Add the RdcSmiDiagnostic module

Provides a RdcSmiDiagnostic module, which will call rocm_smi_lib.

It will support following diagnostics: Get GPU Topology, Check GPU
parameters and check processes running on the GPUs.

The grpc client and server side diagnostics function is added.

The diag module is added to the rdci.

Change-Id: I10a0cf3c20556a61373ab686f82cae75acaa40dd
このコミットが含まれているのは:
Bill(Shuzhou) Liu committed 2021-07-26 14:56:17 -04:00
1 親 5a4bf97327
コミット 76ccf58008
22ファイル変更
+1330 -13

検索条件に一致するファイルはありません

+22 -1
ファイルの表示
@@ -67,7 +67,7 @@ namespace amd {
namespace rdc {
// TODO(bill_liu): make it configurable
const uint32_t METIC_UPDATE_FREQUENCY = 100; // 100ms by default
const uint32_t METIC_UPDATE_FREQUENCY = 1000; // 1000 microseconds by default
RdcEmbeddedHandler::RdcEmbeddedHandler(rdc_operation_mode_t mode):
group_settings_(new RdcGroupSettingsImpl())
@@ -386,6 +386,9 @@ rdc_status_t RdcEmbeddedHandler::rdc_diagnostic_run(
rdc_gpu_group_t group_id,
rdc_diag_level_t level,
rdc_diag_response_t* response) {
if (!response) {
return RDC_ST_BAD_PARAMETER;
}
// Get GPU group information
rdc_group_info_t rdc_group_info;
@@ -397,6 +400,24 @@ rdc_status_t RdcEmbeddedHandler::rdc_diagnostic_run(
return diag->rdc_diagnostic_run(rdc_group_info, level, response);
}
rdc_status_t RdcEmbeddedHandler::rdc_test_case_run(
rdc_gpu_group_t group_id,
rdc_diag_test_cases_t test_case,
rdc_diag_test_result_t* result) {
if (!result) {
return RDC_ST_BAD_PARAMETER;
}
// Get GPU group information
rdc_group_info_t rdc_group_info;
rdc_status_t status = rdc_group_gpu_get_info(
group_id, &rdc_group_info);
if (status != RDC_ST_OK) return status;
auto diag = rdc_module_mgr_->get_diagnostic_module();
return diag->rdc_test_case_run(test_case, rdc_group_info.entity_ids,
rdc_group_info.count, result);
}
// Control API
rdc_status_t RdcEmbeddedHandler::rdc_field_update_all(
uint32_t wait_for_update) {