Merge tag 'drm-next-2025-01-17' of https://gitlab.freedesktop.org/drm/kernel

Pull drm updates from Dave Airlie:
 "There are two external interactions of note, the msm tree pull in some
  opp tree, hopefully the opp tree arrives from the same git tree
  however it normally does.

  There is also a new cgroup controller for device memory, that is used
  by drm, so is merging through my tree. This will hopefully help open
  up gpu cgroup usage a bit more and move us forward.

  There is a new accelerator driver for the AMD XDNA Ryzen AI NPUs.

  Then the usual xe/amdgpu/i915/msm leaders and lots of changes and
  refactors across the board:

  core:
   - device memory cgroup controller added
   - Remove driver date from drm_driver
   - Add drm_printer based hex dumper
   - drm memory stats docs update
   - scheduler documentation improvements

  new driver:
   - amdxdna - Ryzen AI NPU support

  connector:
   - add a mutex to protect ELD
   - make connector setup two-step

  panels:
   - Introduce backlight quirks infrastructure
   - New panels: KDB KD116N2130B12, Tianma TM070JDHG34-00,
   - Multi-Inno Technology MI1010Z1T-1CP11

  bridge:
   - ti-sn65dsi83: Add ti,lvds-vod-swing optional properties
   - Provide default implementation of atomic_check for HDMI bridges
   - it605: HDCP improvements, MCCS Support

  xe:
   - make OA buffer size configurable
   - GuC capture fixes
   - add ufence and g2h flushes
   - restore system memory GGTT mappings
   - ioctl fixes
   - SRIOV PF scheduling priority
   - allow fault injection
   - lots of improvements/refactors
   - Enable GuC's WA_DUAL_QUEUE for newer platforms
   - IRQ related fixes and improvements

  i915:
   - More accurate engine busyness metrics with GuC submission
   - Ensure partial BO segment offset never exceeds allowed max
   - Flush GuC CT receive tasklet during reset preparation
   - Some DG2 refactor to fix DG2 bugs when operating with certain CPUs
   - Fix DG1 power gate sequence
   - Enabling uncompressed 128b/132b UHBR SST
   - Handle hdmi connector init failures, and no HDMI/DP cases
   - More robust engine resets on Haswell and older

  i915/xe display:
   - HDCP fixes for Xe3Lpd
   - New GSC FW ARL-H/ARL-U
   - support 3 VDSC engines 12 slices
   - MBUS joining sanitisation
   - reconcile i915/xe display power mgmt
   - Xe3Lpd fixes
   - UHBR rates for Thunderbolt

  amdgpu:
   - DRM panic support
   - track BO memory stats at runtime
   - Fix max surface handling in DC
   - Cleaner shader support for gfx10.3 dGPUs
   - fix drm buddy trim handling
   - SDMA engine reset updates
   - Fix doorbell ttm cleanup
   - RAS updates
   - ISP updates
   - SDMA queue reset support
   - Rework DPM powergating interfaces
   - Documentation updates and cleanups
   - DCN 3.5 updates
   - Use a pm notifier to more gracefully handle VRAM eviction on
     suspend or hibernate
   - Add debugfs interfaces for forcing scheduling to specific engine
     instances
   - GG 9.5 updates
   - IH 4.4 updates
   - Make missing optional firmware less noisy
   - PSP 13.x updates
   - SMU 13.x updates
   - VCN 5.x updates
   - JPEG 5.x updates
   - GC 12.x updates
   - DC FAMS updates

  amdkfd:
   - GG 9.5 updates
   - Logging improvements
   - Shader debugger fixes
   - Trap handler cleanup
   - Cleanup includes
   - Eviction fence wq fix

  msm:
   - MDSS:
      - properly described UBWC registers
      - added SM6150 (aka QCS615) support
   - DPU:
      - added SM6150 (aka QCS615) support
      - enabled wide planes if virtual planes are enabled (by using two
        SSPPs for a single plane)
      - added CWB hardware blocks support
   - DSI:
      - added SM6150 (aka QCS615) support
   - GPU:
      - Print GMU core fw version
      - GMU bandwidth voting for a740 and a750
      - Expose uche trap base via uapi
      - UAPI error reporting

  rcar-du:
   - Add r8a779h0 Support

  ivpu:
   - Fix qemu crash when using passthrough

  nouveau:
   - expose GSP-RM logging buffers via debugfs

  panfrost:
   - Add MT8188 Mali-G57 MC3 support

  rockchip:
   - Gamma LUT support

  hisilicon:
   - new HIBMC support

  virtio-gpu:
   - convert to helpers
   - add prime support for scanout buffers

  v3d:
   - Add DRM_IOCTL_V3D_PERFMON_SET_GLOBAL

  vc4:
   - Add support for BCM2712

  vkms:
   - line-per-line compositing algorithm to improve performance

  zynqmp:
   - Add DP audio support

  mediatek:
   - dp: Add sdp path reset
   - dp: Support flexible length of DP calibration data

  etnaviv:
   - add fdinfo memory support
   - add explicit reset handling"

* tag 'drm-next-2025-01-17' of https://gitlab.freedesktop.org/drm/kernel: (1070 commits)
  drm/bridge: fix documentation for the hdmi_audio_prepare() callback
  doc/cgroup: Fix title underline length
  drm/doc: Include new drm-compute documentation
  cgroup/dmem: Fix parameters documentation
  cgroup/dmem: Select PAGE_COUNTER
  kernel/cgroup: Remove the unused variable climit
  drm/display: hdmi: Do not read EDID on disconnected connectors
  drm/tests: hdmi: Add connector disablement test
  drm/connector: hdmi: Do atomic check when necessary
  drm/amd/display: 3.2.316
  drm/amd/display: avoid reset DTBCLK at clock init
  drm/amd/display: improve dpia pre-train
  drm/amd/display: Apply DML21 Patches
  drm/amd/display: Use HW lock mgr for PSR1
  drm/amd/display: Revised for Replay Pseudo vblank control
  drm/amd/display: Add a new flag for replay low hz
  drm/amd/display: Remove unused read_ono_state function from Hwss module
  drm/amd/display: Do not elevate mem_type change to full update
  drm/amd/display: Do not wait for PSR disable on vbl enable
  drm/amd/display: Remove unnecessary eDP power down
  ...

1 files changed, 308 insertions, 37 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index 4c9fa24dd972..f0924aa3f4e4 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -36,6 +36,7 @@
 #include "amdgpu_xgmi.h"
 #include "ivsrcid/nbio/irqsrcs_nbif_7_4.h"
 #include "nbio_v4_3.h"
+#include "nbif_v6_3_1.h"
 #include "nbio_v7_9.h"
 #include "atom.h"
 #include "amdgpu_reset.h"
@@ -192,7 +193,7 @@ static int amdgpu_reserve_page_direct(struct amdgpu_device *adev, uint64_t addre
 
 	if (amdgpu_bad_page_threshold != 0) {
 		amdgpu_ras_add_bad_pages(adev, err_data.err_addr,
-					 err_data.err_addr_cnt);
+					 err_data.err_addr_cnt, false);
 		amdgpu_ras_save_bad_pages(adev, NULL);
 	}
 
@@ -2015,6 +2016,7 @@ static bool amdgpu_ras_aca_is_supported(struct amdgpu_device *adev)
 
 	switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
 	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 12):
 	case IP_VERSION(13, 0, 14):
 		ret = true;
 		break;
@@ -2156,6 +2158,16 @@ void amdgpu_ras_interrupt_fatal_error_handler(struct amdgpu_device *adev)
 	/* Fatal error events are handled on host side */
 	if (amdgpu_sriov_vf(adev))
 		return;
+	/**
+	 * If the current interrupt is caused by a non-fatal RAS error, skip
+	 * check for fatal error. For fatal errors, FED status of all devices
+	 * in XGMI hive gets set when the first device gets fatal error
+	 * interrupt. The error gets propagated to other devices as well, so
+	 * make sure to ack the interrupt regardless of FED status.
+	 */
+	if (!amdgpu_ras_get_fed_status(adev) &&
+	    amdgpu_ras_is_err_state(adev, AMDGPU_RAS_BLOCK__ANY))
+		return;
 
 	if (adev->nbio.ras &&
 	    adev->nbio.ras->handle_ras_controller_intr_no_bifring)
@@ -2185,6 +2197,7 @@ static void amdgpu_ras_interrupt_poison_consumption_handler(struct ras_manager *
 	if (ret)
 		return;
 
+	amdgpu_ras_set_err_poison(adev, block_obj->ras_comm.block);
 	/* both query_poison_status and handle_poison_consumption are optional,
 	 * but at least one of them should be implemented if we need poison
 	 * consumption handler
@@ -2717,40 +2730,203 @@ static int amdgpu_ras_realloc_eh_data_space(struct amdgpu_device *adev,
 	return 0;
 }
 
+static int amdgpu_ras_mca2pa_by_idx(struct amdgpu_device *adev,
+			struct eeprom_table_record *bps,
+			struct ras_err_data *err_data)
+{
+	struct ta_ras_query_address_input addr_in;
+	uint32_t socket = 0;
+	int ret = 0;
+
+	if (adev->smuio.funcs && adev->smuio.funcs->get_socket_id)
+		socket = adev->smuio.funcs->get_socket_id(adev);
+
+	/* reinit err_data */
+	err_data->err_addr_cnt = 0;
+	err_data->err_addr_len = adev->umc.retire_unit;
+
+	memset(&addr_in, 0, sizeof(addr_in));
+	addr_in.ma.err_addr = bps->address;
+	addr_in.ma.socket_id = socket;
+	addr_in.ma.ch_inst = bps->mem_channel;
+	/* tell RAS TA the node instance is not used */
+	addr_in.ma.node_inst = TA_RAS_INV_NODE;
+
+	if (adev->umc.ras && adev->umc.ras->convert_ras_err_addr)
+		ret = adev->umc.ras->convert_ras_err_addr(adev, err_data,
+				&addr_in, NULL, false);
+
+	return ret;
+}
+
+static int amdgpu_ras_mca2pa(struct amdgpu_device *adev,
+			struct eeprom_table_record *bps,
+			struct ras_err_data *err_data)
+{
+	struct ta_ras_query_address_input addr_in;
+	uint32_t die_id, socket = 0;
+
+	if (adev->smuio.funcs && adev->smuio.funcs->get_socket_id)
+		socket = adev->smuio.funcs->get_socket_id(adev);
+
+	/* although die id is gotten from PA in nps1 mode, the id is
+	 * fitable for any nps mode
+	 */
+	if (adev->umc.ras && adev->umc.ras->get_die_id_from_pa)
+		die_id = adev->umc.ras->get_die_id_from_pa(adev, bps->address,
+					bps->retired_page << AMDGPU_GPU_PAGE_SHIFT);
+	else
+		return -EINVAL;
+
+	/* reinit err_data */
+	err_data->err_addr_cnt = 0;
+	err_data->err_addr_len = adev->umc.retire_unit;
+
+	memset(&addr_in, 0, sizeof(addr_in));
+	addr_in.ma.err_addr = bps->address;
+	addr_in.ma.ch_inst = bps->mem_channel;
+	addr_in.ma.umc_inst = bps->mcumc_id;
+	addr_in.ma.node_inst = die_id;
+	addr_in.ma.socket_id = socket;
+
+	if (adev->umc.ras && adev->umc.ras->convert_ras_err_addr)
+		return adev->umc.ras->convert_ras_err_addr(adev, err_data,
+					&addr_in, NULL, false);
+	else
+		return  -EINVAL;
+}
+
 /* it deal with vram only. */
 int amdgpu_ras_add_bad_pages(struct amdgpu_device *adev,
-		struct eeprom_table_record *bps, int pages)
+		struct eeprom_table_record *bps, int pages, bool from_rom)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct ras_err_handler_data *data;
+	struct ras_err_data err_data;
+	struct eeprom_table_record *err_rec;
+	struct amdgpu_ras_eeprom_control *control =
+			&adev->psp.ras_context.ras->eeprom_control;
+	enum amdgpu_memory_partition nps = AMDGPU_NPS1_PARTITION_MODE;
 	int ret = 0;
-	uint32_t i;
+	uint32_t i, j, loop_cnt = 1;
+	bool find_pages_per_pa = false;
 
 	if (!con || !con->eh_data || !bps || pages <= 0)
 		return 0;
 
+	if (from_rom) {
+		err_data.err_addr =
+			kcalloc(adev->umc.retire_unit,
+				sizeof(struct eeprom_table_record), GFP_KERNEL);
+		if (!err_data.err_addr) {
+			dev_warn(adev->dev, "Failed to alloc UMC error address record in mca2pa conversion!\n");
+			ret = -ENOMEM;
+			goto out;
+		}
+
+		err_rec = err_data.err_addr;
+		loop_cnt = adev->umc.retire_unit;
+		if (adev->gmc.gmc_funcs->query_mem_partition_mode)
+			nps = adev->gmc.gmc_funcs->query_mem_partition_mode(adev);
+	}
+
 	mutex_lock(&con->recovery_lock);
 	data = con->eh_data;
-	if (!data)
-		goto out;
+	if (!data) {
+		/* Returning 0 as the absence of eh_data is acceptable */
+		goto free;
+	}
 
 	for (i = 0; i < pages; i++) {
-		if (amdgpu_ras_check_bad_page_unlock(con,
-			bps[i].retired_page << AMDGPU_GPU_PAGE_SHIFT))
-			continue;
+		if (from_rom &&
+		    control->rec_type == AMDGPU_RAS_EEPROM_REC_MCA) {
+			if (!find_pages_per_pa) {
+				if (amdgpu_ras_mca2pa_by_idx(adev, &bps[i], &err_data)) {
+					if (!i && nps == AMDGPU_NPS1_PARTITION_MODE) {
+						/* may use old RAS TA, use PA to find pages in
+						 * one row
+						 */
+						if (amdgpu_umc_pages_in_a_row(adev, &err_data,
+									      bps[i].retired_page <<
+									      AMDGPU_GPU_PAGE_SHIFT)) {
+							ret = -EINVAL;
+							goto free;
+						} else {
+							find_pages_per_pa = true;
+						}
+					} else {
+						/* unsupported cases */
+						ret = -EOPNOTSUPP;
+						goto free;
+					}
+				}
+			} else {
+				if (amdgpu_umc_pages_in_a_row(adev, &err_data,
+						bps[i].retired_page << AMDGPU_GPU_PAGE_SHIFT)) {
+					ret = -EINVAL;
+					goto free;
+				}
+			}
+		} else {
+			if (from_rom && !find_pages_per_pa) {
+				if (bps[i].retired_page & UMC_CHANNEL_IDX_V2) {
+					/* bad page in any NPS mode in eeprom */
+					if (amdgpu_ras_mca2pa_by_idx(adev, &bps[i], &err_data)) {
+						ret = -EINVAL;
+						goto free;
+					}
+				} else {
+					/* legacy bad page in eeprom, generated only in
+					 * NPS1 mode
+					 */
+					if (amdgpu_ras_mca2pa(adev, &bps[i], &err_data)) {
+						/* old RAS TA or ASICs which don't support to
+						 * convert addrss via mca address
+						 */
+						if (!i && nps == AMDGPU_NPS1_PARTITION_MODE) {
+							find_pages_per_pa = true;
+							err_rec = &bps[i];
+							loop_cnt = 1;
+						} else {
+							/* non-nps1 mode, old RAS TA
+							 * can't support it
+							 */
+							ret = -EOPNOTSUPP;
+							goto free;
+						}
+					}
+				}
 
-		if (!data->space_left &&
-			amdgpu_ras_realloc_eh_data_space(adev, data, 256)) {
-			ret = -ENOMEM;
-			goto out;
+				if (!find_pages_per_pa)
+					i += (adev->umc.retire_unit - 1);
+			} else {
+				err_rec = &bps[i];
+			}
 		}
 
-		amdgpu_ras_reserve_page(adev, bps[i].retired_page);
+		for (j = 0; j < loop_cnt; j++) {
+			if (amdgpu_ras_check_bad_page_unlock(con,
+				err_rec[j].retired_page << AMDGPU_GPU_PAGE_SHIFT))
+				continue;
+
+			if (!data->space_left &&
+			    amdgpu_ras_realloc_eh_data_space(adev, data, 256)) {
+				ret = -ENOMEM;
+				goto free;
+			}
 
-		memcpy(&data->bps[data->count], &bps[i], sizeof(*data->bps));
-		data->count++;
-		data->space_left--;
+			amdgpu_ras_reserve_page(adev, err_rec[j].retired_page);
+
+			memcpy(&data->bps[data->count], &(err_rec[j]),
+					sizeof(struct eeprom_table_record));
+			data->count++;
+			data->space_left--;
+		}
 	}
+
+free:
+	if (from_rom)
+		kfree(err_data.err_addr);
 out:
 	mutex_unlock(&con->recovery_lock);
 
@@ -2768,7 +2944,7 @@ int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev,
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	struct ras_err_handler_data *data;
 	struct amdgpu_ras_eeprom_control *control;
-	int save_count;
+	int save_count, unit_num, bad_page_num, i;
 
 	if (!con || !con->eh_data) {
 		if (new_cnt)
@@ -2780,19 +2956,32 @@ int amdgpu_ras_save_bad_pages(struct amdgpu_device *adev,
 	mutex_lock(&con->recovery_lock);
 	control = &con->eeprom_control;
 	data = con->eh_data;
-	save_count = data->count - control->ras_num_recs;
+	bad_page_num = control->ras_num_bad_pages;
+	save_count = data->count - bad_page_num;
 	mutex_unlock(&con->recovery_lock);
 
+	unit_num = save_count / adev->umc.retire_unit;
 	if (new_cnt)
-		*new_cnt = save_count / adev->umc.retire_unit;
+		*new_cnt = unit_num;
 
 	/* only new entries are saved */
 	if (save_count > 0) {
-		if (amdgpu_ras_eeprom_append(control,
-					     &data->bps[control->ras_num_recs],
-					     save_count)) {
-			dev_err(adev->dev, "Failed to save EEPROM table data!");
-			return -EIO;
+		if (control->rec_type == AMDGPU_RAS_EEPROM_REC_PA) {
+			if (amdgpu_ras_eeprom_append(control,
+						     &data->bps[control->ras_num_recs],
+						     save_count)) {
+				dev_err(adev->dev, "Failed to save EEPROM table data!");
+				return -EIO;
+			}
+		} else {
+			for (i = 0; i < unit_num; i++) {
+				if (amdgpu_ras_eeprom_append(control,
+						&data->bps[bad_page_num + i * adev->umc.retire_unit],
+						1)) {
+					dev_err(adev->dev, "Failed to save EEPROM table data!");
+					return -EIO;
+				}
+			}
 		}
 
 		dev_info(adev->dev, "Saved %d pages to EEPROM table.\n", save_count);
@@ -2821,11 +3010,32 @@ static int amdgpu_ras_load_bad_pages(struct amdgpu_device *adev)
 		return -ENOMEM;
 
 	ret = amdgpu_ras_eeprom_read(control, bps, control->ras_num_recs);
-	if (ret)
+	if (ret) {
 		dev_err(adev->dev, "Failed to load EEPROM table records!");
-	else
-		ret = amdgpu_ras_add_bad_pages(adev, bps, control->ras_num_recs);
+	} else {
+		if (control->ras_num_recs > 1 &&
+		    adev->umc.ras && adev->umc.ras->convert_ras_err_addr) {
+			if ((bps[0].address == bps[1].address) &&
+			    (bps[0].mem_channel == bps[1].mem_channel))
+				control->rec_type = AMDGPU_RAS_EEPROM_REC_PA;
+			else
+				control->rec_type = AMDGPU_RAS_EEPROM_REC_MCA;
+		}
+
+		ret = amdgpu_ras_eeprom_check(control);
+		if (ret)
+			goto out;
+
+		/* HW not usable */
+		if (amdgpu_ras_is_rma(adev)) {
+			ret = -EHWPOISON;
+			goto out;
+		}
 
+		ret = amdgpu_ras_add_bad_pages(adev, bps, control->ras_num_recs, true);
+	}
+
+out:
 	kfree(bps);
 	return ret;
 }
@@ -3205,31 +3415,36 @@ static int amdgpu_ras_page_retirement_thread(void *param)
 int amdgpu_ras_init_badpage_info(struct amdgpu_device *adev)
 {
 	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
+	struct amdgpu_ras_eeprom_control *control;
 	int ret;
 
 	if (!con || amdgpu_sriov_vf(adev))
 		return 0;
 
-	ret = amdgpu_ras_eeprom_init(&con->eeprom_control);
-
+	control = &con->eeprom_control;
+	ret = amdgpu_ras_eeprom_init(control);
 	if (ret)
 		return ret;
 
-	/* HW not usable */
-	if (amdgpu_ras_is_rma(adev))
-		return -EHWPOISON;
+	if (!adev->umc.ras || !adev->umc.ras->convert_ras_err_addr)
+		control->rec_type = AMDGPU_RAS_EEPROM_REC_PA;
+
+	/* default status is MCA storage */
+	if (control->ras_num_recs <= 1 &&
+	    adev->umc.ras && adev->umc.ras->convert_ras_err_addr)
+		control->rec_type = AMDGPU_RAS_EEPROM_REC_MCA;
 
-	if (con->eeprom_control.ras_num_recs) {
+	if (control->ras_num_recs) {
 		ret = amdgpu_ras_load_bad_pages(adev);
 		if (ret)
 			return ret;
 
 		amdgpu_dpm_send_hbm_bad_pages_num(
-			adev, con->eeprom_control.ras_num_recs);
+			adev, control->ras_num_bad_pages);
 
 		if (con->update_channel_flag == true) {
 			amdgpu_dpm_send_hbm_bad_channel_flag(
-				adev, con->eeprom_control.bad_channel_bitmap);
+				adev, control->bad_channel_bitmap);
 			con->update_channel_flag = false;
 		}
 	}
@@ -3366,6 +3581,7 @@ static bool amdgpu_ras_asic_supported(struct amdgpu_device *adev)
 		switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
 		case IP_VERSION(13, 0, 2):
 		case IP_VERSION(13, 0, 6):
+		case IP_VERSION(13, 0, 12):
 		case IP_VERSION(13, 0, 14):
 			return true;
 		default:
@@ -3378,7 +3594,9 @@ static bool amdgpu_ras_asic_supported(struct amdgpu_device *adev)
 		case IP_VERSION(13, 0, 0):
 		case IP_VERSION(13, 0, 6):
 		case IP_VERSION(13, 0, 10):
+		case IP_VERSION(13, 0, 12):
 		case IP_VERSION(13, 0, 14):
+		case IP_VERSION(14, 0, 3):
 			return true;
 		default:
 			return false;
@@ -3629,6 +3847,7 @@ static void amdgpu_ras_init_reserved_vram_size(struct amdgpu_device *adev)
 	switch (amdgpu_ip_version(adev, MP0_HWIP, 0)) {
 	case IP_VERSION(13, 0, 2):
 	case IP_VERSION(13, 0, 6):
+	case IP_VERSION(13, 0, 12):
 	case IP_VERSION(13, 0, 14):
 		con->reserved_pages_in_bytes = AMDGPU_RAS_RESERVED_VRAM_SIZE;
 		break;
@@ -3704,7 +3923,19 @@ int amdgpu_ras_init(struct amdgpu_device *adev)
 			 * check DF RAS */
 			adev->nbio.ras = &nbio_v4_3_ras;
 		break;
+	case IP_VERSION(6, 3, 1):
+		if (adev->ras_hw_enabled & (1 << AMDGPU_RAS_BLOCK__DF))
+			/* unlike other generation of nbio ras,
+			 * nbif v6_3_1 only support fatal error interrupt
+			 * to inform software that DF is freezed due to
+			 * system fatal error event. driver should not
+			 * enable nbio ras in such case. Instead,
+			 * check DF RAS
+			 */
+			adev->nbio.ras = &nbif_v6_3_1_ras;
+		break;
 	case IP_VERSION(7, 9, 0):
+	case IP_VERSION(7, 9, 1):
 		if (!adev->gmc.is_app_apu)
 			adev->nbio.ras = &nbio_v7_9_ras;
 		break;
@@ -4083,7 +4314,7 @@ bool amdgpu_ras_get_fed_status(struct amdgpu_device *adev)
 	if (!ras)
 		return false;
 
-	return atomic_read(&ras->fed);
+	return test_bit(AMDGPU_RAS_BLOCK__LAST, &ras->ras_err_state);
 }
 
 void amdgpu_ras_set_fed(struct amdgpu_device *adev, bool status)
@@ -4091,8 +4322,48 @@ void amdgpu_ras_set_fed(struct amdgpu_device *adev, bool status)
 	struct amdgpu_ras *ras;
 
 	ras = amdgpu_ras_get_context(adev);
+	if (ras) {
+		if (status)
+			set_bit(AMDGPU_RAS_BLOCK__LAST, &ras->ras_err_state);
+		else
+			clear_bit(AMDGPU_RAS_BLOCK__LAST, &ras->ras_err_state);
+	}
+}
+
+void amdgpu_ras_clear_err_state(struct amdgpu_device *adev)
+{
+	struct amdgpu_ras *ras;
+
+	ras = amdgpu_ras_get_context(adev);
+	if (ras)
+		ras->ras_err_state = 0;
+}
+
+void amdgpu_ras_set_err_poison(struct amdgpu_device *adev,
+			       enum amdgpu_ras_block block)
+{
+	struct amdgpu_ras *ras;
+
+	ras = amdgpu_ras_get_context(adev);
 	if (ras)
-		atomic_set(&ras->fed, !!status);
+		set_bit(block, &ras->ras_err_state);
+}
+
+bool amdgpu_ras_is_err_state(struct amdgpu_device *adev, int block)
+{
+	struct amdgpu_ras *ras;
+
+	ras = amdgpu_ras_get_context(adev);
+	if (ras) {
+		if (block == AMDGPU_RAS_BLOCK__ANY)
+			return (ras->ras_err_state != 0);
+		else
+			return test_bit(block, &ras->ras_err_state) ||
+			       test_bit(AMDGPU_RAS_BLOCK__LAST,
+					&ras->ras_err_state);
+	}
+
+	return false;
 }
 
 static struct ras_event_manager *__get_ras_event_mgr(struct amdgpu_device *adev)