From d6df3d50d285bc0aab7a4e1bb7513eb2c18f09a5 Mon Sep 17 00:00:00 2001 From: t Date: Mon, 9 Nov 2026 07:20:58 -0400 Subject: [PATCH] docs: the AXB35-02 port page gets an inbound link, so it stops being one of the two pages nobody links `docs/axb35-usb-ports.md` landed in #1567 with no page pointing at it. While reviewing #1650 (which links `docs/tokenizers.md` because it was "the last page with no authored inbound link") I measured the actual set, by grepping every `*.md`, `*.astro`, `*.tsx`, `*.ts` and `*.html` in the tree for each page's filename: architecture.md <- README.md install.md <- README.md lora.md <- README.md moe-streaming.md <- docs/serve.md tokenizers.md <- NONE axb35-usb-ports.md <- NONE <- mine So there were two, and this closes mine. The link goes in `docs/kernel.md`, next to the out-of-tree module paragraph, because the factor is board-firmware rather than a module: three of the four xHCI controllers park in `D3cold`, the rear USB-A and both Type-C ports never enumerate, and the fix that ships is a udev rule. The clause points at the page that carries the ACPI evidence, the measured power states, the rule itself and the one-command check. Verified: the relative link resolves (`docs/axb35-usb-ports.md` exists), `tools/copyright.py --check` rc 0 with the tree staged, `git add ` only, one file. refs #823 (the port finding), #1567 (the page), #1650 (whose claim this corrects by one). --- docs/kernel.md | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/docs/kernel.md b/docs/kernel.md index 97fd8448e..49c9cfa33 100644 --- a/docs/kernel.md +++ b/docs/kernel.md @@ -44,7 +44,10 @@ list, with the facts that go stale marked as facts and the commands that do not. Both are also packaged as dkms packages (`amdxdna-1bit-dkms`, `ec-su-axb35-dkms`) which do that rebuild on any machine that installs them — `scripts/mk-kernel-packages.sh` builds the set for deb, rpm and Arch, and [install-any-distro.md](install-any-distro.md), "Kernel and NPU driver packages", - is the install. + is the install. The board's other firmware-side quirk is not a module at all: three of its four xHCI + controllers park in `D3cold`, so the rear USB-A and both Type-C ports never enumerate, and the fix + that ships for it is a udev rule — written up, with the ACPI evidence and the one-command check, in + [axb35-usb-ports.md](axb35-usb-ports.md). ## 7.4 on the box: what changed, and what to watch