extension/llm/server: warm append-only session resume (V2b.1)

[mergennachin]

Builds on the isolated named sessions: a named session now keeps its decoded
context across requests. When the next request's prompt tokens are an exact
prefix extension of the session's resident tokens (the same conversation plus a
new turn), the worker prefills ONLY the new suffix -- continuing the
KV/recurrent state in place -- instead of resetting and re-prefilling the whole
prompt. The match is exact-token (never re-tokenized text), so it is always
correct: a token mismatch, a stop-string trim, or a prior error falls back to a
full reset + prefill. This is per-session resume, not global prefix caching.

The decision lives in a dependency-free pure helper (worker_prefill_plan.h,
unit-tested standalone); worker_loop.h tracks each session's resident token ids
(invariant: resident size == session position) and executes the plan, and the
done event reports reused_prompt_tokens / prefilled_prompt_tokens /
session_reset_reason for measuring the hit rate. POST /v1/sessions/{id}/reset
clears a session's context while keeping its slot. The qwen worker's
--warm_resume (serve.py --no-warm-resume) gates the behavior for A/B measurement.

Review order: worker_prefill_plan.h + its test; then worker_loop.h (resident
tracking + plan execution); then the control-plane reset op + metrics; then docs.

Part of #20001

[mergennachin]

Stack from ghstack (oldest at bottom):

• extension/llm/server: token-ID prompt segments for tool-use resume (V2b.1.5) #20161
• -> extension/llm/server: warm append-only session resume (V2b.1) #20160
• extension/llm/server: isolated multi-session serving (V2a) #20159
• Qwen3.5-MoE CUDA V2 foundation: one model, many isolated sessions #20117
• examples/models/qwen3_5_moe: CUDA Engine/Session adapter + OpenAI serving #20043
• extension/llm/server: document pi integration #19999
• extension/llm/server: worker-based OpenAI-compatible HTTP server #19994
• extension/llm/server: serving foundations (schemas, errors, templating, tools) #19993
• extension/llm/runner: Engine/Session C++ core + token-step primitives #19991

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/20160

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ 6 New Failures, 3 Pending

As of commit b930b9f with merge base eeb0646 ():

NEW FAILURES - The following jobs have failed:

• pull / test-llama-runner-qnn-linux (fp32, qnn_16a16w, qnn) / linux-job (gh)
  RuntimeError: Command docker exec -t 0f6f17a0f3974026c429abbbeed1bad55780f2605fd25428bd1ca3cf9012482e /exec failed with exit code 92
• pull / unittest / linux / linux-job (gh)
  RuntimeError: Command docker exec -t 0c86ab577016019fafc1aeda5f1a3d39579fe39ffe62fc5d2d936c4bd17be9f5 /exec failed with exit code 1
• pull / unittest / macos / macos-job (gh)
  RuntimeError: Command bash /Users/ec2-user/runner/_work/_temp/exec_script failed with exit code 1
• pull / unittest-editable / linux / linux-job (gh)
  RuntimeError: Command docker exec -t 069bfa93fb96275cb56393b9a6ebc0b635bd51bd53c663894cf07e852fad4062 /exec failed with exit code 1
• pull / unittest-editable / macos / macos-job (gh)
  RuntimeError: Command bash /Users/ec2-user/runner/_work/_temp/exec_script failed with exit code 1
• pull / unittest-nxp-neutron / linux-job (gh)
  RuntimeError: Command docker exec -t 35221cd6b022a631bfc59a5a81c442a980f1c3fdfef403ffc5cb3cd6dfc8c953 /exec failed with exit code 1

This comment was automatically generated by Dr. CI and updates every 15 minutes.