cudaq-guide
CUDA-Q onboarding guide for installation, test programs, GPU simulation, QPU hardware, and quantum applications.
jetson-customize-usb
Enable/disable Jetson USB2/USB3 SS ports via kernel-DT overlay. Do NOT use for UPHY lane allocation or ODMDATA edits.
Full skill instructions
Enable, disable, or change the role of USB2 / USB3 SS ports on a
Jetson Thor (Tegra264) or Orin (Tegra234) custom carrier. Captures
per-port wiring (role, max speed, VBUS-EN / OC GPIOs, CC1/CC2 GPIOs
for Type-C, USB3 SS UPHY lane), resolves the SS to USB2 companion
graph from the in-tree DTB, then renders a self-contained kernel-DT
overlay that flips every port action in three places in lockstep
(lane status, port status, host xHCI phys + phy-names).
UPHY lane allocation belongs to jetson-customize-uphy. No ODMDATA
edit. Output is one commit to the composite custom overlay .dts in
the bsp_sources/ hardware repo.
reference_devkit: + custom_carrier: blocks.<source.root_path>/Linux_for_Tegra/.git exists
(/jetson-init-source)./jetson-derive-carrier has run — carrier flash-conf fork in the
overlay tracker./jetson-customize-uphy has run when any enabled USB3 SS port
needs a non-stock UPHY lane allocation. Its JSON sidecar at
<workspace>/target-platform/<profile-stem>.jetson-customize-uphy.json
is consulted for SS lane allocation.custom_carrier: is present, both
documents.custom_carrier_schematic AND
documents.custom_carrier_pinmux_xls are REQUIRED. Refuse the run
if either is missing — per-port routing (VBUS-EN / OC / CC GPIOs, SS
lane wiring, hub fan-out) on a custom carrier cannot be guessed.
Reference-devkit-only profiles skip this check.dtc, fdtoverlay on PATH.USB on Tegra spans three IP surfaces: the xusb_padctl block (USB2
OTG + USB3 SS PHYs), the tegra-xusb xHCI host controller, and an
optional tegra-xudc device controller attached to the single
OTG-capable USB2 port (usb2-0).
A per-port flip MUST touch three kernel-DT places in lockstep.
Anything less crashes the host xHCI probe and leaves lsusb empty
on every port (collateral damage to stock-okay ports):
| # | Place | Path | What it controls |
|---|---|---|---|
| 1 | Lane (PHY provider) | xusb_padctl/pads/usb<2|3>/lanes/usb<2|3>-N | SS / OTG PHY hardware-binding. status="disabled" then lane stops providing a PHY. |
| 2 | Port (controller-binding) | xusb_padctl/ports/usb<2|3>-N | Per-port mode (host/device/otg), companion link, VBUS / OC / CC pin refs. status="disabled" then port removed from user-facing topology. |
| 3 | Host xHCI phys-list | bus@0/usb@<addr>.phys + .phy-names | Array of phandles + names the xHCI driver iterates. A ref to a disabled PHY returns -ENODEV and aborts the whole host probe. |
NVIDIA's stock-disabled usb3-3 in the Thor base DTB is the canonical
pattern — all three places flipped in lockstep.
Two extra rules ride on top of the three-place pattern:
xusb_padctl/ports/usb3-N.nvidia,usb2-companion references a USB2
port phandle. Disabling that USB2 without cascading to its SS
companion then tegra-xusb: failed to enable PHYs: -19.Agentic, not table-driven — every port, controller, lane,
companion link, phandle, and __symbols__ lookup is resolved at
runtime from docs + DTB + carrier pinmap + schematic.
lsusb empty after a previous
jetson-customize-usb attempt) needs to be fixed.jetson-customize-uphy re-allocated UPHY lanes affecting USB3 SS
ports and per-port DT now needs to follow.The full step-by-step procedure lives in references/procedure.md.
AskUserQuestion for port(s) to enable / disable;
surface companion cascade + on-carrier hub fan-out explicitly.pin_verifier.py).usb:padctl, usb:xhci, optional
usb:xudc) to the composite custom overlay .dts, run
fdtoverlay + the three post-merge invariants, commit to
bsp_sources/.references/run-state-sidecar.md), emit headline, then drive the
downstream next-step chain via sequential AskUserQuestion prompts
per references/procedure.md Step 6. Never substitute a printed
"Next step: …" line for the prompts.See references/gotchas.md for the load-bearing failure modes.
status knob; do not edit it.jetson-customize-uphy owns that.
Refuse to commit an SS-enable until uphy run-state shows the lane
allocated./jetson-customize-pinmux set-pin..dtbo or register OVERLAY_DTB_FILE+= —
/jetson-build-source owns build + flash-conf registration.usb2-0 is OTG-capable; xudc
attaches there only. All other USB2 ports and all USB3 SS ports
are host-only.lsusb on every port, USB-eth at 192.168.55.1 still up:
host xHCI bailed; three-place lockstep was broken. Inspect merged
DTB; verify post-merge invariants in references/procedure.md
Step 5d.tegra-xusb: failed to enable PHYs: -19: companion cascade
(Rule B) violated — a USB2 was disabled without its SS companion.no port found or Requested PHY is disabled: Rule A
violated — lane status and port status are mismatched.FDT_ERR_NOTFOUND from fdtoverlay: a fragment used
target = <&label> for a node whose label is not in
__symbols__ (typical for host xHCI / tegra-xudc). Switch to
target-path = "/bus@0/usb@<addr>".phys_property: cell 0 is not a phandle reference:
benign; expected when using raw integer phandles in the host phys
override.vbus-supply references a
regulator parent node that does not exist. Ensure the fixed
regulator node is present before referencing it.phys element order
was not preserved. Only elide disabled entries; never reorder
kept ones.references/procedure.md — full step-by-step procedure.references/gotchas.md — load-bearing failure modes.references/run-state-sidecar.md — run-state JSON shape.references/usb-architecture.md — Tegra USB IP architecture
notes (xusb_padctl, tegra-xusb, xudc).references/usb-dt-bindings.md — USB DT binding cheatsheet
(lane / port / phys-list shapes).../../scripts/pin_verifier.py — shared HSIO pin verifier.../../references/platform_template.yaml — documents: block
consumed by Step 1.../../context/bsp-customization-workflow.md — overlay edit
protocol.../../references/bsp-customization-kernel-dtb.md — composite
overlay append protocol.../jetson-customize-uphy/SKILL.md — sibling skill that owns
UPHY lane allocation.../jetson-customize-pinmux/SKILL.md — sibling skill for
VBUS-EN / OC / CC SFIO fixes.../jetson-derive-carrier/SKILL.md — must run first.../jetson-init-source/SKILL.md — produces the overlay tracker
bsp_sources repo.CUDA-Q onboarding guide for installation, test programs, GPU simulation, QPU hardware, and quantum applications.
Use when the user wants to create a dataset, generate synthetic data, or build a data generation pipeline.
NVIDIA DeepStream SDK development with Python pyservicemaker API. Use when building video analytics pipelines, GStreamer-based video processing, TensorRT inference integration, object detection/tracking, or Kafka/message broker integration.
Filesystem RAG benchmarks: corpus/, train.json, evaluate_rag.py (RAGAS quality). Not for prod monitoring, latency/throughput benchmarking (use rag-perf), or evals outside this repo layout.
Use when the user wants to search, query, extract, transcribe, describe, quote, filter, or aggregate across documents — PDFs, scanned forms / images (`.jpg` `.png` `.tiff`), Office (`.docx` `.pptx`), text (`.html` `.txt`), audio (`.mp3` `.wav` `.m4a`), or video (`.mp4` `.mov`). Prefer this over native Read / Grep for multi-file or non-PDF corpora. Not for: editing files, web browsing, single-file plain-text lookups, fine-tuning.
Coordinate the end-to-end CAD/source-asset to SimReady workflow. Use for broad requests such as CAD to SimReady, source asset to simulation-ready USD, or prop packaging that require conversion, material/physics assignment, SimReady conformance, validation, and optional package creation; deploy or verify Content Agents services first when property assignment is enabled; route single-stage work through nested references.
Use this skill to bring any vision model from HuggingFace or NVIDIA NGC into an NVIDIA DeepStream pipeline with end-to-end automation: ONNX download, SafeTensors export, TRT engine build, custom nvinfer bbox parser, multi-stream benchmark, and PDF report. Object detection models only.
Top-level workflow skill for USD performance diagnosis and optimization. Handles slow loading, high memory, low FPS, and broad scene-optimization requests; delegates auth/runtime setup to Phase 0 owners.
Vehicle routing (VRP, TSP, PDP) with cuOpt — Python API only. Use when the user is building or solving routing in Python.
Use only to generate or update a governance skill card for a specified existing agent skill directory. Do not use for explaining, listing, comparing, or discussing skill capabilities.
Use as the top-level router for Omniverse Realtime Viewer USD app requests and focused viewer reference documents.
Performance benchmarking for a deployed NVIDIA RAG Blueprint server: profiling pass + aiperf load test driven by a single YAML config. Not for accuracy / RAGAS scoring (use rag-eval) or for deploying / repairing services (use rag-blueprint).
Generate, edit, and beat-sync AI video with leading models in one workspace.
The world's fastest calendar for remote work
Transform Your Design with AI Designer by ImgCreator.ai
Revolutionizing Video Production with AI-Powered Creativity
Extend an image past the frame and let AI fill the new aspect ratio.
Discover your celebrity doppelgänger with StarByFace!
ChainClarity explains 700+ crypto whitepapers in plain English, with layered summaries, comparisons, research tools, alerts, and a $4.99 Pro plan.
Opus.ai: Revolutionize Your Web Experience