drm/amdgpu: message smu to update hbm bad page number

Use SMU to update the bad pages rather than directly
accessing the EEPROM from the driver.

Signed-off-by: Stanley.Yang <Stanley.Yang@amd.com>
Reviewed-by: John Clements <john.clements@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

4 files changed, 28 insertions, 0 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
index 6d1e6005bd87..c13b02caf8c3 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c
@@ -1984,6 +1984,9 @@ int amdgpu_ras_recovery_init(struct amdgpu_device *adev)
 		ret = amdgpu_ras_load_bad_pages(adev);
 		if (ret)
 			goto free;
+
+		if (adev->smu.ppt_funcs && adev->smu.ppt_funcs->send_hbm_bad_pages_num)
+			adev->smu.ppt_funcs->send_hbm_bad_pages_num(&adev->smu, con->eeprom_control.num_recs);
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
index ea6f99be070b..f4489773715e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_umc.c
@@ -94,6 +94,7 @@ int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev,
 		struct amdgpu_iv_entry *entry)
 {
 	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);
 	if (adev->umc.ras_funcs &&
@@ -131,6 +132,9 @@ int amdgpu_umc_process_ras_data_cb(struct amdgpu_device *adev,
 			amdgpu_ras_add_bad_pages(adev, err_data->err_addr,
 						err_data->err_addr_cnt);
 			amdgpu_ras_save_bad_pages(adev);
+
+			if (adev->smu.ppt_funcs && adev->smu.ppt_funcs->send_hbm_bad_pages_num)
+				adev->smu.ppt_funcs->send_hbm_bad_pages_num(&adev->smu, con->eeprom_control.num_recs);
 		}
 
 		amdgpu_ras_reset_gpu(adev);
diff --git a/drivers/gpu/drm/amd/pm/inc/amdgpu_smu.h b/drivers/gpu/drm/amd/pm/inc/amdgpu_smu.h
index 6559912ba229..3e89852e4820 100644
--- a/drivers/gpu/drm/amd/pm/inc/amdgpu_smu.h
+++ b/drivers/gpu/drm/amd/pm/inc/amdgpu_smu.h
@@ -1232,6 +1232,12 @@ struct pptable_funcs {
 	 */
 	int (*wait_for_event)(struct smu_context *smu,
 			      enum smu_event_type event, uint64_t event_arg);
+
+	/**
+	 * @sned_hbm_bad_pages_num:  message SMU to update bad page number
+	 *										of SMUBUS table.
+	 */
+	int (*send_hbm_bad_pages_num)(struct smu_context *smu, uint32_t size);
 };
 
 typedef enum {
diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c
index 82099c528ccb..9316a726195c 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c
@@ -1923,6 +1923,20 @@ static int aldebaran_set_mp1_state(struct smu_context *smu,
 	}
 }
 
+static int aldebaran_smu_send_hbm_bad_page_num(struct smu_context *smu,
+		uint32_t size)
+{
+	int ret = 0;
+
+	/* message SMU to update the bad page number on SMUBUS */
+	ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetNumBadHbmPagesRetired, size, NULL);
+	if (ret)
+		dev_err(smu->adev->dev, "[%s] failed to message SMU to update HBM bad pages number\n",
+				__func__);
+
+	return ret;
+}
+
 static const struct pptable_funcs aldebaran_ppt_funcs = {
 	/* init dpm */
 	.get_allowed_feature_mask = aldebaran_get_allowed_feature_mask,
@@ -1985,6 +1999,7 @@ static const struct pptable_funcs aldebaran_ppt_funcs = {
 	.wait_for_event = smu_v13_0_wait_for_event,
 	.i2c_init = aldebaran_i2c_control_init,
 	.i2c_fini = aldebaran_i2c_control_fini,
+	.send_hbm_bad_pages_num = aldebaran_smu_send_hbm_bad_page_num,
 };
 
 void aldebaran_set_ppt_funcs(struct smu_context *smu)