| Bug ID | 1577 |
|---|---|
| Summary | dts: Request for memory usage test in CI |
| Product | DPDK |
| Version | unspecified |
| Hardware | All |
| OS | All |
| Status | UNCONFIRMED |
| Severity | normal |
| Priority | Normal |
| Component | DTS |
| Assignee | dev@dpdk.org |
| Reporter | probb@iol.unh.edu |
| CC | juraj.linkes@pantheon.tech, probb@iol.unh.edu |
| Target Milestone | --- |
On 11/13/24 Tech board discussed the idea of tracking hugepage memory usage for EAL bringup and basic DPDK workloads for per patch CI testing. Ideally UNH could track a baseline for each DPDK tree, and then for per patch testing compare the baseline for the tree a patch is applied on with the memory usage test result for that patch. Testcases would be: 1. Start testpmd (EAL starts) and check memory usage. 2. Start some basic l2/l3 forwarding and check memory usage during. 3. (possible) Create flow rules for a device, check memory usage. 4. etc. Prereqs to adding this test: 1. Add a runtime command to testpmd which will return hugepage memory usage. Other notes: 1. UNH would need to track tree memory usage for each device/driver, as we can expect usage will not be consistent across devices. In principle it is similar to how UNH stores performance baselines for NIC/tree combinations right now, so no DB schema updates should be required.