Expose CP shared-KV allocation fallback as a warning

The compute-owner allocation path is an intended hot path for CP shared KV. Falling back to legacy page allocation changes the behavior profile enough that INFO/no-marker logging is too easy to miss during profiling.

This makes the fallback warning-visible and gives it the standard CP shared-KV fallback marker while preserving the existing behavior of logging every fallback event.

Constraint: Production fallbacks from intended CP shared-KV hot paths must be visible and grep-able.

Rejected: Keep INFO logging | historical silent or low-visibility fallbacks hid inactive optimized paths.

Confidence: medium

Scope-risk: narrow

Directive: Keep fallback logs warning-level unless the path is proven to be expected steady state and separately observable.

Tested: Local py_compile for common.py and test_cp_shared_kv_layout.py.

Tested: Local extracted assertLogs check for _log_cp_shared_kv_alloc_fallback warning prefix and reason formatting.

Tested: Remote g0034 py_compile for common.py and test_cp_shared_kv_layout.py.

Not-tested: Full remote pytest for test_cp_shared_kv_layout.py; current g0034 container aborts while importing installed sgl_kernel common_ops before test execution.

Co-authored-by: OmX <omx@oh-my-codex.dev>
This commit is contained in:
laoyao0822
2026-05-29 05:59:02 +08:00
parent 38532a1fc9
commit e09c8256d7
3 changed files with 18 additions and 3 deletions

View File

@@ -782,6 +782,10 @@ Current state:
- Some prefetch misses use debug logs or normal fallback paths.
- Historical silent fallbacks made it hard to identify that async/per-layer or
page-aligned paths were not actually active.
- The CP shared-KV compute-owner allocation fallback currently logs at INFO and
does not use the `[CP_SHARED_KV_FALLBACK]` marker. This is an intended hot
path fallback from owner-lane/page-aligned allocation to legacy allocation, so
it must be visible at warning level.
Correction:
@@ -798,6 +802,15 @@ Tests:
- MLA prefix not page-aligned;
- index prefetch consume miss after start layer;
- CP HiCache reservation/load capacity failure.
- Update compute-owner allocation fallback coverage to require warning level and
the `[CP_SHARED_KV_FALLBACK][compute_owner_alloc]` prefix.
Implemented C11 slice:
- Changed CP shared-KV compute-owner allocation fallback logging from INFO to
WARNING and added the `[CP_SHARED_KV_FALLBACK][compute_owner_alloc]` marker.
- Tightened the existing unit expectation so repeated fallback events are still
emitted, but now require the production-visible fallback prefix.
### C12. Verification gaps before claiming completion