From: Tao Zhou <tao.zhou1@amd.com>
Date: Mon, 8 Sep 2025 12:39:49 +0000 (+0800)
Subject: drm/amdgpu: skip writing eeprom when PMFW manages RAS data
X-Git-Url: http://git.ipfire.org/gitweb.cgi?a=commitdiff_plain;h=541414065c59db785000d7661d3d07184e104ec2;p=thirdparty%2Fkernel%2Flinux.git

drm/amdgpu: skip writing eeprom when PMFW manages RAS data

Only update bad page number in legacy eeprom write path.

v2: add null pointer check for con.

Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
---

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
index 9854238ce7bfa..670c0dedf4e92 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c
@@ -871,6 +871,18 @@ Out:
 	return res;
 }
 
+int amdgpu_ras_eeprom_update_record_num(struct amdgpu_ras_eeprom_control *control)
+{
+	struct amdgpu_device *adev = to_amdgpu_device(control);
+
+	if (!amdgpu_ras_smu_eeprom_supported(adev))
+		return 0;
+
+	control->ras_num_recs_old = control->ras_num_recs;
+	return amdgpu_ras_smu_get_badpage_count(adev,
+			&(control->ras_num_recs), 12);
+}
+
 /**
  * amdgpu_ras_eeprom_append -- append records to the EEPROM RAS table
  * @control: pointer to control structure
@@ -889,12 +901,18 @@ int amdgpu_ras_eeprom_append(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 res, i;
 	uint64_t nps = AMDGPU_NPS1_PARTITION_MODE;
 
-	if (!__is_ras_eeprom_supported(adev))
+	if (!__is_ras_eeprom_supported(adev) || !con)
 		return 0;
 
+	if (amdgpu_ras_smu_eeprom_supported(adev)) {
+		control->ras_num_bad_pages = con->bad_page_num;
+		return 0;
+	}
+
 	if (num == 0) {
 		dev_err(adev->dev, "will not append 0 records\n");
 		return -EINVAL;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
index e881007f715b0..2e5d63957e714 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.h
@@ -82,6 +82,7 @@ struct amdgpu_ras_eeprom_control {
 	/* Number of records in the table.
 	 */
 	u32 ras_num_recs;
+	u32 ras_num_recs_old;
 
 	/* the bad page number is ras_num_recs or
 	 * ras_num_recs * umc.retire_unit
@@ -190,6 +191,8 @@ int amdgpu_ras_eeprom_read_idx(struct amdgpu_ras_eeprom_control *control,
 			struct eeprom_table_record *record, u32 rec_idx,
 			const u32 num);
 
+int amdgpu_ras_eeprom_update_record_num(struct amdgpu_ras_eeprom_control *control);
+
 extern const struct file_operations amdgpu_ras_debugfs_eeprom_size_ops;
 extern const struct file_operations amdgpu_ras_debugfs_eeprom_table_ops;