New Sysman API for VF telemetry #254
Conversation
| - if count is zero, the driver shall update the value with the total number of memory stats available. | ||
| - if count is greater than the total number of memory stats available, the driver shall update the value with the correct number of memory stats available. | ||
| - The count returned is the sum of number of VF instances currently available and the PF instance. | ||
| - type: $s_vf_util_mem_exp_t* |
There was a problem hiding this comment.
Mem type can be MEM_TYPE_SYSTEM and DEVICE right ?
So each VF can have MEM_TYPE_SYSTEM and DEVICE utilizations seperately ?
Please share whether we would be able to retrieve both information for each VF and the PF using this API.
There was a problem hiding this comment.
@aravindksg @sanilkumar0 this clarification still seems valid.
Could you see if there is a misunderstanding.
Thanks
There was a problem hiding this comment.
yes, it is intended that from API PoV we have capability to expose system and device mem utilizations for each VF from the PF.
There was a problem hiding this comment.
Since the count is"count returned is the sum of number of VF instances currently available and the PF instance", please share how this API would return both in a single call ?
There was a problem hiding this comment.
@sanilkumar0 I think Santosh's point is valid, in these APIs, the input is a VF handle so we return the utilization stats wrt that VF. So for PF the expectation might be use existing zesMemoryGetState call for gathering utilization info. @sanilkumar0 could you please verify that.
We might need to clarify the description around these APIs if that's the case.
| name: "flag" | ||
| desc: "[in] Use bitf field of utilization flags to set sampling interval." | ||
| - type: uint64_t | ||
| name: "samplingInterval" |
There was a problem hiding this comment.
We could mention units as well
| name: "activeCounterValue" | ||
| desc: "[out] Represents active counter." | ||
| - type: uint64_t | ||
| name: "samplingCounterValue" |
There was a problem hiding this comment.
since "timestamp" already has the wall clock time, please share if we need the "samplingCounterValue" ?
wdamon-intel
force-pushed
the
VirtualFunctionManagement
branch
from
bc137e1 to
8430c9c
Compare
|
Moving forward with merge. As this is an experimental extension, any necessary adjustments may be done in 1.9.x. |
Resolves: #248