test_get_named_events: Don't try and maintain the set of all possible perf stat keys here.

[wash-amzn]

When I ran the tests on a codebuild github runner (I don't actually know the instance type because they don't let you choose), this test was failing. I modified it to dump the event keys, and the failure was because data-st-tlb-mpki and data-st-tlb-tw-pki were being returned, but hadn't been added to this test's allow-list.

The values that the test failure saw

Event values: ["ipc", "stall-frontend-pkc", "branch-mpki", "inst-tlb-mpki", "inst-tlb-tw-pki", "inst-l1-mpki", "stall-backend-pkc", "l3-mpki", "l2-mpki", "data-tlb-mpki", "data-tlb-tw-pki", "data-l1-mpki", "data-st-tlb-mpki", "data-st-tlb-tw-pki", "ipc", "stall-frontend-pkc", "branch-mpki", "inst-tlb-mpki", "inst-tlb-tw-pki", "inst-l1-mpki", "stall-backend-pkc", "l3-mpki", "l2-mpki", "data-tlb-mpki", "data-tlb-tw-pki", "data-l1-mpki", "data-st-tlb-mpki", "data-st-tlb-tw-pki"]

and failed with

thread 'data::perf_stat::tests::test_get_named_events' panicked at src/data/perf_stat.rs:620:21:
assertion failed: key_map.contains_key(&event.to_string())

Instead of trying to maintain a superset of all possible keys in the test, now it just asserts for a simple case (ipc) and doesn't try to defend against 'unexpected' ones.

This now passes on those codebuild github runners.

test data::perf_stat::tests::test_get_named_events ... ok