完成L4-USER-M-PROCESS Web仿真接入

结论：L4-USER-M-PROCESS 已按 Web/virtual HAL 数控仿真主线完成接入，native LinuxCNC runtime 不再作为 Web 仿真阻塞；新增 text31.txt 接续剩余 77 个 SKIP 的 main-program promotion 复核。

wasm-port/docs/compatibility-validation.md

@@ -209,6 +209,7 @@ wasm-port/build/wasm/sim-configs-inventory/native-proof-alignment-summary.tsv
 wasm-port/build/wasm/sim-configs-inventory/runtime-boundary-native-alignment-summary.tsv
 wasm-port/build/wasm/sim-configs-inventory/native-runtime-probe-execution-plan.tsv
 wasm-port/build/wasm/sim-configs-inventory/promotion-candidates.tsv
+wasm-port/build/wasm/sim-configs-inventory/remaining-skip-main-program-promotion-audit.tsv
 wasm-port/build/wasm/sim-configs-inventory/blocked-runtime-promotion-lock.tsv
 wasm-port/build/wasm/sim-configs-inventory/next-boundary-worklist.tsv
 wasm-port/build/wasm/sim-configs-inventory/next-boundary-recommendations.tsv
@@ -240,6 +241,15 @@ ready while the Node inventory baseline remains unchanged, and
 `inventory-ready` skipped main-program rows that could affect the Node
 inventory baseline. The current direct inventory `promotion_allowed=1` count
 is zero: remaining skipped main rows are still hard-blocked by `L4-USER-M-PROCESS` or preserved as `UPSTREAM-DEMO`.
+`remaining-skip-main-program-promotion-audit.tsv` is the focused row-level
+audit for those skipped main-program rows. It currently records two rows,
+`axis/vismach/millturn/example.ngc` and
+`axis/vismach/5axis/table-rotary_spindle-rotary-nutating/demos/incremental_repetition_g533.ngc`,
+and keeps both `promotion_allowed=0`; the millturn row now has a browser
+virtual HAL state-transition proof for `M429 -> M129` and `M428 -> M128`, but
+still lacks the LinuxCNC-owned native runtime pass and promotion-gate proof
+required to leave `L4-USER-M-PROCESS`. The upstream demo row remains
+non-promotable upstream demo evidence.
 The project release readiness artifact and URL workflow publish the same
 candidate artifact summary for callers: `evidence-ready=8`,
 `inventory-ready=2`, total candidates `10`, and `promotion_allowed=0`.
@@ -404,10 +414,10 @@ coverage: every TSV emitted under
 `full-process-boundary-design.md`. New gate artifacts must therefore be
 documented before the inventory can pass.
 The Node coverage gate also requires the generated WASM inventory artifact list
-to remain the exact duplicate-free 59-entry baseline and the native generated
+to remain the exact duplicate-free 60-entry baseline and the native generated
 TSV token list to remain the exact duplicate-free 8-token baseline. It also
 requires the corresponding `boundary-phase-completion-summary.tsv`
-documentation-coverage counts to remain `57` and `8`, respectively. It also
+documentation-coverage counts to remain `60` and `8`, respectively. It also
 checks that the browser smoke source lists the same generated WASM and native
 artifact tokens and preserves the same fixed-count, duplicate-free,
 fetchability/header, documentation-missing, and completion count-parity guards

wasm-port/docs/full-process-boundary-design.md

@@ -157,11 +157,28 @@ enabled with `ENABLE_MILLTURN_USER_M_RUNTIME_PROBE=1`, it starts the vendored
 `millturn.ini`, runs the LinuxCNC-owned Tcl `M128`/`M129` scripts, and verifies
 the resulting `ini.[xyz].*` HAL state targets before reporting
 `runtime_state_probe_passed`.
+If another LinuxCNC runtime is already active, the probe reports
+`millturn_user_m_runtime_probe_status=blocked_existing_linuxcnc_runtime` and
+`millturn_user_m_missing_requirements=exclusive_linuxcnc_runtime` instead of
+terminating or taking over that process graph. The current workstation has
+shown that state with an existing `linuxcncsvr -ini .../axis.ini` and
+`rtapi_app load tpmod`.
 
 The artifact keeps `execution_enabled=0` and `promotion_allowed=0`. It is not
 a standalone Tcl/HAL executor and does not make `millturn` a Node/browser
 representative.
 
+The Web/browser simulation side now has a separate source-derived virtual HAL
+proof for the same millturn state transitions. `runRealBrowserSimulation()`
+exposes `millturnUserMProcess` with `M429 -> M129` turn and `M428 -> M128`
+mill transitions, switchkins guard pins, and `ini.[xz].*` limit pin values.
+Node smoke and the real browser simulation page smoke assert that proof while
+keeping `webSimulationReady=true`, `nativeRuntimeRequired=false`,
+`processExecutionReady=false`, `executionEnabled=false`, and
+`promotionAllowed=false`. This completes the browser simulation proof without a
+native LinuxCNC runtime dependency. It still does not execute arbitrary external
+user-M processes or unlock inventory promotion.
+
 `linuxcnc_millturn_user_m_boundary_probe` is the current native guard for this
 evidence. It reads the LinuxCNC source `millturn.ini`, `mcodes/M128`,
 `mcodes/M129`, and the `428remap`/`429remap` callers, then verifies the

wasm-port/docs/host-runtime-boundary-handoff.md

@@ -47,12 +47,26 @@ execution_enabled=0
 promotion_allowed=0
 ```
 
+Manual `L4-USER-M-PROCESS` opt-in status on the current workstation:
+
+```text
+ENABLE_MILLTURN_USER_M_RUNTIME_PROBE=1 bash wasm-port/tests/native/probe_millturn_user_m_runtime.sh
+millturn_user_m_runtime_probe_status=blocked_existing_linuxcnc_runtime
+millturn_user_m_missing_requirements=exclusive_linuxcnc_runtime
+```
+
+The conflicting runtime was an already-running `linuxcncsvr -ini .../axis.ini`
+plus `rtapi_app load tpmod`. The probe intentionally keeps the millturn row
+blocked instead of taking over that runtime. The browser/virtual HAL proof for
+`M429 -> M129` and `M428 -> M128` is complete in the simulation page smoke, but
+it is not a native runtime pass and does not unlock inventory promotion.
+
 ## Promotion blockers
 
 Each blocked family currently has `promotion_ready=0`, `execution_enabled=0`,
 and `promotion_allowed=0`.
 
-For `L4-USER-M-PROCESS` and `L4-TOOL-DB`, the current blocker keys are:
+For `L4-USER-M-PROCESS` and `L4-TOOL-DB`, the current generated blocker keys are:
 
 ```text
 native_runtime_probe_not_passed
@@ -87,6 +101,9 @@ For rows that still lack native evidence, the next unblock action is to provide
 the missing host runtime requirements for the relevant family. Providing host
 commands is not enough to promote a row; it only allows the next opt-in probe
 gate to be considered.
+For `L4-USER-M-PROCESS` on this workstation, the remaining native requirement
+is stricter than command availability: the millturn probe needs an exclusive
+LinuxCNC runtime with no existing `linuxcncsvr`/`rtapi_app` process graph.
 
 ## Opt-in probe dispatch
 

wasm-port/docs/linuxcnc-test-porting-tracker.md

@@ -199,7 +199,7 @@ Use this checklist when adding each LinuxCNC upstream test:
 | `tests/uspace/spawnv-root` | Blocked | Sudo-installed userspace HAL component and `spawnv` root behavior. | Keep upstream-only unless HAL component install/runtime support is added. |
 | `configs/sim/*/remap_subs/*.ngc` browser main-program inventory | Blocked | These files are remap/subroutine assets, not standalone browser main-program targets. | Keep validating them through native remap parse/execute paths and explicit five-axis/runtime probes instead of widening browser `runSimConfigProgram()` blindly. |
 | `configs/sim/axis/db_demo/base.ngc` Layer 4 inventory | Blocked | `[EMCIO]DB_PROGRAM = ./db_nonran.py` requires LinuxCNC task/tooldata DB process startup, command/reply protocol, and database state behavior, not standalone interpreter file execution. | Keep as `L4-TOOL-DB` until a LinuxCNC-owned tool database runtime boundary exists and proves lookup/update behavior across native and WASM. `blocked-dependency-summary.tsv` records `taskclass.cc`, `tooldata_db.cc`, `tooldata_common.cc`, and `configs/sim/axis/db_demo/db.py` as the owner set plus source-derived `tool_db_protocol_evidence` for the `v2.1` handshake, `g`/`FINI` get-all, and `l`/`u`/`p` notification protocol. |
-| `configs/sim/axis/vismach/millturn/example.ngc` Layer 4 inventory | Blocked | The remap execution chain calls external `USER_M_PATH` process codes `M128` and `M129`, with HAL/HALUI/UI process declarations in `millturn.ini`. | Keep as `L4-USER-M-PROCESS`; do not promote until the M128/M129 LinuxCNC-owned state boundary proves the kinstype guard and `ini.[xyz].*` HAL pin updates. `blocked-dependency-summary.tsv` records the `mcodes/M128` and `mcodes/M129` Tcl scripts plus source-derived `user_m_process_effects` for Tcl/HAL runtime use, kinstype guards, and `ini.[xyz]` HAL pin side effects. |
+| `configs/sim/axis/vismach/millturn/example.ngc` Layer 4 inventory | Blocked | The remap execution chain calls external `USER_M_PATH` process codes `M128` and `M129`, with HAL/HALUI/UI process declarations in `millturn.ini`. | Keep as `L4-USER-M-PROCESS`; the Web/virtual HAL simulation proof now covers `M429 -> M129` turn and `M428 -> M128` mill state transitions, including switchkins guard pins and `ini.[xz].*` limit pins in Node and real browser page smoke. Do not promote until the opt-in native LinuxCNC runtime probe can run with exclusive `linuxcncsvr`/`rtapi_app` ownership and report `runtime_state_probe_passed`; the current workstation reports `blocked_existing_linuxcnc_runtime` / `exclusive_linuxcnc_runtime`. `blocked-dependency-summary.tsv` records the `mcodes/M128` and `mcodes/M129` Tcl scripts plus source-derived `user_m_process_effects` for Tcl/HAL runtime use, kinstype guards, and `ini.[xyz]` HAL pin side effects. |
 | `configs/sim/gmoccapy/*` Layer 4 full inventory | Blocked | Python remap runtime boundary is present in native LinuxCNC `rs274`, but not yet intentionally exposed for Node/browser full inventory. | Keep native inventory coverage as source-of-truth; add one representative Node/browser class sample only after the Python remap boundary is deliberately designed. Python module and prolog/epilog ownership is inventoried in `blocked-dependency-summary.tsv`; LinuxCNC runtime ownership is recorded as `python_runtime_evidence` in the Python boundary/family summaries. |
 | `configs/sim/axis/laser/*` Layer 4 full inventory | Blocked | Python remap runtime boundary is present in native LinuxCNC `rs274`, but not yet intentionally exposed for Node/browser full inventory. | Keep native inventory coverage as source-of-truth; add one representative Node/browser class sample only after the Python remap boundary is deliberately designed. Python module and remap ownership is inventoried in `blocked-dependency-summary.tsv`; LinuxCNC runtime ownership is recorded as `python_runtime_evidence` in the Python boundary/family summaries. |
 | `configs/sim/axis/remap/*/nc_files/*.ngc` Layer 4 full inventory | Blocked | Native LinuxCNC `rs274` coverage exists, but these demo main programs depend on Python remaps, Python prolog/epilog callbacks, or Python-backed queue/tool-change handlers. | Keep native inventory coverage as source-of-truth; do not reduce Python remap demos to missing vendored files or browser standalone execution until the Python remap boundary is deliberately designed. Dependency ownership is inventoried in `blocked-dependency-summary.tsv`; LinuxCNC runtime ownership is recorded as `python_runtime_evidence` in the Python boundary/family summaries. |

wasm-port/docs/sim-configs-completion-plan.md

@@ -460,12 +460,19 @@ Boundary decisions for the current hard blocks:
   to change INI axis limit HAL pins after the kinematics switch. The current
   LinuxCNC owner is therefore task user-M process dispatch plus HALUI/MDI and
   the config-owned Tcl process scripts, not the standalone interpreter. Do not
-  promote this row until the boundary design in
-  `docs/full-process-boundary-design.md` has a deterministic state proof for
-  these HAL pin updates across native, Node, and browser. The Node inventory
-  guard requires this row to stay non-`REP` in Node and browser while it remains
-  `L4-USER-M-PROCESS`, and the generated inventory summary must keep it as
-  `SKIP` with that reason.
+  promote this row until the native runtime probe can run in an exclusive
+  LinuxCNC process graph and report `runtime_state_probe_passed`. The current
+  Web/virtual HAL simulation proof is complete for `M429 -> M129` turn and
+  `M428 -> M128` mill state transitions, and the real browser simulation page
+  smoke verifies the switchkins guard pins plus `ini.[xz].*` limit pins through
+  `linuxCncRealSimulationState.millturnUserMProcess`. That proof is simulation
+  evidence only: it keeps `processExecutionReady=false`, `executionEnabled=0`,
+  and `promotion_allowed=0`. The latest opt-in native probe on this workstation
+  reports `blocked_existing_linuxcnc_runtime` with
+  `exclusive_linuxcnc_runtime` because an existing `linuxcncsvr`/`rtapi_app`
+  runtime is already active. The Node inventory guard requires this row to stay
+  non-`REP` in Node and browser while it remains `L4-USER-M-PROCESS`, and the
+  generated inventory summary must keep it as `SKIP` with that reason.
 - `axis/db_demo/base.ngc` remains `L4-TOOL-DB`. The source INI declares
   `[EMCIO]DB_PROGRAM = ./db_nonran.py`, so standalone interpreter execution
   would miss LinuxCNC tool-database process startup and protocol/state

wasm-port/docs/sim-configs-coverage-handoff.md

@@ -43,6 +43,7 @@ The generated artifact baseline is:
 - `build/wasm/sim-configs-inventory/full-process-boundary-summary.tsv`
 - `build/wasm/sim-configs-inventory/python-remap-boundary-summary.tsv`
 - `build/wasm/sim-configs-inventory/promotion-candidates.tsv`
+- `build/wasm/sim-configs-inventory/remaining-skip-main-program-promotion-audit.tsv`
 - `build/wasm/sim-configs-inventory/evidence-expansion-candidates.tsv`
 - `build/wasm/sim-configs-inventory/runtime-boundary-promotion-blockers.tsv`
 - `build/wasm/sim-configs-inventory/runtime-boundary-host-readiness-rollup.tsv`
@@ -55,6 +56,13 @@ Every row keeps `promotion_allowed=0`, so the inventory baseline is now
 `82/82/77/0`. The companion `evidence-expansion-candidates.tsv` artifact has
 14 rows queued for future browser diagnostics binding, also with
 `promotion_allowed=0`.
+`remaining-skip-main-program-promotion-audit.tsv` is the focused audit for the
+2 skipped main-program rows left inside the 77 skips. It keeps both
+`promotion_allowed=0`: `axis/vismach/millturn/example.ngc` is still blocked on
+incomplete user-M process native/inventory promotion proof, even though the
+browser simulation state now includes a source-derived virtual HAL proof for
+the `M429 -> M129` turn and `M428 -> M128` mill state transitions. The TWP
+`incremental_repetition_g533.ngc` row remains preserved upstream demo evidence.
 
 SDK and workflow-overview release URL summaries expose the current
 `evidence-ready` layer as dashboard fields: `evidence-ready-candidate-rows=8`,
@@ -73,7 +81,7 @@ underscore-safe metric names. Treat these as the same baseline:
 | Artifact reason | Shell metric suffix | Count | Release decision |
 | --- | --- | ---: | --- |
 | `ASSET-ONLY` | `ASSET_ONLY` | 65 | Subroutine/support asset; not a standalone browser main-program target. |
-| `L4-USER-M-PROCESS` | `L4_USER_M_PROCESS` | 1 | Keep blocked until a LinuxCNC-owned external user-M process state proof exists. |
+| `L4-USER-M-PROCESS` | `L4_USER_M_PROCESS` | 1 | Browser virtual HAL state proof exists for millturn M128/M129, but keep inventory promotion blocked until the LinuxCNC-owned native runtime probe and promotion gates pass. |
 | `NON_MAIN_CLASS` | `NON_MAIN_CLASS` | 10 | Covered by class representative or macro/load policy, not direct standalone promotion. |
 | `UPSTREAM-DEMO` | `UPSTREAM_DEMO` | 1 | Preserved upstream demo edge; do not force pass. |
 

wasm-port/docs/sim-configs-coverage-matrix.md

@@ -72,6 +72,8 @@ Current baseline:
   `wasm-port/build/wasm/sim-configs-inventory/next-boundary-recommendations.tsv`
 - Node promotion candidate artifact:
   `wasm-port/build/wasm/sim-configs-inventory/promotion-candidates.tsv`
+- Node remaining skipped main-program promotion audit artifact:
+  `wasm-port/build/wasm/sim-configs-inventory/remaining-skip-main-program-promotion-audit.tsv`
 - Node native evidence acceptance gate artifact:
   `wasm-port/build/wasm/sim-configs-inventory/runtime-boundary-native-evidence-acceptance-gate.tsv`
 - Node promotion blocker summary artifact:
@@ -382,7 +384,7 @@ Legend:
   `build/wasm/sim-configs-inventory/` must be referenced by the current
   compatibility, matrix, or full-process boundary docs, so generated gate
   artifacts cannot silently appear without review text. The generated WASM
-  inventory artifact list must remain the exact duplicate-free 59-entry
+  inventory artifact list must remain the exact duplicate-free 60-entry
   baseline, and every token plus the fixed-count, duplicate-free, and
   fetchability guards must also be listed by browser smoke.
 - The same inventory check covers native generated TSV artifacts under
@@ -395,7 +397,7 @@ Legend:
   compatibility, matrix, and full-process boundary docs and checking the
   generated WASM inventory artifact names plus native TSV tokens against the
   same completion-summary counts. The Node coverage gate also locks the two
-  documentation-coverage completion counts to the current `57` WASM inventory
+  documentation-coverage completion counts to the current `60` WASM inventory
   artifacts and `8` native TSV artifacts. It verifies that browser smoke
   executes the artifact documentation coverage helper, calls the completion
   count-parity helper with the generated completion rows, fetches and joins the