
 }
 
-static void gfx_v9_4_3_update_spm_vmid(struct amdgpu_device *adev,
+static void gfx_v9_4_3_update_spm_vmid(struct amdgpu_device *adev, struct amdgpu_ring *ring,
 				       unsigned vmid)
 {
 	u32 reg, data;
@@ -4242,9 +4307,32 @@ struct amdgpu_ras_block_hw_ops  gfx_v9_4_3_ras_ops = {
 	.reset_ras_error_count = &gfx_v9_4_3_reset_ras_error_count,
 };
 
+static int gfx_v9_4_3_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int r;
+
+	r = amdgpu_ras_block_late_init(adev, ras_block);
+	if (r)
+		return r;
+
+	r = amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__GFX,
+				&gfx_v9_4_3_aca_info,
+				NULL);
+	if (r)
+		goto late_fini;
+
+	return 0;
+
+late_fini:
+	amdgpu_ras_block_late_fini(adev, ras_block);
+
+	return r;
+}
+
 struct amdgpu_gfx_ras gfx_v9_4_3_ras = {
 	.ras_block = {
 		.hw_ops = &gfx_v9_4_3_ras_ops,
+		.ras_late_init = &gfx_v9_4_3_ras_late_init,
 	},
 	.enable_watchdog_timer = &gfx_v9_4_3_enable_watchdog_timer,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
index 6c5185608854..db89d13bd80d 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c
@@ -262,16 +262,17 @@ static void gmc_v10_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 	/* flush hdp cache */
 	adev->hdp.funcs->flush_hdp(adev, NULL);
 
-	/* For SRIOV run time, driver shouldn't access the register through MMIO
-	 * Directly use kiq to do the vm invalidation instead
+	/* This is necessary for SRIOV as well as for GFXOFF to function
+	 * properly under bare metal
 	 */
 	if (adev->gfx.kiq[0].ring.sched.ready && !adev->enable_mes &&
 	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev))) {
-		amdgpu_virt_kiq_reg_write_reg_wait(adev, req, ack, inv_req,
-				1 << vmid, GET_INST(GC, 0));
+		amdgpu_gmc_fw_reg_write_reg_wait(adev, req, ack, inv_req,
+						 1 << vmid, GET_INST(GC, 0));
 		return;
 	}
 
+	/* This path is needed before KIQ/MES/GFXOFF are set up */
 	hub_ip = (vmhub == AMDGPU_GFXHUB(0)) ? GC_HWIP : MMHUB_HWIP;
 
 	spin_lock(&adev->gmc.invalidate_lock);
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
index c9c653cfc765..6c68135cac9f 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c
@@ -223,16 +223,17 @@ static void gmc_v11_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 	/* flush hdp cache */
 	adev->hdp.funcs->flush_hdp(adev, NULL);
 
-	/* For SRIOV run time, driver shouldn't access the register through MMIO
-	 * Directly use kiq to do the vm invalidation instead
+	/* This is necessary for SRIOV as well as for GFXOFF to function
+	 * properly under bare metal
 	 */
 	if ((adev->gfx.kiq[0].ring.sched.ready || adev->mes.ring.sched.ready) &&
 	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev))) {
-		amdgpu_virt_kiq_reg_write_reg_wait(adev, req, ack, inv_req,
-				1 << vmid, GET_INST(GC, 0));
+		amdgpu_gmc_fw_reg_write_reg_wait(adev, req, ack, inv_req,
+						 1 << vmid, GET_INST(GC, 0));
 		return;
 	}
 
+	/* This path is needed before KIQ/MES/GFXOFF are set up */
 	hub_ip = (vmhub == AMDGPU_GFXHUB(0)) ? GC_HWIP : MMHUB_HWIP;
 
 	spin_lock(&adev->gmc.invalidate_lock);
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
index 59d9215e5556..23b478639921 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v6_0.c
@@ -435,9 +435,10 @@ static void gmc_v6_0_set_prt(struct amdgpu_device *adev, bool enable)
 	WREG32(mmVM_PRT_CNTL, tmp);
 
 	if (enable) {
-		uint32_t low = AMDGPU_VA_RESERVED_SIZE >> AMDGPU_GPU_PAGE_SHIFT;
+		uint32_t low = AMDGPU_VA_RESERVED_BOTTOM >>
+			AMDGPU_GPU_PAGE_SHIFT;
 		uint32_t high = adev->vm_manager.max_pfn -
-			(AMDGPU_VA_RESERVED_SIZE >> AMDGPU_GPU_PAGE_SHIFT);
+			(AMDGPU_VA_RESERVED_TOP >> AMDGPU_GPU_PAGE_SHIFT);
 
 		WREG32(mmVM_PRT_APERTURE0_LOW_ADDR, low);
 		WREG32(mmVM_PRT_APERTURE1_LOW_ADDR, low);
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
index 45a2f8e031a2..3da7b6a2b00d 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c
@@ -563,9 +563,10 @@ static void gmc_v7_0_set_prt(struct amdgpu_device *adev, bool enable)
 	WREG32(mmVM_PRT_CNTL, tmp);
 
 	if (enable) {
-		uint32_t low = AMDGPU_VA_RESERVED_SIZE >> AMDGPU_GPU_PAGE_SHIFT;
+		uint32_t low = AMDGPU_VA_RESERVED_BOTTOM >>
+			AMDGPU_GPU_PAGE_SHIFT;
 		uint32_t high = adev->vm_manager.max_pfn -
-			(AMDGPU_VA_RESERVED_SIZE >> AMDGPU_GPU_PAGE_SHIFT);
+			(AMDGPU_VA_RESERVED_TOP >> AMDGPU_GPU_PAGE_SHIFT);
 
 		WREG32(mmVM_PRT_APERTURE0_LOW_ADDR, low);
 		WREG32(mmVM_PRT_APERTURE1_LOW_ADDR, low);
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
index 4422b27a3cc2..969a9e867170 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c
@@ -777,9 +777,10 @@ static void gmc_v8_0_set_prt(struct amdgpu_device *adev, bool enable)
 	WREG32(mmVM_PRT_CNTL, tmp);
 
 	if (enable) {
-		uint32_t low = AMDGPU_VA_RESERVED_SIZE >> AMDGPU_GPU_PAGE_SHIFT;
+		uint32_t low = AMDGPU_VA_RESERVED_BOTTOM >>
+			AMDGPU_GPU_PAGE_SHIFT;
 		uint32_t high = adev->vm_manager.max_pfn -
-			(AMDGPU_VA_RESERVED_SIZE >> AMDGPU_GPU_PAGE_SHIFT);
+			(AMDGPU_VA_RESERVED_TOP >> AMDGPU_GPU_PAGE_SHIFT);
 
 		WREG32(mmVM_PRT_APERTURE0_LOW_ADDR, low);
 		WREG32(mmVM_PRT_APERTURE1_LOW_ADDR, low);
diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
index 40a00ea0009f..d442ae85162d 100644
--- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c
@@ -829,23 +829,25 @@ static void gmc_v9_0_flush_gpu_tlb(struct amdgpu_device *adev, uint32_t vmid,
 	req = hub->vm_inv_eng0_req + hub->eng_distance * eng;
 	ack = hub->vm_inv_eng0_ack + hub->eng_distance * eng;
 
-	/* This is necessary for a HW workaround under SRIOV as well
-	 * as GFXOFF under bare metal
-	 */
 	if (vmhub >= AMDGPU_MMHUB0(0))
 		inst = GET_INST(GC, 0);
 	else
 		inst = vmhub;
+
+	/* This is necessary for SRIOV as well as for GFXOFF to function
+	 * properly under bare metal
+	 */
 	if (adev->gfx.kiq[inst].ring.sched.ready &&
 	    (amdgpu_sriov_runtime(adev) || !amdgpu_sriov_vf(adev))) {
 		uint32_t req = hub->vm_inv_eng0_req + hub->eng_distance * eng;
 		uint32_t ack = hub->vm_inv_eng0_ack + hub->eng_distance * eng;
 
-		amdgpu_virt_kiq_reg_write_reg_wait(adev, req, ack, inv_req,
-						   1 << vmid, inst);
+		amdgpu_gmc_fw_reg_write_reg_wait(adev, req, ack, inv_req,
+						 1 << vmid, inst);
 		return;
 	}
 
+	/* This path is needed before KIQ/MES/GFXOFF are set up */
 	spin_lock(&adev->gmc.invalidate_lock);
 
 	/*
@@ -1947,14 +1949,6 @@ static int gmc_v9_0_init_mem_ranges(struct amdgpu_device *adev)
 
 static void gmc_v9_4_3_init_vram_info(struct amdgpu_device *adev)
 {
-	static const u32 regBIF_BIOS_SCRATCH_4 = 0x50;
-	u32 vram_info;
-
-	/* Only for dGPU, vendor informaton is reliable */
-	if (!amdgpu_sriov_vf(adev) && !(adev->flags & AMD_IS_APU)) {
-		vram_info = RREG32(regBIF_BIOS_SCRATCH_4);
-		adev->gmc.vram_vendor = vram_info & 0xF;
-	}
 	adev->gmc.vram_type = AMDGPU_VRAM_TYPE_HBM;
 	adev->gmc.vram_width = 128 * 64;
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c
index e67a337457ed..99cd49ee8ef6 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v2_5.c
@@ -551,7 +551,7 @@ static int jpeg_v2_5_set_powergating_state(void *handle,
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 	int ret;
 
-	if(state == adev->jpeg.cur_state)
+	if (state == adev->jpeg.cur_state)
 		return 0;
 
 	if (state == AMD_PG_STATE_GATE)
@@ -559,7 +559,7 @@ static int jpeg_v2_5_set_powergating_state(void *handle,
 	else
 		ret = jpeg_v2_5_start(adev);
 
-	if(!ret)
+	if (!ret)
 		adev->jpeg.cur_state = state;
 
 	return ret;
@@ -754,8 +754,7 @@ static void jpeg_v2_5_set_irq_funcs(struct amdgpu_device *adev)
 	}
 }
 
-const struct amdgpu_ip_block_version jpeg_v2_5_ip_block =
-{
+const struct amdgpu_ip_block_version jpeg_v2_5_ip_block = {
 		.type = AMD_IP_BLOCK_TYPE_JPEG,
 		.major = 2,
 		.minor = 5,
@@ -763,8 +762,7 @@ const struct amdgpu_ip_block_version jpeg_v2_5_ip_block =
 		.funcs = &jpeg_v2_5_ip_funcs,
 };
 
-const struct amdgpu_ip_block_version jpeg_v2_6_ip_block =
-{
+const struct amdgpu_ip_block_version jpeg_v2_6_ip_block = {
 		.type = AMD_IP_BLOCK_TYPE_JPEG,
 		.major = 2,
 		.minor = 6,
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c
index bc38b90f8cf8..88ea58d5c4ab 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0.c
@@ -674,14 +674,6 @@ static int jpeg_v4_0_set_powergating_state(void *handle,
 	return ret;
 }
 
-static int jpeg_v4_0_set_interrupt_state(struct amdgpu_device *adev,
-					struct amdgpu_irq_src *source,
-					unsigned type,
-					enum amdgpu_interrupt_state state)
-{
-	return 0;
-}
-
 static int jpeg_v4_0_set_ras_interrupt_state(struct amdgpu_device *adev,
 					struct amdgpu_irq_src *source,
 					unsigned int type,
@@ -765,7 +757,6 @@ static void jpeg_v4_0_set_dec_ring_funcs(struct amdgpu_device *adev)
 }
 
 static const struct amdgpu_irq_src_funcs jpeg_v4_0_irq_funcs = {
-	.set = jpeg_v4_0_set_interrupt_state,
 	.process = jpeg_v4_0_process_interrupt,
 };
 
diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c
index 6ede85b28cc8..78b74daf4eeb 100644
--- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c
+++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c
@@ -181,7 +181,6 @@ static int jpeg_v4_0_5_hw_fini(void *handle)
 			RREG32_SOC15(JPEG, 0, regUVD_JRBC_STATUS))
 			jpeg_v4_0_5_set_powergating_state(adev, AMD_PG_STATE_GATE);
 	}
-	amdgpu_irq_put(adev, &adev->jpeg.inst->irq, 0);
 
 	return 0;
 }
@@ -516,14 +515,6 @@ static int jpeg_v4_0_5_set_powergating_state(void *handle,
 	return ret;
 }
 
-static int jpeg_v4_0_5_set_interrupt_state(struct amdgpu_device *adev,
-					struct amdgpu_irq_src *source,
-					unsigned type,
-					enum amdgpu_interrupt_state state)
-{
-	return 0;
-}
-
 static int jpeg_v4_0_5_process_interrupt(struct amdgpu_device *adev,
 				      struct amdgpu_irq_src *source,
 				      struct amdgpu_iv_entry *entry)
@@ -603,7 +594,6 @@ static void jpeg_v4_0_5_set_dec_ring_funcs(struct amdgpu_device *adev)
 }
 
 static const struct amdgpu_irq_src_funcs jpeg_v4_0_5_irq_funcs = {
-	.set = jpeg_v4_0_5_set_interrupt_state,
 	.process = jpeg_v4_0_5_process_interrupt,
 };
 
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c
index fb53aacdcba2..c0fc44cdd658 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c
+++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c
@@ -33,6 +33,7 @@
 
 #define regVM_L2_CNTL3_DEFAULT	0x80100007
 #define regVM_L2_CNTL4_DEFAULT	0x000000c1
+#define mmSMNAID_AID0_MCA_SMU 0x03b30400
 
 static u64 mmhub_v1_8_get_fb_location(struct amdgpu_device *adev)
 {
@@ -705,8 +706,94 @@ static const struct amdgpu_ras_block_hw_ops mmhub_v1_8_ras_hw_ops = {
 	.reset_ras_error_count = mmhub_v1_8_reset_ras_error_count,
 };
 
+static int mmhub_v1_8_aca_bank_generate_report(struct aca_handle *handle,
+					       struct aca_bank *bank, enum aca_error_type type,
+					       struct aca_bank_report *report, void *data)
+{
+	u64 status, misc0;
+	int ret;
+
+	status = bank->regs[ACA_REG_IDX_STATUS];
+	if ((type == ACA_ERROR_TYPE_UE &&
+	     ACA_REG__STATUS__ERRORCODEEXT(status) == ACA_EXTERROR_CODE_FAULT) ||
+	    (type == ACA_ERROR_TYPE_CE &&
+	     ACA_REG__STATUS__ERRORCODEEXT(status) == ACA_EXTERROR_CODE_CE)) {
+
+		ret = aca_bank_info_decode(bank, &report->info);
+		if (ret)
+			return ret;
+
+		misc0 = bank->regs[ACA_REG_IDX_MISC0];
+		report->count[type] = ACA_REG__MISC0__ERRCNT(misc0);
+	}
+
+	return 0;
+}
+
+/* reference to smu driver if header file */
+static int mmhub_v1_8_err_codes[] = {
+	0, 1, 2, 3, 4, /* CODE_DAGB0 - 4 */
+	5, 6, 7, 8, 9, /* CODE_EA0 - 4 */
+	10, /* CODE_UTCL2_ROUTER */
+	11, /* CODE_VML2 */
+	12, /* CODE_VML2_WALKER */
+	13, /* CODE_MMCANE */
+};
+
+static bool mmhub_v1_8_aca_bank_is_valid(struct aca_handle *handle, struct aca_bank *bank,
+					 enum aca_error_type type, void *data)
+{
+	u32 instlo;
+
+	instlo = ACA_REG__IPID__INSTANCEIDLO(bank->regs[ACA_REG_IDX_IPID]);
+	instlo &= GENMASK(31, 1);
+
+	if (instlo != mmSMNAID_AID0_MCA_SMU)
+		return false;
+
+	if (aca_bank_check_error_codes(handle->adev, bank,
+				       mmhub_v1_8_err_codes,
+				       ARRAY_SIZE(mmhub_v1_8_err_codes)))
+		return false;
+
+	return true;
+}
+
+static const struct aca_bank_ops mmhub_v1_8_aca_bank_ops = {
+	.aca_bank_generate_report = mmhub_v1_8_aca_bank_generate_report,
+	.aca_bank_is_valid = mmhub_v1_8_aca_bank_is_valid,
+};
+
+static const struct aca_info mmhub_v1_8_aca_info = {
+	.hwip = ACA_HWIP_TYPE_SMU,
+	.mask = ACA_ERROR_UE_MASK,
+	.bank_ops = &mmhub_v1_8_aca_bank_ops,
+};
+
+static int mmhub_v1_8_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int r;
+
+	r = amdgpu_ras_block_late_init(adev, ras_block);
+	if (r)
+		return r;
+
+	r = amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__MMHUB,
+				&mmhub_v1_8_aca_info, NULL);
+	if (r)
+		goto late_fini;
+
+	return 0;
+
+late_fini:
+	amdgpu_ras_block_late_fini(adev, ras_block);
+
+	return r;
+}
+
 struct amdgpu_mmhub_ras mmhub_v1_8_ras = {
 	.ras_block = {
 		.hw_ops = &mmhub_v1_8_ras_hw_ops,
+		.ras_late_init = mmhub_v1_8_ras_late_init,
 	},
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
index 63725b2ebc03..a2bd2c3b1ef9 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_ai.c
@@ -404,7 +404,8 @@ static int xgpu_ai_request_init_data(struct amdgpu_device *adev)
 	return xgpu_ai_send_access_requests(adev, IDH_REQ_GPU_INIT_DATA);
 }
 
-static void xgpu_ai_ras_poison_handler(struct amdgpu_device *adev)
+static void xgpu_ai_ras_poison_handler(struct amdgpu_device *adev,
+					enum amdgpu_ras_block block)
 {
 	xgpu_ai_send_access_requests(adev, IDH_RAS_POISON);
 }
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c
index 6a68ee946f1c..77f5b55decf9 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.c
@@ -152,14 +152,14 @@ static void xgpu_nv_mailbox_trans_msg (struct amdgpu_device *adev,
 	xgpu_nv_mailbox_set_valid(adev, false);
 }
 
-static int xgpu_nv_send_access_requests(struct amdgpu_device *adev,
-					enum idh_request req)
+static int xgpu_nv_send_access_requests_with_param(struct amdgpu_device *adev,
+			enum idh_request req, u32 data1, u32 data2, u32 data3)
 {
 	int r, retry = 1;
 	enum idh_event event = -1;
 
 send_request:
-	xgpu_nv_mailbox_trans_msg(adev, req, 0, 0, 0);
+	xgpu_nv_mailbox_trans_msg(adev, req, data1, data2, data3);
 
 	switch (req) {
 	case IDH_REQ_GPU_INIT_ACCESS:
@@ -170,6 +170,10 @@ send_request:
 	case IDH_REQ_GPU_INIT_DATA:
 		event = IDH_REQ_GPU_INIT_DATA_READY;
 		break;
+	case IDH_RAS_POISON:
+		if (data1 != 0)
+			event = IDH_RAS_POISON_READY;
+		break;
 	default:
 		break;
 	}
@@ -206,6 +210,13 @@ send_request:
 	return 0;
 }
 
+static int xgpu_nv_send_access_requests(struct amdgpu_device *adev,
+					enum idh_request req)
+{
+	return xgpu_nv_send_access_requests_with_param(adev,
+						req, 0, 0, 0);
+}
+
 static int xgpu_nv_request_reset(struct amdgpu_device *adev)
 {
 	int ret, i = 0;
@@ -424,9 +435,17 @@ void xgpu_nv_mailbox_put_irq(struct amdgpu_device *adev)
 	amdgpu_irq_put(adev, &adev->virt.rcv_irq, 0);
 }
 
-static void xgpu_nv_ras_poison_handler(struct amdgpu_device *adev)
+static void xgpu_nv_ras_poison_handler(struct amdgpu_device *adev,
+		enum amdgpu_ras_block block)
 {
-	xgpu_nv_send_access_requests(adev, IDH_RAS_POISON);
+	if (amdgpu_ip_version(adev, UMC_HWIP, 0) < IP_VERSION(12, 0, 0)) {
+		xgpu_nv_send_access_requests(adev, IDH_RAS_POISON);
+	} else {
+		amdgpu_virt_fini_data_exchange(adev);
+		xgpu_nv_send_access_requests_with_param(adev,
+					IDH_RAS_POISON,	block, 0, 0);
+		amdgpu_virt_init_data_exchange(adev);
+	}
 }
 
 const struct amdgpu_virt_ops xgpu_nv_virt_ops = {
diff --git a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.h b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.h
index d0221ce08769..1e8fd90cab43 100644
--- a/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.h
+++ b/drivers/gpu/drm/amd/amdgpu/mxgpu_nv.h
@@ -51,6 +51,7 @@ enum idh_event {
 	IDH_FAIL,
 	IDH_QUERY_ALIVE,
 	IDH_REQ_GPU_INIT_DATA_READY,
+	IDH_RAS_POISON_READY,
 
 	IDH_TEXT_MESSAGE = 255,
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
index de93614726c9..4178f4e5dad7 100644
--- a/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
+++ b/drivers/gpu/drm/amd/amdgpu/navi10_ih.c
@@ -728,8 +728,7 @@ static void navi10_ih_set_interrupt_funcs(struct amdgpu_device *adev)
 		adev->irq.ih_funcs = &navi10_ih_funcs;
 }
 
-const struct amdgpu_ip_block_version navi10_ih_ip_block =
-{
+const struct amdgpu_ip_block_version navi10_ih_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_IH,
 	.major = 5,
 	.minor = 0,
diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c
index e90f33780803..b4723d68eab0 100644
--- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c
+++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c
@@ -431,6 +431,12 @@ static void nbio_v7_9_init_registers(struct amdgpu_device *adev)
 	u32 inst_mask;
 	int i;
 
+	if (amdgpu_sriov_vf(adev))
+		adev->rmmio_remap.reg_offset =
+			SOC15_REG_OFFSET(
+				NBIO, 0,
+				regBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL)
+			<< 2;
 	WREG32_SOC15(NBIO, 0, regXCC_DOORBELL_FENCE,
 		0xff & ~(adev->gfx.xcc_mask));
 
diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
index df1844d0800f..722b6066ce07 100644
--- a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c
@@ -27,6 +27,7 @@
 #include "amdgpu_ucode.h"
 #include "soc15_common.h"
 #include "psp_v13_0.h"
+#include "amdgpu_ras.h"
 
 #include "mp/mp_13_0_2_offset.h"
 #include "mp/mp_13_0_2_sh_mask.h"
@@ -187,11 +188,18 @@ static int psp_v13_0_wait_for_bootloader(struct psp_context *psp)
 static int psp_v13_0_wait_for_bootloader_steady_state(struct psp_context *psp)
 {
 	struct amdgpu_device *adev = psp->adev;
+	int ret;
 
 	if (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 6)) {
-		psp_v13_0_wait_for_vmbx_ready(psp);
+		ret = psp_v13_0_wait_for_vmbx_ready(psp);
+		if (ret)
+			amdgpu_ras_query_boot_status(adev, 4);
+
+		ret = psp_v13_0_wait_for_bootloader(psp);
+		if (ret)
+			amdgpu_ras_query_boot_status(adev, 4);
 
-		return psp_v13_0_wait_for_bootloader(psp);
+		return ret;
 	}
 
 	return 0;
@@ -763,81 +771,28 @@ static int psp_v13_0_fatal_error_recovery_quirk(struct psp_context *psp)
 	return 0;
 }
 
-
-static void psp_v13_0_boot_error_reporting(struct amdgpu_device *adev,
-					   uint32_t inst,
-					   uint32_t boot_error)
-{
-	uint32_t socket_id;
-	uint32_t aid_id;
-	uint32_t hbm_id;
-	uint32_t reg_data;
-
-	socket_id = REG_GET_FIELD(boot_error, MP0_SMN_C2PMSG_126, SOCKET_ID);
-	aid_id = REG_GET_FIELD(boot_error, MP0_SMN_C2PMSG_126, AID_ID);
-	hbm_id = REG_GET_FIELD(boot_error, MP0_SMN_C2PMSG_126, HBM_ID);
-
-	reg_data = RREG32_SOC15(MP0, inst, regMP0_SMN_C2PMSG_109);
-	dev_info(adev->dev, "socket: %d, aid: %d, firmware boot failed, fw status is 0x%x\n",
-		 socket_id, aid_id, reg_data);
-
-	if (REG_GET_FIELD(boot_error, MP0_SMN_C2PMSG_126, GPU_ERR_MEM_TRAINING))
-		dev_info(adev->dev, "socket: %d, aid: %d, hbm: %d, memory training failed\n",
-			 socket_id, aid_id, hbm_id);
-
-	if (REG_GET_FIELD(boot_error, MP0_SMN_C2PMSG_126, GPU_ERR_FW_LOAD))
-		dev_info(adev->dev, "socket: %d, aid: %d, firmware load failed at boot time\n",
-			 socket_id, aid_id);
-
-	if (REG_GET_FIELD(boot_error, MP0_SMN_C2PMSG_126, GPU_ERR_WAFL_LINK_TRAINING))
-		dev_info(adev->dev, "socket: %d, aid: %d, wafl link training failed\n",
-			 socket_id, aid_id);
-
-	if (REG_GET_FIELD(boot_error, MP0_SMN_C2PMSG_126, GPU_ERR_XGMI_LINK_TRAINING))
-		dev_info(adev->dev, "socket: %d, aid: %d, xgmi link training failed\n",
-			 socket_id, aid_id);
-
-	if (REG_GET_FIELD(boot_error, MP0_SMN_C2PMSG_126, GPU_ERR_USR_CP_LINK_TRAINING))
-		dev_info(adev->dev, "socket: %d, aid: %d, usr cp link training failed\n",
-			 socket_id, aid_id);
-
-	if (REG_GET_FIELD(boot_error, MP0_SMN_C2PMSG_126, GPU_ERR_USR_DP_LINK_TRAINING))
-		dev_info(adev->dev, "socket: %d, aid: %d, usr dp link training failed\n",
-			 socket_id, aid_id);
-
-	if (REG_GET_FIELD(boot_error, MP0_SMN_C2PMSG_126, GPU_ERR_HBM_MEM_TEST))
-		dev_info(adev->dev, "socket: %d, aid: %d, hbm: %d, hbm memory test failed\n",
-			 socket_id, aid_id, hbm_id);
-
-	if (REG_GET_FIELD(boot_error, MP0_SMN_C2PMSG_126, GPU_ERR_HBM_BIST_TEST))
-		dev_info(adev->dev, "socket: %d, aid: %d, hbm: %d, hbm bist test failed\n",
-			 socket_id, aid_id, hbm_id);
-}
-
-static int psp_v13_0_query_boot_status(struct psp_context *psp)
+static bool psp_v13_0_get_ras_capability(struct psp_context *psp)
 {
 	struct amdgpu_device *adev = psp->adev;
-	int inst_mask = adev->aid_mask;
-	uint32_t reg_data;
-	uint32_t i;
-	int ret = 0;
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	u32 reg_data;
 
-	if (amdgpu_ip_version(adev, MP0_HWIP, 0) != IP_VERSION(13, 0, 6))
-		return 0;
+	/* query ras cap should be done from host side */
+	if (amdgpu_sriov_vf(adev))
+		return false;
 
-	if (RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_59) < 0x00a10109)
-		return 0;
+	if (!con)
+		return false;
 
-	for_each_inst(i, inst_mask) {
-		reg_data = RREG32_SOC15(MP0, i, regMP0_SMN_C2PMSG_126);
-		if (!REG_GET_FIELD(reg_data, MP0_SMN_C2PMSG_126, BOOT_STATUS)) {
-			psp_v13_0_boot_error_reporting(adev, i, reg_data);
-			ret = -EINVAL;
-			break;
-		}
+	if ((amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 6)) &&
+	    (!(adev->flags & AMD_IS_APU))) {
+		reg_data = RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_127);
+		adev->ras_hw_enabled = (reg_data & GENMASK_ULL(23, 0));
+		con->poison_supported = ((reg_data & GENMASK_ULL(24, 24)) >> 24) ? true : false;
+		return true;
+	} else {
+		return false;
 	}
-
-	return ret;
 }
 
 static const struct psp_funcs psp_v13_0_funcs = {
@@ -862,7 +817,7 @@ static const struct psp_funcs psp_v13_0_funcs = {
 	.update_spirom = psp_v13_0_update_spirom,
 	.vbflash_stat = psp_v13_0_vbflash_status,
 	.fatal_error_recovery_quirk = psp_v13_0_fatal_error_recovery_quirk,
-	.query_boot_status = psp_v13_0_query_boot_status,
+	.get_ras_capability = psp_v13_0_get_ras_capability,
 };
 
 void psp_v13_0_set_psp_funcs(struct psp_context *psp)
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c
index 8d5d86675a7f..07e19caf2bc1 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c
@@ -57,22 +57,19 @@ static void sdma_v2_4_set_irq_funcs(struct amdgpu_device *adev);
 MODULE_FIRMWARE("amdgpu/topaz_sdma.bin");
 MODULE_FIRMWARE("amdgpu/topaz_sdma1.bin");
 
-static const u32 sdma_offsets[SDMA_MAX_INSTANCE] =
-{
+static const u32 sdma_offsets[SDMA_MAX_INSTANCE] = {
 	SDMA0_REGISTER_OFFSET,
 	SDMA1_REGISTER_OFFSET
 };
 
-static const u32 golden_settings_iceland_a11[] =
-{
+static const u32 golden_settings_iceland_a11[] = {
 	mmSDMA0_CHICKEN_BITS, 0xfc910007, 0x00810007,
 	mmSDMA0_CLK_CTRL, 0xff000fff, 0x00000000,
 	mmSDMA1_CHICKEN_BITS, 0xfc910007, 0x00810007,
 	mmSDMA1_CLK_CTRL, 0xff000fff, 0x00000000,
 };
 
-static const u32 iceland_mgcg_cgcg_init[] =
-{
+static const u32 iceland_mgcg_cgcg_init[] = {
 	mmSDMA0_CLK_CTRL, 0xff000ff0, 0x00000100,
 	mmSDMA1_CLK_CTRL, 0xff000ff0, 0x00000100
 };
@@ -142,7 +139,8 @@ static int sdma_v2_4_init_microcode(struct amdgpu_device *adev)
 	case CHIP_TOPAZ:
 		chip_name = "topaz";
 		break;
-	default: BUG();
+	default:
+		BUG();
 	}
 
 	for (i = 0; i < adev->sdma.num_instances; i++) {
@@ -1258,8 +1256,7 @@ static void sdma_v2_4_set_vm_pte_funcs(struct amdgpu_device *adev)
 	adev->vm_manager.vm_pte_num_scheds = adev->sdma.num_instances;
 }
 
-const struct amdgpu_ip_block_version sdma_v2_4_ip_block =
-{
+const struct amdgpu_ip_block_version sdma_v2_4_ip_block = {
 	.type = AMD_IP_BLOCK_TYPE_SDMA,
 	.major = 2,
 	.minor = 4,
diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
index 2d688dca26be..fec5a3d1c4bc 100644
--- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
+++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_4_2.c
@@ -45,6 +45,8 @@
 
 MODULE_FIRMWARE("amdgpu/sdma_4_4_2.bin");
 
+#define mmSMNAID_AID0_MCA_SMU 0x03b30400
+
 #define WREG32_SDMA(instance, offset, value) \
 	WREG32(sdma_v4_4_2_get_reg_offset(adev, (instance), (offset)), value)
 #define RREG32_SDMA(instance, offset) \
@@ -2204,9 +2206,79 @@ static const struct amdgpu_ras_block_hw_ops sdma_v4_4_2_ras_hw_ops = {
 	.reset_ras_error_count = sdma_v4_4_2_reset_ras_error_count,
 };
 
+static int sdma_v4_4_2_aca_bank_generate_report(struct aca_handle *handle,
+						struct aca_bank *bank, enum aca_error_type type,
+						struct aca_bank_report *report, void *data)
+{
+	u64 status, misc0;
+	int ret;
+
+	status = bank->regs[ACA_REG_IDX_STATUS];
+	if ((type == ACA_ERROR_TYPE_UE &&
+	     ACA_REG__STATUS__ERRORCODEEXT(status) == ACA_EXTERROR_CODE_FAULT) ||
+	    (type == ACA_ERROR_TYPE_CE &&
+	     ACA_REG__STATUS__ERRORCODEEXT(status) == ACA_EXTERROR_CODE_CE)) {
+
+		ret = aca_bank_info_decode(bank, &report->info);
+		if (ret)
+			return ret;
+
+		misc0 = bank->regs[ACA_REG_IDX_MISC0];
+		report->count[type] = ACA_REG__MISC0__ERRCNT(misc0);
+	}
+
+	return 0;
+}
+
+/* CODE_SDMA0 - CODE_SDMA4, reference to smu driver if header file */
+static int sdma_v4_4_2_err_codes[] = { 33, 34, 35, 36 };
+
+static bool sdma_v4_4_2_aca_bank_is_valid(struct aca_handle *handle, struct aca_bank *bank,
+					  enum aca_error_type type, void *data)
+{
+	u32 instlo;
+
+	instlo = ACA_REG__IPID__INSTANCEIDLO(bank->regs[ACA_REG_IDX_IPID]);
+	instlo &= GENMASK(31, 1);
+
+	if (instlo != mmSMNAID_AID0_MCA_SMU)
+		return false;
+
+	if (aca_bank_check_error_codes(handle->adev, bank,
+				       sdma_v4_4_2_err_codes,
+				       ARRAY_SIZE(sdma_v4_4_2_err_codes)))
+		return false;
+
+	return true;
+}
+
+static const struct aca_bank_ops sdma_v4_4_2_aca_bank_ops = {
+	.aca_bank_generate_report = sdma_v4_4_2_aca_bank_generate_report,
+	.aca_bank_is_valid = sdma_v4_4_2_aca_bank_is_valid,
+};
+
+static const struct aca_info sdma_v4_4_2_aca_info = {
+	.hwip = ACA_HWIP_TYPE_SMU,
+	.mask = ACA_ERROR_UE_MASK,
+	.bank_ops = &sdma_v4_4_2_aca_bank_ops,
+};
+
+static int sdma_v4_4_2_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int r;
+
+	r = amdgpu_sdma_ras_late_init(adev, ras_block);
+	if (r)
+		return r;
+
+	return amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__SDMA,
+				   &sdma_v4_4_2_aca_info, NULL);
+}
+
 static struct amdgpu_sdma_ras sdma_v4_4_2_ras = {
 	.ras_block = {
 		.hw_ops = &sdma_v4_4_2_ras_hw_ops,
+		.ras_late_init = sdma_v4_4_2_ras_late_init,
 	},
 };
 
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c
index 15033efec2ba..c64c01e2944a 100644
--- a/drivers/gpu/drm/amd/amdgpu/soc15.c
+++ b/drivers/gpu/drm/amd/amdgpu/soc15.c
@@ -1298,10 +1298,32 @@ static int soc15_common_suspend(void *handle)
 	return soc15_common_hw_fini(adev);
 }
 
+static bool soc15_need_reset_on_resume(struct amdgpu_device *adev)
+{
+	u32 sol_reg;
+
+	sol_reg = RREG32_SOC15(MP0, 0, mmMP0_SMN_C2PMSG_81);
+
+	/* Will reset for the following suspend abort cases.
+	 * 1) Only reset limit on APU side, dGPU hasn't checked yet.
+	 * 2) S3 suspend abort and TOS already launched.
+	 */
+	if (adev->flags & AMD_IS_APU && adev->in_s3 &&
+			!adev->suspend_complete &&
+			sol_reg)
+		return true;
+
+	return false;
+}
+
 static int soc15_common_resume(void *handle)
 {
 	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
 
+	if (soc15_need_reset_on_resume(adev)) {
+		dev_info(adev->dev, "S3 suspend abort case, let's reset ASIC.\n");
+		soc15_asic_reset(adev);
+	}
 	return soc15_common_hw_init(adev);
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h b/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h
index 879bb7af297c..056d4df8fa1f 100644
--- a/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h
+++ b/drivers/gpu/drm/amd/amdgpu/ta_ras_if.h
@@ -36,6 +36,9 @@ enum ras_command {
 	TA_RAS_COMMAND__ENABLE_FEATURES = 0,
 	TA_RAS_COMMAND__DISABLE_FEATURES,
 	TA_RAS_COMMAND__TRIGGER_ERROR,
+	TA_RAS_COMMAND__QUERY_BLOCK_INFO,
+	TA_RAS_COMMAND__QUERY_SUB_BLOCK_INFO,
+	TA_RAS_COMMAND__QUERY_ADDRESS,
 };
 
 enum ta_ras_status {
@@ -105,6 +108,11 @@ enum ta_ras_error_type {
 	TA_RAS_ERROR__POISON			= 8,
 };
 
+enum ta_ras_address_type {
+	TA_RAS_MCA_TO_PA,
+	TA_RAS_PA_TO_MCA,
+};
+
 /* Input/output structures for RAS commands */
 /**********************************************************/
 
@@ -133,12 +141,38 @@ struct ta_ras_init_flags {
 	uint8_t channel_dis_num;
 };
 
+struct ta_ras_mca_addr {
+	uint64_t err_addr;
+	uint32_t ch_inst;
+	uint32_t umc_inst;
+	uint32_t node_inst;
+};
+
+struct ta_ras_phy_addr {
+	uint64_t pa;
+	uint32_t bank;
+	uint32_t channel_idx;
+};
+
+struct ta_ras_query_address_input {
+	enum ta_ras_address_type addr_type;
+	struct ta_ras_mca_addr ma;
+	struct ta_ras_phy_addr pa;
+};
+
 struct ta_ras_output_flags {
 	uint8_t ras_init_success_flag;
 	uint8_t err_inject_switch_disable_flag;
 	uint8_t reg_access_failure_flag;
 };
 
+struct ta_ras_query_address_output {
+	/* don't use the flags here */
+	struct ta_ras_output_flags flags;
+	struct ta_ras_mca_addr ma;
+	struct ta_ras_phy_addr pa;
+};
+
 /* Common input structure for RAS callbacks */
 /**********************************************************/
 union ta_ras_cmd_input {
@@ -146,12 +180,14 @@ union ta_ras_cmd_input {
 	struct ta_ras_enable_features_input	enable_features;
 	struct ta_ras_disable_features_input	disable_features;
 	struct ta_ras_trigger_error_input	trigger_error;
+	struct ta_ras_query_address_input	address;
 
 	uint32_t reserve_pad[256];
 };
 
 union ta_ras_cmd_output {
 	struct ta_ras_output_flags flags;
+	struct ta_ras_query_address_output address;
 
 	uint32_t reserve_pad[256];
 };
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c
index 7458a218e89d..14ef7a24be7b 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.c
@@ -89,12 +89,28 @@ static void umc_v12_0_reset_error_count(struct amdgpu_device *adev)
 		umc_v12_0_reset_error_count_per_channel, NULL);
 }
 
+bool umc_v12_0_is_deferred_error(struct amdgpu_device *adev, uint64_t mc_umc_status)
+{
+	dev_info(adev->dev,
+		"MCA_UMC_STATUS(0x%llx): Val:%llu, Poison:%llu, Deferred:%llu, PCC:%llu, UC:%llu, TCC:%llu\n",
+		mc_umc_status,
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val),
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Poison),
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred),
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, PCC),
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UC),
+		REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, TCC)
+	);
+
+	return (amdgpu_ras_is_poison_mode_supported(adev) &&
+		(REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1) &&
+		(REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred) == 1));
+}
+
 bool umc_v12_0_is_uncorrectable_error(struct amdgpu_device *adev, uint64_t mc_umc_status)
 {
-	if (amdgpu_ras_is_poison_mode_supported(adev) &&
-	    (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1) &&
-	    (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred) == 1))
-		return true;
+	if (umc_v12_0_is_deferred_error(adev, mc_umc_status))
+		return false;
 
 	return ((REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1) &&
 		(REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, PCC) == 1 ||
@@ -104,9 +120,7 @@ bool umc_v12_0_is_uncorrectable_error(struct amdgpu_device *adev, uint64_t mc_um
 
 bool umc_v12_0_is_correctable_error(struct amdgpu_device *adev, uint64_t mc_umc_status)
 {
-	if (amdgpu_ras_is_poison_mode_supported(adev) &&
-	    (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1) &&
-	    (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Deferred) == 1))
+	if (umc_v12_0_is_deferred_error(adev, mc_umc_status))
 		return false;
 
 	return (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
@@ -119,9 +133,10 @@ bool umc_v12_0_is_correctable_error(struct amdgpu_device *adev, uint64_t mc_umc_
 		!(umc_v12_0_is_uncorrectable_error(adev, mc_umc_status)))));
 }
 
-static void umc_v12_0_query_correctable_error_count(struct amdgpu_device *adev,
+static void umc_v12_0_query_error_count_per_type(struct amdgpu_device *adev,
 						   uint64_t umc_reg_offset,
-						   unsigned long *error_count)
+						   unsigned long *error_count,
+						   check_error_type_func error_type_func)
 {
 	uint64_t mc_umc_status;
 	uint64_t mc_umc_status_addr;
@@ -129,31 +144,11 @@ static void umc_v12_0_query_correctable_error_count(struct amdgpu_device *adev,
 	mc_umc_status_addr =
 		SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_STATUST0);
 
-	/* Rely on MCUMC_STATUS for correctable error counter
-	 * MCUMC_STATUS is a 64 bit register
-	 */
+	/* Check MCUMC_STATUS */
 	mc_umc_status =
 		RREG64_PCIE_EXT((mc_umc_status_addr + umc_reg_offset) * 4);
 
-	if (umc_v12_0_is_correctable_error(adev, mc_umc_status))
-		*error_count += 1;
-}
-
-static void umc_v12_0_query_uncorrectable_error_count(struct amdgpu_device *adev,
-						      uint64_t umc_reg_offset,
-						      unsigned long *error_count)
-{
-	uint64_t mc_umc_status;
-	uint64_t mc_umc_status_addr;
-
-	mc_umc_status_addr =
-		SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_STATUST0);
-
-	/* Check the MCUMC_STATUS. */
-	mc_umc_status =
-		RREG64_PCIE_EXT((mc_umc_status_addr + umc_reg_offset) * 4);
-
-	if (umc_v12_0_is_uncorrectable_error(adev, mc_umc_status))
+	if (error_type_func(adev, mc_umc_status))
 		*error_count += 1;
 }
 
@@ -162,7 +157,7 @@ static int umc_v12_0_query_error_count(struct amdgpu_device *adev,
 					uint32_t ch_inst, void *data)
 {
 	struct ras_err_data *err_data = (struct ras_err_data *)data;
-	unsigned long ue_count = 0, ce_count = 0;
+	unsigned long ue_count = 0, ce_count = 0, de_count = 0;
 
 	/* NOTE: node_inst is converted by adev->umc.active_mask and the range is [0-3],
 	 * which can be used as die ID directly */
@@ -174,11 +169,16 @@ static int umc_v12_0_query_error_count(struct amdgpu_device *adev,
 	uint64_t umc_reg_offset =
 		get_umc_v12_0_reg_offset(adev, node_inst, umc_inst, ch_inst);
 
-	umc_v12_0_query_correctable_error_count(adev, umc_reg_offset, &ce_count);
-	umc_v12_0_query_uncorrectable_error_count(adev, umc_reg_offset, &ue_count);
+	umc_v12_0_query_error_count_per_type(adev, umc_reg_offset,
+					    &ce_count, umc_v12_0_is_correctable_error);
+	umc_v12_0_query_error_count_per_type(adev, umc_reg_offset,
+					    &ue_count, umc_v12_0_is_uncorrectable_error);
+	umc_v12_0_query_error_count_per_type(adev, umc_reg_offset,
+					    &de_count, umc_v12_0_is_deferred_error);
 
 	amdgpu_ras_error_statistic_ue_count(err_data, &mcm_info, NULL, ue_count);
 	amdgpu_ras_error_statistic_ce_count(err_data, &mcm_info, NULL, ce_count);
+	amdgpu_ras_error_statistic_de_count(err_data, &mcm_info, NULL, de_count);
 
 	return 0;
 }
@@ -203,14 +203,14 @@ static bool umc_v12_0_bit_wise_xor(uint32_t val)
 	return result;
 }
 
-static void umc_v12_0_convert_error_address(struct amdgpu_device *adev,
-					    struct ras_err_data *err_data, uint64_t err_addr,
-					    uint32_t ch_inst, uint32_t umc_inst,
-					    uint32_t node_inst)
+static void umc_v12_0_mca_addr_to_pa(struct amdgpu_device *adev,
+					uint64_t err_addr, uint32_t ch_inst, uint32_t umc_inst,
+					uint32_t node_inst,
+					struct ta_ras_query_address_output *addr_out)
 {
 	uint32_t channel_index, i;
-	uint64_t soc_pa, na, retired_page, column;
-	uint32_t bank_hash0, bank_hash1, bank_hash2, bank_hash3, col, row, row_xor;
+	uint64_t na, soc_pa;
+	uint32_t bank_hash0, bank_hash1, bank_hash2, bank_hash3, col, row;
 	uint32_t bank0, bank1, bank2, bank3, bank;
 
 	bank_hash0 = (err_addr >> UMC_V12_0_MCA_B0_BIT) & 0x1ULL;
@@ -260,12 +260,44 @@ static void umc_v12_0_convert_error_address(struct amdgpu_device *adev,
 	/* the umc channel bits are not original values, they are hashed */
 	UMC_V12_0_SET_CHANNEL_HASH(channel_index, soc_pa);
 
+	addr_out->pa.pa = soc_pa;
+	addr_out->pa.bank = bank;
+	addr_out->pa.channel_idx = channel_index;
+}
+
+static void umc_v12_0_convert_error_address(struct amdgpu_device *adev,
+					    struct ras_err_data *err_data, uint64_t err_addr,
+					    uint32_t ch_inst, uint32_t umc_inst,
+					    uint32_t node_inst)
+{
+	uint32_t col, row, row_xor, bank, channel_index;
+	uint64_t soc_pa, retired_page, column;
+	struct ta_ras_query_address_input addr_in;
+	struct ta_ras_query_address_output addr_out;
+
+	addr_in.addr_type = TA_RAS_MCA_TO_PA;
+	addr_in.ma.err_addr = err_addr;
+	addr_in.ma.ch_inst = ch_inst;
+	addr_in.ma.umc_inst = umc_inst;
+	addr_in.ma.node_inst = node_inst;
+
+	if (psp_ras_query_address(&adev->psp, &addr_in, &addr_out))
+		/* fallback to old path if fail to get pa from psp */
+		umc_v12_0_mca_addr_to_pa(adev, err_addr, ch_inst, umc_inst,
+				node_inst, &addr_out);
+
+	soc_pa = addr_out.pa.pa;
+	bank = addr_out.pa.bank;
+	channel_index = addr_out.pa.channel_idx;
+
+	col = (err_addr >> 1) & 0x1fULL;
+	row = (err_addr >> 10) & 0x3fffULL;
+	row_xor = row ^ (0x1ULL << 13);
 	/* clear [C3 C2] in soc physical address */
 	soc_pa &= ~(0x3ULL << UMC_V12_0_PA_C2_BIT);
 	/* clear [C4] in soc physical address */
 	soc_pa &= ~(0x1ULL << UMC_V12_0_PA_C4_BIT);
 
-	row_xor = row ^ (0x1ULL << 13);
 	/* loop for all possibilities of [C4 C3 C2] */
 	for (column = 0; column < UMC_V12_0_NA_MAP_PA_NUM; column++) {
 		retired_page = soc_pa | ((column & 0x3) << UMC_V12_0_PA_C2_BIT);
@@ -316,10 +348,7 @@ static int umc_v12_0_query_error_address(struct amdgpu_device *adev,
 	}
 
 	/* calculate error address if ue error is detected */
-	if (REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, Val) == 1 &&
-	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, AddrV) == 1 &&
-	    REG_GET_FIELD(mc_umc_status, MCA_UMC_UMC0_MCUMC_STATUST0, UC) == 1) {
-
+	if (umc_v12_0_is_uncorrectable_error(adev, mc_umc_status)) {
 		mc_umc_addrt0 =
 			SOC15_REG_OFFSET(UMC, 0, regMCA_UMC_UMC0_MCUMC_ADDRT0);
 
@@ -385,45 +414,69 @@ static void umc_v12_0_ecc_info_query_ras_error_address(struct amdgpu_device *ade
 {
 	struct ras_err_node *err_node;
 	uint64_t mc_umc_status;
+	struct ras_err_info *err_info;
+	struct ras_err_addr *mca_err_addr, *tmp;
 	struct ras_err_data *err_data = (struct ras_err_data *)ras_error_status;
 
 	for_each_ras_error(err_node, err_data) {
-		mc_umc_status = err_node->err_info.err_addr.err_status;
-		if (!mc_umc_status)
+		err_info = &err_node->err_info;
+		if (list_empty(&err_info->err_addr_list))
 			continue;
 
-		if (umc_v12_0_is_uncorrectable_error(adev, mc_umc_status)) {
-			uint64_t mca_addr, err_addr, mca_ipid;
-			uint32_t InstanceIdLo;
-			struct amdgpu_smuio_mcm_config_info *mcm_info;
-
-			mcm_info = &err_node->err_info.mcm_info;
-			mca_addr = err_node->err_info.err_addr.err_addr;
-			mca_ipid = err_node->err_info.err_addr.err_ipid;
-
-			err_addr =  REG_GET_FIELD(mca_addr, MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr);
-			InstanceIdLo = REG_GET_FIELD(mca_ipid, MCMP1_IPIDT0, InstanceIdLo);
-
-			dev_info(adev->dev, "UMC:IPID:0x%llx, aid:%d, inst:%d, ch:%d, err_addr:0x%llx\n",
-				mca_ipid,
-				mcm_info->die_id,
-				MCA_IPID_LO_2_UMC_INST(InstanceIdLo),
-				MCA_IPID_LO_2_UMC_CH(InstanceIdLo),
-				err_addr);
-
-			umc_v12_0_convert_error_address(adev,
-				err_data, err_addr,
-				MCA_IPID_LO_2_UMC_CH(InstanceIdLo),
-				MCA_IPID_LO_2_UMC_INST(InstanceIdLo),
-				mcm_info->die_id);
-
-			/* Clear umc error address content */
-			memset(&err_node->err_info.err_addr,
-				0, sizeof(err_node->err_info.err_addr));
+		list_for_each_entry_safe(mca_err_addr, tmp, &err_info->err_addr_list, node) {
+			mc_umc_status = mca_err_addr->err_status;
+			if (mc_umc_status &&
+				(umc_v12_0_is_uncorrectable_error(adev, mc_umc_status) ||
+				 umc_v12_0_is_deferred_error(adev, mc_umc_status))) {
+				uint64_t mca_addr, err_addr, mca_ipid;
+				uint32_t InstanceIdLo;
+
+				mca_addr = mca_err_addr->err_addr;
+				mca_ipid = mca_err_addr->err_ipid;
+
+				err_addr = REG_GET_FIELD(mca_addr,
+							MCA_UMC_UMC0_MCUMC_ADDRT0, ErrorAddr);
+				InstanceIdLo = REG_GET_FIELD(mca_ipid, MCMP1_IPIDT0, InstanceIdLo);
+
+				dev_info(adev->dev, "UMC:IPID:0x%llx, aid:%d, inst:%d, ch:%d, err_addr:0x%llx\n",
+					mca_ipid,
+					err_info->mcm_info.die_id,
+					MCA_IPID_LO_2_UMC_INST(InstanceIdLo),
+					MCA_IPID_LO_2_UMC_CH(InstanceIdLo),
+					err_addr);
+
+				umc_v12_0_convert_error_address(adev,
+					err_data, err_addr,
+					MCA_IPID_LO_2_UMC_CH(InstanceIdLo),
+					MCA_IPID_LO_2_UMC_INST(InstanceIdLo),
+					err_info->mcm_info.die_id);
+			}
+
+			/* Delete error address node from list and free memory */
+			amdgpu_ras_del_mca_err_addr(err_info, mca_err_addr);
 		}
 	}
 }
 
+static bool umc_v12_0_check_ecc_err_status(struct amdgpu_device *adev,
+			enum amdgpu_mca_error_type type, void *ras_error_status)
+{
+	uint64_t mc_umc_status = *(uint64_t *)ras_error_status;
+
+	switch (type) {
+	case AMDGPU_MCA_ERROR_TYPE_UE:
+		return umc_v12_0_is_uncorrectable_error(adev, mc_umc_status);
+	case AMDGPU_MCA_ERROR_TYPE_CE:
+		return umc_v12_0_is_correctable_error(adev, mc_umc_status);
+	case AMDGPU_MCA_ERROR_TYPE_DE:
+		return umc_v12_0_is_deferred_error(adev, mc_umc_status);
+	default:
+		return false;
+	}
+
+	return false;
+}
+
 static void umc_v12_0_err_cnt_init(struct amdgpu_device *adev)
 {
 	amdgpu_umc_loop_channels(adev,
@@ -444,12 +497,71 @@ const struct amdgpu_ras_block_hw_ops umc_v12_0_ras_hw_ops = {
 	.query_ras_error_address = umc_v12_0_query_ras_error_address,
 };
 
+static int umc_v12_0_aca_bank_generate_report(struct aca_handle *handle, struct aca_bank *bank, enum aca_error_type type,
+					      struct aca_bank_report *report, void *data)
+{
+	struct amdgpu_device *adev = handle->adev;
+	u64 status;
+	int ret;
+
+	ret = aca_bank_info_decode(bank, &report->info);
+	if (ret)
+		return ret;
+
+	status = bank->regs[ACA_REG_IDX_STATUS];
+	switch (type) {
+	case ACA_ERROR_TYPE_UE:
+		if (umc_v12_0_is_uncorrectable_error(adev, status)) {
+			report->count[type] = 1;
+		}
+		break;
+	case ACA_ERROR_TYPE_CE:
+		if (umc_v12_0_is_correctable_error(adev, status)) {
+			report->count[type] = 1;
+		}
+		break;
+	default:
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+static const struct aca_bank_ops umc_v12_0_aca_bank_ops = {
+	.aca_bank_generate_report = umc_v12_0_aca_bank_generate_report,
+};
+
+const struct aca_info umc_v12_0_aca_info = {
+	.hwip = ACA_HWIP_TYPE_UMC,
+	.mask = ACA_ERROR_UE_MASK | ACA_ERROR_CE_MASK,
+	.bank_ops = &umc_v12_0_aca_bank_ops,
+};
+
+static int umc_v12_0_ras_late_init(struct amdgpu_device *adev, struct ras_common_if *ras_block)
+{
+	int ret;
+
+	ret = amdgpu_umc_ras_late_init(adev, ras_block);
+	if (ret)
+		return ret;
+
+	ret = amdgpu_ras_bind_aca(adev, AMDGPU_RAS_BLOCK__UMC,
+				  &umc_v12_0_aca_info, NULL);
+	if (ret)
+		return ret;
+
+	return 0;
+}
+
 struct amdgpu_umc_ras umc_v12_0_ras = {
 	.ras_block = {
 		.hw_ops = &umc_v12_0_ras_hw_ops,
+		.ras_late_init = umc_v12_0_ras_late_init,
 	},
 	.err_cnt_init = umc_v12_0_err_cnt_init,
 	.query_ras_poison_mode = umc_v12_0_query_ras_poison_mode,
 	.ecc_info_query_ras_error_count = umc_v12_0_ecc_info_query_ras_error_count,
 	.ecc_info_query_ras_error_address = umc_v12_0_ecc_info_query_ras_error_address,
+	.check_ecc_err_status = umc_v12_0_check_ecc_err_status,
 };
+
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.h b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.h
index e8de3a92251a..5973bfb14fce 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v12_0.h
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v12_0.h
@@ -121,9 +121,12 @@
 			(((_ipid_lo) >> 12) & 0xF))
 #define MCA_IPID_LO_2_UMC_INST(_ipid_lo) (((_ipid_lo) >> 21) & 0x7)
 
+bool umc_v12_0_is_deferred_error(struct amdgpu_device *adev, uint64_t mc_umc_status);
 bool umc_v12_0_is_uncorrectable_error(struct amdgpu_device *adev, uint64_t mc_umc_status);
 bool umc_v12_0_is_correctable_error(struct amdgpu_device *adev, uint64_t mc_umc_status);
 
+typedef bool (*check_error_type_func)(struct amdgpu_device *adev, uint64_t mc_umc_status);
+
 extern const uint32_t
 	umc_v12_0_channel_idx_tbl[]
 			[UMC_V12_0_UMC_INSTANCE_NUM]
diff --git a/drivers/gpu/drm/amd/amdgpu/umc_v6_0.c b/drivers/gpu/drm/amd/amdgpu/umc_v6_0.c
index 0d6b50528d76..97fa88ed770c 100644
--- a/drivers/gpu/drm/amd/amdgpu/umc_v6_0.c
+++ b/drivers/gpu/drm/amd/amdgpu/umc_v6_0.c
@@ -25,7 +25,7 @@
 
 static void umc_v6_0_init_registers(struct amdgpu_device *adev)
 {
-	unsigned i,j;
+	unsigned i, j;
 
 	for (i = 0; i < 4; i++)
 		for (j = 0; j < 4; j++)