Skip to content

feat(kokoro): own opt-in CUDA worker admission and cleanup - #55

Merged
sushiHex merged 4 commits into
mainfrom
feat/kokoro-owned-worker-20261106
Nov 6, 2026
Merged

sushiHex merged 4 commits into
mainfrom
feat/kokoro-owned-worker-20261106

Conversation

@sushiHex

@sushiHex sushiHex commented Nov 6, 2026 •

Copy link
Copy Markdown
Owner

Kokoro CUDA could only be launched directly by its caller, so vram-mcp could reserve capacity without owning the load. Add three opt-in, Windows-only tools for one operator-configured reviewed worker: load_kokoro, kokoro_status, and stop_kokoro.

The server reserves and serializes admission, checks fresh selected-device capacity, launches the actual verified base Python image, and retains custody until child exit and reservation/admission release are proven. Callers supply no command or paths. The existing pinned worker starts its Windows parent-handle watchdog before CUDA initialization. Device visibility is restricted to the admitted UUID. Unknown launch or cleanup blocks duplicate work; an unpublished admission failure explicitly requires closing the scoped server. There is no force, eviction, direct-launch fallback or default enablement.

A successful load means launch acceptance only: gpu_ready remains false. The trusted client must independently validate the existing authenticated CUDA-ready/profile handshake. Operator capacity estimates are not hard memory caps or attributable peaks. This does not reload existing MCP servers or alter the CPU voice baseline.

Validation:

  • Final full suite: 794 passed, one skipped. Final focused loader/schema checks: 77 passed.
  • 34 isolated consequential guard mutants fail their targeted cases; sources restored. Independent capacity/device/privacy and process/lifecycle reviews are GO.
  • Actual model-free Windows checks verified direct-base/venv identity, redirector refusal, actual-child exit before capacity release, and reviewed watchdog exit on actual parent death.
  • Actual SDK malformed-input tests caught and repaired token-prefix reflection before the handler; new calls return fixed refusals without argument echoes.

One bounded integration trial used a separate real MCP server and the new load/status/stop tools. Launch was accepted and shared-ledger ownership was verified, but the authenticated CUDA-ready handshake was not reached, so no speech timings were measured. The owned worker stopped, its reservation was released, and the scoped server closed. Sampled pressure remained healthy; attributable peak memory and desktop usability were not measured. The probe retained only a generic startup failure, so its cause remains unproven. No retry or production activation is queued; the CPU voice baseline is unchanged. CUDA readiness and performance remain open under Hermes #445.

Refs Hermes optional-loader dependency. That record retains host-integration and hardware qualification ownership; this PR does not close it automatically.

@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Nov 6, 2026 •

Copy link
Copy Markdown
Codex Review Summary: ✅ Completed

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-11-06T22:02:04.444469Z 8609dcc PR opened
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@sushiHex
sushiHex merged commit 8a1bae0 into main Nov 6, 2026
15 checks passed
@sushiHex

sushiHex commented Nov 6, 2026

Copy link
Copy Markdown
Owner Author

Merged as 8a1bae0. Verified that its tree exactly matches reviewed head 8609dcc and its sole parent is the captured previous main 5d07c37. Both PR workflows passed on attempt 1. Post-merge main CI/build are running; this does not activate the loader in existing servers or qualify CUDA speech readiness.

The bounded real MCP trial accepted the launch and cleaned up the owned worker/reservation, but failed before authenticated CUDA readiness. No speed result exists. Hermes #445 retains readiness and host-integration work; no additional GPU trial is queued.

@sushiHex

sushiHex commented Nov 6, 2026

Copy link
Copy Markdown
Owner Author

Post-merge validation is complete: main CI run 42302851706 and package build run 42302851797 both passed on attempt 1 for 8a1bae0. Release attachment and PyPI publication were correctly skipped. No loader activation or further GPU trial occurred.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant