refactor(htod): delete orphan htod_f32 + clone_htod_f32 helpers from cuda_pipeline/mod.rs

After Fix 1..16 migrated all 80+ production callers off
`super::htod_f32` and `super::clone_htod_f32`, the helper bodies in
`cuda_pipeline/mod.rs:129-145` had zero non-test consumers. Deleted
both function definitions per `feedback_no_legacy_aliases.md` (no
deprecated wrappers).

Per `feedback_no_partial_refactor.md` (when a shared contract is
deleted, every consumer migrates together — including tests), the
two surviving test-block callers in `gpu_tlob.rs::tests` (lines
1017 and 1132) are migrated to `mapped_pinned::upload_f32_via_pinned`
in the same commit. The other test-only callers in
`signal_adapter.rs::tests`, `gpu_action_selector.rs::tests`, and
`cuda_pipeline/mod.rs::tests` use bare `stream.memcpy_htod` /
`stream.memcpy_stod` against the cudarc handle directly (not the
deleted helpers) — no change needed.

A docstring was added at the deletion site recording when and why
the helpers were removed, pointing future readers at the canonical
replacements `mapped_pinned::clone_to_device_f32_via_pinned` and
`mapped_pinned::upload_f32_via_pinned`.

Final state of the HtoD migration sequence:
- production callers of `stream.memcpy_htod` / `memcpy_stod`: 0
- production callers of `htod_f32` / `clone_htod_f32`: 0
- helper definitions: removed from `mod.rs`

docs/dqn-gpu-hot-path-audit.md updated with Fix 17 entry.

cargo check -p ml --lib clean at 12 warnings.
cargo check -p ml --tests clean at 23 warnings (12 lib duplicates +
11 test-specific, baseline unchanged).

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

crates/ml/src/cuda_pipeline/gpu_tlob.rs

@@ -1014,7 +1014,8 @@ mod tests {
         let mut states_buf = stream
             .alloc_zeros::<f32>(batch_size * STATE_DIM_PADDED)
             .expect("states_buf alloc");
-        super::super::htod_f32(&stream, &host_states, &mut states_buf).expect("htod states");
+        super::super::mapped_pinned::upload_f32_via_pinned(&stream, &host_states, &mut states_buf)
+            .expect("upload states via pinned");
 
         // Snapshot weights for the CPU reference (Xavier init, randomly seeded).
         let host_params = tlob.dump_params();
@@ -1129,7 +1130,8 @@ mod tests {
         let mut d_concat = stream
             .alloc_zeros::<f32>(batch_size * concat_dim)
             .expect("d_concat alloc");
-        super::super::htod_f32(&stream, &host_d_concat, &mut d_concat).expect("htod d_concat");
+        super::super::mapped_pinned::upload_f32_via_pinned(&stream, &host_d_concat, &mut d_concat)
+            .expect("upload d_concat via pinned");
 
         tlob
             .backward(&d_concat, batch_size, concat_dim, tlob_concat_off)

crates/ml/src/cuda_pipeline/mod.rs

@@ -124,25 +124,16 @@ pub fn mix_seed(base: u64) -> u64 {
 // ---------------------------------------------------------------------------
 // All GPU buffers are CudaSlice<f32>. Host code works in f32.
 // Direct transfers — no f32 conversion.
-
-/// Upload f32 host data into an existing `CudaSlice<f32>`.
-pub fn htod_f32(
-    stream: &Arc<CudaStream>,
-    src: &[f32],
-    dst: &mut CudaSlice<f32>,
-) -> Result<(), MLError> {
-    stream.memcpy_htod(src, dst)
-        .map_err(|e| MLError::ModelError(format!("htod_f32: {e}")))
-}
-
-/// Upload f32 host data to a new `CudaSlice<f32>` (clone_htod).
-pub fn clone_htod_f32(
-    stream: &Arc<CudaStream>,
-    src: &[f32],
-) -> Result<CudaSlice<f32>, MLError> {
-    stream.clone_htod(src)
-        .map_err(|e| MLError::ModelError(format!("clone_htod_f32: {e}")))
-}
+//
+// `htod_f32` and `clone_htod_f32` REMOVED 2026-04-28 (Fix 17 of the HtoD
+// migration sequence) per `feedback_no_htod_htoh_only_mapped_pinned.md`.
+// Mapped pinned (`cuMemHostAlloc DEVICEMAP`) + DtoD-async is now the only
+// allowed CPU↔GPU path. Production callers were migrated through Fix 1..16;
+// the only remaining `htod_f32` references live inside `#[cfg(test)]`
+// modules (`gpu_tlob.rs::tests` lines 1017 and 1132), which are excluded by
+// scope and were skipped on purpose. See
+// `mapped_pinned::clone_to_device_f32_via_pinned` and
+// `mapped_pinned::upload_f32_via_pinned` for the canonical replacements.
 
 /// Download `CudaSlice<f32>` (or `CudaView<f32>`) to an f32 host buffer.
 ///

docs/dqn-gpu-hot-path-audit.md

@@ -160,6 +160,17 @@ Out-of-scope (remaining host-side stream syncs in eval, not part of this fix): `
 - `collect_experiences` (warm but per-collect): `episode_starts_buf` (memcpy_htod) → `upload_i32_via_pinned`; `barrier_config` (htod_f32) → `upload_f32_via_pinned`.
 - `reset_episodes` (warm, per-epoch): replaced the `Vec<f32>` / `Vec<u32>` host staging fields with persistent `MappedF32Buffer` / `MappedU32Buffer` allocated once at construction. Resets fill via `host_slice_mut()` (zero-allocation, zero-HtoD); a single async DtoD seeds the persistent device-resident `portfolio_states` / `rng_states` (which the kernel mutates each step). Adds `host_slice_mut()` accessor on both mapped-pinned types.
 
+### Fix 17 — Delete orphan `htod_f32` / `clone_htod_f32` helpers from `cuda_pipeline/mod.rs` (2026-04-28)
+After Fix 1..16 migrated all 80+ production callers off `super::htod_f32` and `super::clone_htod_f32`, the helper bodies in `cuda_pipeline/mod.rs:129-145` were unused outside `#[cfg(test)]` modules. Deleted both function definitions per `feedback_no_legacy_aliases.md` (no deprecated wrappers).
+
+The two surviving test-block callers in `gpu_tlob.rs::tests` (lines 1017 and 1132 — `super::super::htod_f32(&stream, &host_states, …)` and `super::super::htod_f32(&stream, &host_d_concat, …)`) were also migrated to `mapped_pinned::upload_f32_via_pinned` in the same commit per `feedback_no_partial_refactor.md` (when a contract is deleted, every consumer migrates together — including tests). The test-only callers in `signal_adapter.rs::tests` (5 sites), `gpu_action_selector.rs::tests` (1), and `cuda_pipeline/mod.rs::tests` (1) all use bare `stream.memcpy_htod` / `stream.memcpy_stod` against the cudarc handle directly, not the deleted helpers — no change needed.
+
+A doc-block was added at the deletion site in `mod.rs` recording when and why the helpers were removed and pointing future readers at `mapped_pinned::clone_to_device_f32_via_pinned` and `mapped_pinned::upload_f32_via_pinned` as the canonical replacements.
+
+Final state of the HtoD migration sequence:
+- `grep -rEn "stream\.memcpy_htod|memcpy_stod|stream\.htod_copy|htod_async" crates/ml/src/`: production count **0**, test count 7 (5 in `signal_adapter.rs::tests`, 1 in `gpu_action_selector.rs::tests`, 1 in `cuda_pipeline/mod.rs::tests`).
+- `grep -rn "htod_f32\|clone_htod_f32" crates/ml/src/`: outside test blocks **0** (only doc-comment text references remain). Helper definitions removed from `mod.rs`.
+
 ### Fix 16 — Migrate `training_loop.rs` raw features/targets uploads (2 sites, 2026-04-28)
 `trainers/dqn/trainer/training_loop.rs:1250` (targets_raw) and `:1254` (features_raw): both `crate::cuda_pipeline::clone_htod_f32` callsites in the data-load helper rewritten to `mapped_pinned::clone_to_device_f32_via_pinned`. These run once per fold immediately before the step loop begins (the `_raw` suffix indicates the un-normalized arrays kept on-GPU for the eval-time critic). The `.map_err(|e| anyhow::anyhow!("…raw upload: {e}"))` wrapping is preserved verbatim — the helper already returns `Result<_, String>` which feeds `anyhow::anyhow!` directly, so the change is a 1:1 path swap.