AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[SWDEV-230863] Improve the functionality of RdcSmiHealth module.

Browse Source 
Memory check:get the threshold of retired page number
EEPROM check:read and verify the checksum
Power/Thermal check: power/thermal throttle status counter

Signed-off-by: Meng Li <li.meng@amd.com>
Change-Id: Id2c751416eb5bf007e6e1da8dc05966a6ba1324e
This commit is contained in:
limeng12
2025-01-09 13:52:56 +08:00
committed by Meng, Li (Jassmine)
parent 83f36f1673
commit 016a1d9d39
11 changed files with 307 additions and 72 deletions
Download Patch File Download Diff File Expand all files Collapse all files
include/rdc_lib/impl/RdcWatchTableImpl.h
+7
View File
@@ -137,6 +137,7 @@ class RdcWatchTableImpl : public RdcWatchTable {
rdc_status_t get_start_end_values(rdc_gpu_group_t group_id,
uint32_t gpu_index,
rdc_field_t field,
uint64_t start_timestamp,
rdc_field_value *start_value,
rdc_field_value *end_value);
rdc_status_t pcie_check(rdc_gpu_group_t group_id,
@@ -145,6 +146,12 @@ class RdcWatchTableImpl : public RdcWatchTable {
uint32_t gpu_index, rdc_health_response_t* response);
rdc_status_t memory_check(rdc_gpu_group_t group_id,
uint32_t gpu_index, rdc_health_response_t* response);
rdc_status_t eeprom_check(rdc_gpu_group_t group_id,
uint32_t gpu_index, rdc_health_response_t* response);
rdc_status_t thermal_check(rdc_gpu_group_t group_id,
uint32_t gpu_index, rdc_health_response_t* response);
rdc_status_t power_check(rdc_gpu_group_t group_id,
uint32_t gpu_index, rdc_health_response_t* response);
RdcGroupSettingsPtr group_settings_;
RdcCacheManagerPtr cache_mgr_;