diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index 8177de0c18155..2b0f69760500b 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -2066,6 +2066,7 @@ int amdgpu_ras_recovery_init(struct amdgpu_device *adev)
 	mutex_init(&con->recovery_lock);
 	INIT_WORK(&con->recovery_work, amdgpu_ras_do_recovery);
 	atomic_set(&con->in_recovery, 0);
+	con->eeprom_control.bad_channel_bitmap = 0;
 
 	max_eeprom_records_count = amdgpu_ras_eeprom_max_record_count();
 	amdgpu_ras_validate_threshold(adev, max_eeprom_records_count);
@@ -2090,6 +2091,11 @@ int amdgpu_ras_recovery_init(struct amdgpu_device *adev)
 			goto free;
 
 		amdgpu_dpm_send_hbm_bad_pages_num(adev, con->eeprom_control.ras_num_recs);
+
+		if (con->update_channel_flag == true) {
+			amdgpu_dpm_send_hbm_bad_channel_flag(adev, con->eeprom_control.bad_channel_bitmap);
+			con->update_channel_flag = false;
+		}
 	}
 
 #ifdef HAVE_SMCA_UMC_V2
@@ -2284,6 +2290,7 @@ int amdgpu_ras_init(struct amdgpu_device *adev)
 		goto release_con;
 	}
 
+	con->update_channel_flag = false;
 	con->features = 0;
 	INIT_LIST_HEAD(&con->head);
 	/* Might need get this flag from vbios. */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
index 72352b069a477..a7c522bf58bf9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.h
@@ -375,6 +375,9 @@ struct amdgpu_ras {
 
 	/* record umc error info queried from smu */
 	struct umc_ecc_info umc_ecc;
+
+	/* Indicates smu whether need update bad channel info */
+	bool update_channel_flag;
 };
 
 struct ras_fs_data {
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
index 2b844a5aafdbf..ad5d8667756d6 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
@@ -265,6 +265,7 @@ int amdgpu_ras_eeprom_reset_table(struct amdgpu_ras_eeprom_control *control)
 {
 	struct amdgpu_device *adev = to_amdgpu_device(control);
 	struct amdgpu_ras_eeprom_table_header *hdr = &control->tbl_hdr;
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	u8 csum;
 	int res;
 
@@ -285,6 +286,10 @@ int amdgpu_ras_eeprom_reset_table(struct amdgpu_ras_eeprom_control *control)
 
 	amdgpu_dpm_send_hbm_bad_pages_num(adev, control->ras_num_recs);
 
+	control->bad_channel_bitmap = 0;
+	amdgpu_dpm_send_hbm_bad_channel_flag(adev, control->bad_channel_bitmap);
+	con->update_channel_flag = false;
+
 	amdgpu_ras_debugfs_set_ret_size(control);
 
 	mutex_unlock(&control->ras_tbl_mutex);
@@ -418,6 +423,7 @@ amdgpu_ras_eeprom_append_table(struct amdgpu_ras_eeprom_control *control,
 			       struct eeprom_table_record *record,
 			       const u32 num)
 {
+	struct amdgpu_ras *con = amdgpu_ras_get_context(to_amdgpu_device(control));
 	u32 a, b, i;
 	u8 *buf, *pp;
 	int res;
@@ -429,9 +435,16 @@ amdgpu_ras_eeprom_append_table(struct amdgpu_ras_eeprom_control *control,
 	/* Encode all of them in one go.
 	 */
 	pp = buf;
-	for (i = 0; i < num; i++, pp += RAS_TABLE_RECORD_SIZE)
+	for (i = 0; i < num; i++, pp += RAS_TABLE_RECORD_SIZE) {
 		__encode_table_record_to_buf(control, &record[i], pp);
 
+		/* update bad channel bitmap */
+		if (!(control->bad_channel_bitmap & (1 << record[i].mem_channel))) {
+			control->bad_channel_bitmap |= 1 << record[i].mem_channel;
+			con->update_channel_flag = true;
+		}
+	}
+
 	/* a, first record index to write into.
 	 * b, last record index to write into.
 	 * a = first index to read (fri) + number of records in the table,
@@ -684,6 +697,7 @@ int amdgpu_ras_eeprom_read(struct amdgpu_ras_eeprom_control *control,
 			   const u32 num)
 {
 	struct amdgpu_device *adev = to_amdgpu_device(control);
+	struct amdgpu_ras *con = amdgpu_ras_get_context(adev);
 	int i, res;
 	u8 *buf, *pp;
 	u32 g0, g1;
@@ -751,8 +765,15 @@ int amdgpu_ras_eeprom_read(struct amdgpu_ras_eeprom_control *control,
 	/* Read up everything? Then transform.
 	 */
 	pp = buf;
-	for (i = 0; i < num; i++, pp += RAS_TABLE_RECORD_SIZE)
+	for (i = 0; i < num; i++, pp += RAS_TABLE_RECORD_SIZE) {
 		__decode_table_record_from_buf(control, &record[i], pp);
+
+		/* update bad channel bitmap */
+		if (!(control->bad_channel_bitmap & (1 << record[i].mem_channel))) {
+			control->bad_channel_bitmap |= 1 << record[i].mem_channel;
+			con->update_channel_flag = true;
+		}
+	}
 Out:
 	kfree(buf);
 	mutex_unlock(&control->ras_tbl_mutex);
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
index 6bb00578bfbbc..54d9bfe0881d9 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
@@ -80,6 +80,10 @@ struct amdgpu_ras_eeprom_control {
 	/* Protect table access via this mutex.
 	 */
 	struct mutex ras_tbl_mutex;
+
+	/* Record channel info which occurred bad pages
+	 */
+	u32 bad_channel_bitmap;
 };
 
 /*
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
index 9400260e3263e..d90c0fd0ea043 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
@@ -97,6 +97,11 @@ static int amdgpu_umc_do_page_retirement(struct amdgpu_device *adev,
 			amdgpu_ras_save_bad_pages(adev);
 
 			amdgpu_dpm_send_hbm_bad_pages_num(adev, con->eeprom_control.ras_num_recs);
+
+			if (con->update_channel_flag == true) {
+				amdgpu_dpm_send_hbm_bad_channel_flag(adev, con->eeprom_control.bad_channel_bitmap);
+				con->update_channel_flag = false;
+			}
 		}
 
 		if (reset)