drm/amd/pm: add send bad channel info function

support message SMU update bad channel info to update HBM bad channel
info in OOB table

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

7 files changed, 77 insertions, 2 deletions

diff --git a/drivers/gpu/drm/amd/pm/amdgpu_dpm.c b/drivers/gpu/drm/amd/pm/amdgpu_dpm.c
index 1d63f1e8884c..9a892d6d1d7a 100644
--- a/drivers/gpu/drm/amd/pm/amdgpu_dpm.c
+++ b/drivers/gpu/drm/amd/pm/amdgpu_dpm.c
@@ -507,6 +507,18 @@ int amdgpu_dpm_send_hbm_bad_pages_num(struct amdgpu_device *adev, uint32_t size)
 	return ret;
 }
 
+int amdgpu_dpm_send_hbm_bad_channel_flag(struct amdgpu_device *adev, uint32_t size)
+{
+	struct smu_context *smu = adev->powerplay.pp_handle;
+	int ret = 0;
+
+	mutex_lock(&adev->pm.mutex);
+	ret = smu_send_hbm_bad_channel_flag(smu, size);
+	mutex_unlock(&adev->pm.mutex);
+
+	return ret;
+}
+
 int amdgpu_dpm_get_dpm_freq_range(struct amdgpu_device *adev,
 				  enum pp_clock_type type,
 				  uint32_t *min,
diff --git a/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h b/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h
index ddfa55b59d02..3e78b3057277 100644
--- a/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h
+++ b/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h
@@ -412,6 +412,7 @@ void amdgpu_dpm_enable_jpeg(struct amdgpu_device *adev, bool enable);
 int amdgpu_pm_load_smu_firmware(struct amdgpu_device *adev, uint32_t *smu_version);
 int amdgpu_dpm_handle_passthrough_sbr(struct amdgpu_device *adev, bool enable);
 int amdgpu_dpm_send_hbm_bad_pages_num(struct amdgpu_device *adev, uint32_t size);
+int amdgpu_dpm_send_hbm_bad_channel_flag(struct amdgpu_device *adev, uint32_t size);
 int amdgpu_dpm_get_dpm_freq_range(struct amdgpu_device *adev,
 				       enum pp_clock_type type,
 				       uint32_t *min,
diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
index 7e79a67bb8ef..f1544755d8b4 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
@@ -3052,3 +3052,13 @@ int smu_send_hbm_bad_pages_num(struct smu_context *smu, uint32_t size)
 
 	return ret;
 }
+
+int smu_send_hbm_bad_channel_flag(struct smu_context *smu, uint32_t size)
+{
+	int ret = 0;
+
+	if (smu->ppt_funcs && smu->ppt_funcs->send_hbm_bad_channel_flag)
+		ret = smu->ppt_funcs->send_hbm_bad_channel_flag(smu, size);
+
+	return ret;
+}
diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h b/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h
index fbef3ab8d487..ef57b6089c69 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h
+++ b/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h
@@ -1292,6 +1292,12 @@ struct pptable_funcs {
 	 * @set_config_table: Apply the input DriverSmuConfig table settings.
 	 */
 	int (*set_config_table)(struct smu_context *smu, struct config_table_setting *table);
+
+	/**
+	 * @sned_hbm_bad_channel_flag:  message SMU to update bad channel info
+	 *										of SMUBUS table.
+	 */
+	int (*send_hbm_bad_channel_flag)(struct smu_context *smu, uint32_t size);
 };
 
 typedef enum {
@@ -1428,5 +1434,6 @@ int smu_get_ecc_info(struct smu_context *smu, void *umc_ecc);
 int smu_stb_collect_info(struct smu_context *smu, void *buff, uint32_t size);
 void amdgpu_smu_stb_debug_fs_init(struct amdgpu_device *adev);
 int smu_send_hbm_bad_pages_num(struct smu_context *smu, uint32_t size);
+int smu_send_hbm_bad_channel_flag(struct smu_context *smu, uint32_t size);
 #endif
 #endif
diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/aldebaran_ppsmc.h b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/aldebaran_ppsmc.h
index ab66a4b9e438..0f498baf6838 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/aldebaran_ppsmc.h
+++ b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/aldebaran_ppsmc.h
@@ -103,7 +103,8 @@
 #define PPSMC_MSG_GfxDriverResetRecovery	0x42
 #define PPSMC_MSG_BoardPowerCalibration 	0x43
 #define PPSMC_MSG_HeavySBR                      0x45
-#define PPSMC_Message_Count			0x46
+#define PPSMC_MSG_SetBadHBMPagesRetiredFlagsPerChannel	0x46
+#define PPSMC_Message_Count			0x47
 
 
 //PPSMC Reset Types
diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_types.h b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_types.h
index d787c3b9fc52..9f6f306eeca0 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/inc/smu_types.h
+++ b/drivers/gpu/drm/amd/pm/swsmu/inc/smu_types.h
@@ -232,7 +232,8 @@
 	__SMU_DUMMY_MAP(ForceGfxVid),             \
 	__SMU_DUMMY_MAP(Spare0),                  \
 	__SMU_DUMMY_MAP(UnforceGfxVid),           \
-	__SMU_DUMMY_MAP(HeavySBR),
+	__SMU_DUMMY_MAP(HeavySBR),			\
+	__SMU_DUMMY_MAP(SetBadHBMPagesRetiredFlagsPerChannel),
 
 #undef __SMU_DUMMY_MAP
 #define __SMU_DUMMY_MAP(type)	SMU_MSG_##type
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 890acc4e2cb8..cd81f848d45a 100644
--- a/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c
@@ -82,6 +82,12 @@
  */
 #define SUPPORT_ECCTABLE_SMU_VERSION 0x00442a00
 
+/*
+ * SMU support BAD CHENNEL info MSG since version 68.51.00,
+ * use this to check ECCTALE feature whether support
+ */
+#define SUPPORT_BAD_CHANNEL_INFO_MSG_VERSION 0x00443300
+
 static const struct smu_temperature_range smu13_thermal_policy[] =
 {
 	{-273150,  99000, 99000, -273150, 99000, 99000, -273150, 99000, 99000},
@@ -140,6 +146,7 @@ static const struct cmn2asic_msg_mapping aldebaran_message_map[SMU_MSG_MAX_COUNT
 	MSG_MAP(GfxDriverResetRecovery,		     PPSMC_MSG_GfxDriverResetRecovery,		0),
 	MSG_MAP(BoardPowerCalibration,		     PPSMC_MSG_BoardPowerCalibration,		0),
 	MSG_MAP(HeavySBR,                            PPSMC_MSG_HeavySBR,                        0),
+	MSG_MAP(SetBadHBMPagesRetiredFlagsPerChannel,	PPSMC_MSG_SetBadHBMPagesRetiredFlagsPerChannel,	0),
 };
 
 static const struct cmn2asic_mapping aldebaran_clk_map[SMU_CLK_COUNT] = {
@@ -1997,6 +2004,41 @@ static int aldebaran_smu_send_hbm_bad_page_num(struct smu_context *smu,
 	return ret;
 }
 
+static int aldebaran_check_bad_channel_info_support(struct smu_context *smu)
+{
+	uint32_t if_version = 0xff, smu_version = 0xff;
+	int ret = 0;
+
+	ret = smu_cmn_get_smc_version(smu, &if_version, &smu_version);
+	if (ret) {
+		/* return not support if failed get smu_version */
+		ret = -EOPNOTSUPP;
+	}
+
+	if (smu_version < SUPPORT_BAD_CHANNEL_INFO_MSG_VERSION)
+		ret = -EOPNOTSUPP;
+
+	return ret;
+}
+
+static int aldebaran_send_hbm_bad_channel_flag(struct smu_context *smu,
+		uint32_t size)
+{
+	int ret = 0;
+
+	ret = aldebaran_check_bad_channel_info_support(smu);
+	if (ret)
+		return ret;
+
+	/* message SMU to update the bad channel info on SMUBUS */
+	ret = smu_cmn_send_smc_msg_with_param(smu, SMU_MSG_SetBadHBMPagesRetiredFlagsPerChannel, size, NULL);
+	if (ret)
+		dev_err(smu->adev->dev, "[%s] failed to message SMU to update HBM bad channel info\n",
+				__func__);
+
+	return ret;
+}
+
 static const struct pptable_funcs aldebaran_ppt_funcs = {
 	/* init dpm */
 	.get_allowed_feature_mask = aldebaran_get_allowed_feature_mask,
@@ -2062,6 +2104,7 @@ static const struct pptable_funcs aldebaran_ppt_funcs = {
 	.i2c_fini = aldebaran_i2c_control_fini,
 	.send_hbm_bad_pages_num = aldebaran_smu_send_hbm_bad_page_num,
 	.get_ecc_info = aldebaran_get_ecc_info,
+	.send_hbm_bad_channel_flag = aldebaran_send_hbm_bad_channel_flag,
 };
 
 void aldebaran_set_ppt_funcs(struct smu_context *smu)