Jetlink e0257f4

warp.make_coherent is public, for an adapter's own warp: it moves a QCOM
output to memory the CPU reads cached, as Jetlink's warp does.
CREDITS.md and JETLINK.md name the new revision.

CREDITS.md

@@ -2,7 +2,7 @@
 
 ## Jetlink external inference
 
-The comma-side Jetlink client and USB transport in `jetlink_repo` are by Zeph Leggett and the [Jetlink contributors](https://github.com/zoompilot/jetlink). They are vendored from revision `0605e34b87c56bcaea6b05c757a6889beebf1422` under the MIT license, retained in [jetlink_repo/LICENSE](jetlink_repo/LICENSE). StarPilot's adapter (`openpilot/starpilot/models/jetlink_adapter.py`) builds on Jetlink's openpilot layer and owns its param names, Always On Lateral rule, model runner reset, settings and alerts. Companion applications and inference servers remain distributed by Jetlink.
+The comma-side Jetlink client and USB transport in `jetlink_repo` are by Zeph Leggett and the [Jetlink contributors](https://github.com/zoompilot/jetlink). They are vendored from revision `e0257f46d6a50bdb16f47b946850fc919047aa82` under the MIT license, retained in [jetlink_repo/LICENSE](jetlink_repo/LICENSE). StarPilot's adapter (`openpilot/starpilot/models/jetlink_adapter.py`) builds on Jetlink's openpilot layer and owns its param names, Always On Lateral rule, model runner reset, settings and alerts. Companion applications and inference servers remain distributed by Jetlink.
 
 ## MRR35 Hyundai angle and radar
 

jetlink_repo/jetlink/openpilot/interface.py

@@ -211,7 +211,9 @@ class JoinedModel(Protocol):
 class FrameWarp(Protocol):
   """What an adapter's model_warp returns, warp.Warp's shape, run on modeld's
   frame thread. Only one of the small and the large model runs a frame, so a
-  warp that shares the small model's JIT and buffers is never in use by both."""
+  warp that shares the small model's JIT and buffers is never in use by both.
+  On a comma, warp.make_coherent lets the CPU read a GPU output cached: 0.22
+  ms for the frame where tinygrad's own mapping takes 2.9."""
   output: Any   # the warped frame, (2, 6, model_h/2, model_w/2) uint8 as bytes: read after wait(), until the next start()
 
   def start(self, frame: int, big_frame: int, tfm: Any, big_tfm: Any) -> None:

jetlink_repo/jetlink/openpilot/warp.py

@@ -161,7 +161,7 @@ class Warp:
     out = jit.captured.ret.uop.base.buffer
     self._device = out.device
     if self._device.startswith('QCOM'):
-      _make_coherent(out)
+      make_coherent(out)
     # the two buffers every frame passes, written in place by start()
     self._tensors = {name: Tensor(np.zeros(shape, dtype=dtype), device=device).realize()
                      for name, (shape, dtype, device) in specs.items()}
@@ -191,10 +191,14 @@ class Warp:
     self._replay(self._inputs, {})
 
 
-def _make_coherent(out) -> None:
-  """Move a QCOM buffer to KGSL memory flagged IOCOHERENT and write-back.
-  Raises if the kernel keeps either flag back (the allocation's record holds
-  its answer): write-back memory that is not coherent reads stale frames."""
+def make_coherent(out) -> None:
+  """Move a QCOM buffer to KGSL memory flagged IOCOHERENT and write-back, so
+  the CPU reads it through its cache once the GPU is done (see Warp). Raises
+  if the kernel keeps either flag back (the allocation's record holds its
+  answer): write-back memory that is not coherent reads stale frames.
+
+  For an adapter's model_warp too. A JIT links its buffers' addresses on its
+  first call, so move one before that, or have the JIT link again after."""
   from tinygrad.device import Device
   dev = Device[out.device]
   mem = dev._gpu_alloc(out.nbytes, flags=COHERENT_WRITEBACK)

openpilot/starpilot/models/JETLINK.md

@@ -28,14 +28,14 @@ release still needs a successful compatible handshake and device validation.
 
 ### Jetson and Linux PC
 
-On a supported NVIDIA host, review and run the [upstream installer](https://github.com/zoompilot/jetlink/blob/0605e34/install.sh).
+On a supported NVIDIA host, review and run the [upstream installer](https://github.com/zoompilot/jetlink/blob/e0257f4/install.sh).
 The upstream Jetson guide specifies an Orin Nano Super 8 GB and JetPack 7.2.1;
 the Linux guide specifies an RTX 20-series or newer GPU with driver 580 or newer.
 Keep the computer powered, cool, and awake. Connect its USB-A port to the comma's USB-C port.
 Use `jetlink status` to check the host and `jetlink logs` to inspect problems.
 Host power, sleep, and driver setup details remain in the credited
-[Jetson](https://github.com/zoompilot/jetlink/blob/0605e34/docs/jetson.md) and
-[Linux](https://github.com/zoompilot/jetlink/blob/0605e34/docs/linux-pc.md) references;
+[Jetson](https://github.com/zoompilot/jetlink/blob/e0257f4/docs/jetson.md) and
+[Linux](https://github.com/zoompilot/jetlink/blob/e0257f4/docs/linux-pc.md) references;
 their comma-side ZoomPilot instructions are replaced by **Set up StarPilot** above.
 
 ### Mac
@@ -45,7 +45,7 @@ Use an Apple silicon Mac with macOS 15 or newer. Get the Jetlink DMG from
 in Applications, and start it. Keep it awake and connect by USB 3.
 Use **USB computer** in StarPilot's Galaxy, rather than following the comma-side
 ZoomPilot installation instructions in the credited
-[Mac host reference](https://github.com/zoompilot/jetlink/blob/0605e34/docs/macos-app.md).
+[Mac host reference](https://github.com/zoompilot/jetlink/blob/e0257f4/docs/macos-app.md).
 
 ### iPhone and iPad
 
@@ -54,7 +54,7 @@ It requires a USB-C device supporting USB 3 and iOS/iPadOS 26.1 or newer.
 Choose **iPhone / iPad** in StarPilot, allow the host app's Local Network access,
 connect by USB 3, and keep the app open and the device unlocked. Check sustained
 thermal performance with the app's benchmark; that benchmark excludes cable latency.
-The [credited Apple host reference](https://github.com/zoompilot/jetlink/blob/0605e34/docs/iphone-app.md)
+The [credited Apple host reference](https://github.com/zoompilot/jetlink/blob/e0257f4/docs/iphone-app.md)
 contains hardware details. StarPilot supplies the comma-side instructions above.
 
 ### Android
@@ -64,7 +64,7 @@ upstream Jetlink Android APK from [its releases](https://github.com/zoompilot/je
 Use a recent USB 3-capable Android 12+ phone, a powered USB 3 hub, and separate
 comma power. Choose **USB computer** in StarPilot. Follow the host's USB permission
 prompt and benchmark sustained performance before use.
-[Credited Android host reference](https://github.com/zoompilot/jetlink/blob/0605e34/docs/android-app.md).
+[Credited Android host reference](https://github.com/zoompilot/jetlink/blob/e0257f4/docs/android-app.md).
 
 ## Use and recovery
 
@@ -89,7 +89,7 @@ and defaults off.
 ## Attribution and license
 
 This StarPilot guide adapts the host setup structure from
-[ZoomPilot / Jetlink at 0605e34](https://github.com/zoompilot/jetlink/tree/0605e34/docs).
+[ZoomPilot / Jetlink at e0257f4](https://github.com/zoompilot/jetlink/tree/e0257f4/docs).
 Jetlink is copyright (c) 2026 Zeph Leggett and licensed under the
 [MIT license](../../../jetlink_repo/LICENSE), whose full text is preserved.
 ZoomPilot source links, project names, authorship, and history are attribution;

upstream-sync.json

@@ -65,8 +65,8 @@
     {
       "path": "jetlink_repo",
       "url": "https://github.com/zoompilot/jetlink.git",
-      "commit": "0605e34b87c56bcaea6b05c757a6889beebf1422",
-      "tree": "fcd6e2cb694def9f0417c192018a1debbfebcfc0",
+      "commit": "e0257f46d6a50bdb16f47b946850fc919047aa82",
+      "tree": "921ec9157a6b607f7f0f81c69cd411b5d23cf872",
       "include": [
         "jetlink",
         "LICENSE",