drm/amd/pm: Add support to query partition metrics

Add interfaces to query compute partition related metrics data.

Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Asad Kamal <asad.kamal@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/include/kgd_pp_interface.h b/drivers/gpu/drm/amd/include/kgd_pp_interface.h
index 0f7542d..f4d914d 100644 (file)
--- a/drivers/gpu/drm/amd/include/kgd_pp_interface.h
+++ b/drivers/gpu/drm/amd/include/kgd_pp_interface.h
@@ -494,6 +494,7 @@ struct amd_pm_funcs {
        int (*set_df_cstate)(void *handle, enum pp_df_cstate state);
        int (*set_xgmi_pstate)(void *handle, uint32_t pstate);
        ssize_t (*get_gpu_metrics)(void *handle, void **table);
+       ssize_t (*get_xcp_metrics)(void *handle, int xcp_id, void *table);
        ssize_t (*get_pm_metrics)(void *handle, void *pmmetrics, size_t size);
        int (*set_watermarks_for_clock_ranges)(void *handle,
                                               struct pp_smu_wm_range_sets *ranges);
@@ -1592,4 +1593,27 @@ struct amdgpu_pm_metrics {
        uint8_t data[];
 };
 
+struct amdgpu_partition_metrics_v1_0 {
+       struct metrics_table_header common_header;
+       /* Current clocks (Mhz) */
+       uint16_t current_gfxclk[MAX_XCC];
+       uint16_t current_socclk[MAX_CLKS];
+       uint16_t current_vclk0[MAX_CLKS];
+       uint16_t current_dclk0[MAX_CLKS];
+       uint16_t current_uclk;
+       uint16_t padding;
+
+       /* Utilization Instantaneous (%) */
+       uint32_t gfx_busy_inst[MAX_XCC];
+       uint16_t jpeg_busy[NUM_JPEG_ENG_V1];
+       uint16_t vcn_busy[NUM_VCN];
+       /* Utilization Accumulated (%) */
+       uint64_t gfx_busy_acc[MAX_XCC];
+       /* Total App Clock Counter Accumulated */
+       uint64_t gfx_below_host_limit_ppt_acc[MAX_XCC];
+       uint64_t gfx_below_host_limit_thm_acc[MAX_XCC];
+       uint64_t gfx_low_utilization_acc[MAX_XCC];
+       uint64_t gfx_below_host_limit_total_acc[MAX_XCC];
+};
+
 #endif

diff --git a/drivers/gpu/drm/amd/pm/amdgpu_dpm.c b/drivers/gpu/drm/amd/pm/amdgpu_dpm.c
index 2148c8d..d98c95d 100644 (file)
--- a/drivers/gpu/drm/amd/pm/amdgpu_dpm.c
+++ b/drivers/gpu/drm/amd/pm/amdgpu_dpm.c
@@ -2019,3 +2019,35 @@ int amdgpu_dpm_get_dpm_clock_table(struct amdgpu_device *adev,
 
        return ret;
 }
+
+/**
+ * amdgpu_dpm_get_xcp_metrics - Retrieve metrics for a specific compute
+ * partition
+ * @adev: Pointer to the device.
+ * @xcp_id: Identifier of the XCP for which metrics are to be retrieved.
+ * @table: Pointer to a buffer where the metrics will be stored. If NULL, the
+ * function returns the size of the metrics structure.
+ *
+ * This function retrieves metrics for a specific XCP, including details such as
+ * VCN/JPEG activity, clock frequencies, and other performance metrics. If the
+ * table parameter is NULL, the function returns the size of the metrics
+ * structure without populating it.
+ *
+ * Return: Size of the metrics structure on success, or a negative error code on failure.
+ */
+ssize_t amdgpu_dpm_get_xcp_metrics(struct amdgpu_device *adev, int xcp_id,
+                                  void *table)
+{
+       const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs;
+       int ret = 0;
+
+       if (!pp_funcs->get_xcp_metrics)
+               return 0;
+
+       mutex_lock(&adev->pm.mutex);
+       ret = pp_funcs->get_xcp_metrics(adev->powerplay.pp_handle, xcp_id,
+                                       table);
+       mutex_unlock(&adev->pm.mutex);
+
+       return ret;
+}

diff --git a/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h b/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h
index 2c3c975..c0f9ecb 100644 (file)
--- a/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h
+++ b/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h
@@ -524,6 +524,8 @@ int amdgpu_dpm_get_power_profile_mode(struct amdgpu_device *adev,
 int amdgpu_dpm_set_power_profile_mode(struct amdgpu_device *adev,
                                      long *input, uint32_t size);
 int amdgpu_dpm_get_gpu_metrics(struct amdgpu_device *adev, void **table);
+ssize_t amdgpu_dpm_get_xcp_metrics(struct amdgpu_device *adev, int xcp_id,
+                                  void *table);
 
 /**
  * @get_pm_metrics: Get one snapshot of power management metrics from PMFW. The

diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
index f24a1d8..d79a1d9 100644 (file)
--- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
+++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c
@@ -3758,6 +3758,19 @@ int smu_set_pm_policy(struct smu_context *smu, enum pp_pm_policy p_type,
        return ret;
 }
 
+static ssize_t smu_sys_get_xcp_metrics(void *handle, int xcp_id, void *table)
+{
+       struct smu_context *smu = handle;
+
+       if (!smu->pm_enabled || !smu->adev->pm.dpm_enabled)
+               return -EOPNOTSUPP;
+
+       if (!smu->adev->xcp_mgr || !smu->ppt_funcs->get_xcp_metrics)
+               return -EOPNOTSUPP;
+
+       return smu->ppt_funcs->get_xcp_metrics(smu, xcp_id, table);
+}
+
 static const struct amd_pm_funcs swsmu_pm_funcs = {
        /* export for sysfs */
        .set_fan_control_mode    = smu_set_fan_control_mode,
@@ -3816,6 +3829,7 @@ static const struct amd_pm_funcs swsmu_pm_funcs = {
        .get_uclk_dpm_states              = smu_get_uclk_dpm_states,
        .get_dpm_clock_table              = smu_get_dpm_clock_table,
        .get_smu_prv_buf_details = smu_get_prv_buffer_details,
+       .get_xcp_metrics                  = smu_sys_get_xcp_metrics,
 };
 
 int smu_wait_for_event(struct smu_context *smu, enum smu_event_type event,

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 d47e32a..9aacc7b 100644 (file)
--- a/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h
+++ b/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h
@@ -1466,6 +1466,12 @@ struct pptable_funcs {
         */
        int (*set_wbrf_exclusion_ranges)(struct smu_context *smu,
                                        struct freq_band_range *exclusion_ranges);
+       /**
+        * @get_xcp_metrics: Get a copy of the partition metrics table from SMU.
+        * Return: Size of table
+        */
+       ssize_t (*get_xcp_metrics)(struct smu_context *smu, int xcp_id,
+                                  void *table);
 };
 
 typedef enum {

diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h
index ade36a8..7473672 100644 (file)
--- a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h
+++ b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.h
@@ -52,6 +52,18 @@
                header->structure_size = sizeof(*(ptr));       \
        } while (0)
 
+#define smu_cmn_init_partition_metrics(ptr, frev, crev)                     \
+       do {                                                                \
+               typecheck(struct amdgpu_partition_metrics_v##frev##_##crev, \
+                         typeof(*(ptr)));                                  \
+               struct metrics_table_header *header =                       \
+                       (struct metrics_table_header *)(ptr);               \
+               memset(header, 0xFF, sizeof(*(ptr)));                       \
+               header->format_revision = frev;                             \
+               header->content_revision = crev;                            \
+               header->structure_size = sizeof(*(ptr));                    \
+       } while (0)
+
 extern const int link_speed[];
 
 /* Helper to Convert from PCIE Gen 1/2/3/4/5/6 to 0.1 GT/s speed units */