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

[SWDEV-511234] Added amdsmi_get_gpu_cper_entries & CLI implementation

Browse Source 
Added amdsmi_get_gpu_cper_entries() in the python and C APIs

Signed-off-by: Maisam Arif <Maisam.Arif@amd.com>
Signed-off-by: Oliveira, Daniel <daniel.oliveira@amd.com>
Co-authored-by: Saeed, Oosman <Oosman.Saeed@amd.com>
Co-authored-by: AL Musaffar, Yazen <Yazen.ALMusaffar@amd.com>
This commit is contained in:
Arif, Maisam
2025-04-12 01:54:57 -05:00
committed by GitHub
parent 3f75cd906f
commit d81871ef16
23 changed files with 1532 additions and 127 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/reference/amdsmi-py-api.md
+50
View File
@@ -1123,6 +1123,56 @@ except AmdSmiException as e:
print(e)
```
### amdsmi_get_gpu_cper_entries
Description: Dump CPER entries for a given GPU in a file using from CPER header file from RAS tool.
Input parameters:
* `processor_handle` device which to query
* `severity_mask` the severity mask of the entries to be retrieved
* `buffer_size` pointer to a variable that specifies the size of the cper_data
* `cursor` pointer to a variable that will contain the cursor for the next call
Output: Dictionary with fields
Field | Description
---|---
`error_severity` | The severity of the CPER error ex: `non_fatal_uncorrected`, `fatal`, `non_fatal_corrected`. |
`notify_type` | The notification type associated with the CPER entry. |
`timestamp` | The time when the CPER entry was recorded, formatted as `YYYY/MM/DD HH:MM:SS`. |
`signature` | A 4-byte signature identifying the entry, typically `CPER`. |
`revision` | The revision number of the CPER record format. |
`signature_end` | A marker value (typically `0xFFFFFFFF`) confirming the integrity of the signature. |
`sec_cnt` | The count of sections included in the CPER entry. |
`record_length` | The total length in bytes of the CPER entry. |
`platform_id` | A character array identifying the GPU or platform. |
`creator_id` | A character array indicating the creator of the CPER entry. |
`record_id` | A unique identifier for the CPER entry. |
`flags` | Reserved flags related to the CPER entry. |
`persistence_info` | Reserved information related to persistence. |
Exceptions that can be thrown by `amdsmi_get_gpu_cper_entries` function:
* `AmdSmiLibraryException`
* `AmdSmiParameterException`
Example:
```python
for device in devices:
entries, new_cursor = amdsmi_get_gpu_cper_entries(device, severity_mask, buffer_size, initial_cursor)
print("CPER entries for device", device)
for key, entry in entries.items():
print("Entry", key)
print(" Error Severity:", entry.get("error_severity", "Unknown"))
print(" Notify Type:", entry.get("notify_type", "Unknown"))
print(" Timestamp:", entry.get("timestamp", ""))
print()
print("New Cursor Position:", new_cursor)
except AmdSmiException as e:
print(e)
```
### amdsmi_get_gpu_board_info
Description: Returns board info for the given GPU