Add ProcessorMetrics endpoint for RAS error counts

[raja-gade]

Add a new Redfish OEM endpoint at
/redfish/v1/Systems/{}/Processors/{}/Oem/AMD/ProcessorMetrics that exposes per-socket RAS error counters from the com.amd.RAS D-Bus service.

The endpoint returns four error categories per processor socket:

• CorrectableCPUErrors (array of uint64)
• CorrectableOtherErrors (uint64)
• NoncorrectableCPUErrors (array of uint64)
• NoncorrectableOtherErrors (uint64)

Tested:
Verified GET on the ProcessorMetrics endpoint returns correct error counts from the com.amd.RAS ErrorCount D-Bus object.

'''
curl -s -k -u root:0penBmc -H "Content-type: application/json" -X GET https:///redfish/v1/Systems/system/Processors/P0/Oem/AMD/ProcessorMetrics {
"@odata.id": "/redfish/v1/Systems/system/Processors/P0/Oem/AMD/ProcessorMetrics",
"@odata.type": "#AmdProcessorMetrics.AmdProcessorMetrics",
"CorrectableCPUErrors": [
0,
0,
....
0
],
"CorrectableOtherErrors": 0,
"Id": "ProcessorMetrics",
"Name": "P0 Processor Error Metrics",
"NoncorrectableCPUErrors": [
2,
0,
...
0
],
"NoncorrectableOtherErrors": 0
}
'''

[supven01]

@raja-gade - This is good. However, this was not required for internal reference BMC. G would have done it themselves.

[raja-gade]

@raja-gade - This is good. However, this was not required for internal reference BMC. G would have done it themselves.

Thanks you @supven01
Just to confirm — should I proceed with merging this change,
or should we hold it since it may not be required for the internal reference BMC?
If it’s not needed, I can abandon