Correct the PTP page: no PTP on a 4B, but the CM4 has it

Reader sent Geerling's CM4 PTP write-up, which shows two CM4s syncing to 10-15ns.
Our headline overclaimed ('impossible on a Pi 4') and our mechanism was wrong (we
blamed the bcmgenet MAC for not exposing a PHC).

The real story is better: the CM4 (BCM54210PE) and the 4B (BCM54213PE) report the
IDENTICAL PHY ID 0x600d84a2, so the kernel cannot tell them apart by asking the
chip. bcm-phy-ptp.c disambiguates by MDIO bus address instead — addr 0 = CM4,
proceed; addr 1 = 4B, return NULL. Ours is addr 1.

It is not a kernel config gap and not our RT patch: CONFIG_BCM_NET_PHYPTP=y and
CONFIG_NETWORK_PHY_TIMESTAMPING=y are both set. The driver loads, looks at the
board, and declines.

Correction is visible on the page rather than quietly patched. Credit to Geerling.

README.md

@@ -2,7 +2,7 @@
 
 **What the Raspberry Pi time-server guides get wrong, and the numbers to prove it.**
 
-Docs: **[cuckoo.warehack.ing](https://cuckoo.warehack.ing)**
+Docs: **[escapement.warehack.ing](https://escapement.warehack.ing)**
 
 We built a GPS-disciplined Stratum 1 NTP server on a Raspberry Pi 4, followed the
 published advice, and measured everything. Most of that advice is wrong on this

dashboard/src/gpsntp_dashboard/static/style.css

@@ -100,8 +100,22 @@ svg { width: 1.15em; height: 1.15em; fill: none; stroke: currentColor; stroke-wi
 @keyframes pulse { 70% { box-shadow: 0 0 0 7px rgba(52, 211, 153, 0); } 100% { box-shadow: 0 0 0 0 rgba(52, 211, 153, 0); } }
 
 /* ---------- layout ---------- */
-main { max-width: 1180px; margin: 0 auto; padding: 20px; display: grid; gap: 18px; }
+/* `grid-template-columns: minmax(0, 1fr)` is load-bearing, not decoration.
+ *
+ * A grid column defaults to `auto`, whose MINIMUM is min-content — and grid items
+ * refuse to shrink below it. The sources table has seven columns, giving it a
+ * ~700px min-content, and that number propagates up and pins `main` to 771px
+ * inside a 375px phone. Every panel then overflows and the document gets a
+ * horizontal scrollbar. `.table-hold { overflow-x: auto }` cannot rescue it,
+ * because the wrapper's own ancestors still won't shrink.
+ *
+ * The 0 in minmax(0, 1fr) overrides that automatic min-content floor. Same reason
+ * .panel needs min-width: 0. Remove either and mobile breaks again.
+ */
+main { max-width: 1180px; margin: 0 auto; padding: 20px;
+  display: grid; grid-template-columns: minmax(0, 1fr); gap: 18px; }
 .panel {
+  min-width: 0;
   background: linear-gradient(var(--panel), var(--bg-2));
   border: 1px solid var(--border-soft); border-radius: 16px; padding: 18px;
   box-shadow: var(--shadow);
@@ -254,10 +268,77 @@ main { max-width: 1180px; margin: 0 auto; padding: 20px; display: grid; gap: 18p
 .ss-plate__cta svg { width: 14px; height: 14px; transition: transform .2s; }
 .ss-plate__link:hover .ss-plate__cta svg { transform: translateX(2px); }
 
+/* ---------- mobile ----------
+ *
+ * This is a status page. On a phone you are almost always asking ONE question:
+ * "is the clock still good?" So the hero answer — time, stratum, lock — must be
+ * readable without scrolling, and everything else can wait below the fold.
+ *
+ * The tables and the signal bars scroll sideways WITHIN their panels rather than
+ * being crushed into unreadability. That's deliberate: a squeezed IPv6 address
+ * wrapped over four lines is worse than one you swipe.
+ */
+@media (max-width: 700px) {
+  main { padding: 14px; gap: 14px; }
+  .panel { padding: 14px; border-radius: 14px; }
+
+  /* A panel head is a title plus its readout. Side by side they fight for a
+     narrow line and the title wraps mid-phrase; stacked, both stay whole. */
+  .panel-head { flex-direction: column; align-items: flex-start; gap: 6px; }
+  .spark-meta { font-size: 11px; }
+
+  /* Stack the clock above the stratum badge, and left-align the badge so the
+     eye travels straight down instead of ping-ponging left-right. */
+  .hero { flex-direction: column; align-items: flex-start; gap: 14px; }
+  .verdict { text-align: left; width: 100%; }
+  .ref-line { margin-top: 6px; }
+
+  /* clamp() already scales this, but 11vw on a narrow phone leaves the seconds
+     hard against the edge. Give it a slightly bigger floor and tighter tracking. */
+  .clock { font-size: clamp(38px, 13vw, 60px); letter-spacing: 0; }
+
+  /* Two columns, not one: six stat cards single-file is a lot of thumb. */
+  .cards { grid-template-columns: repeat(2, minmax(0, 1fr)); gap: 10px; }
+  .card { padding: 12px; }
+  .card-value { font-size: 21px; }
+
+  .spark-big { height: 90px; }
+  .snr { min-height: 150px; }
+
+  /* Swipe the source table rather than wrap it. The shadow hint tells the reader
+     there IS more to the right — otherwise a cut-off column just looks broken. */
+  .table-hold {
+    overflow-x: auto;
+    -webkit-overflow-scrolling: touch;
+    background:
+      linear-gradient(to right, var(--panel) 30%, transparent) left / 24px 100% no-repeat,
+      linear-gradient(to left, var(--panel) 30%, transparent) right / 24px 100% no-repeat,
+      radial-gradient(farthest-side at 0 50%, rgba(0,0,0,.35), transparent) left / 12px 100% no-repeat,
+      radial-gradient(farthest-side at 100% 50%, rgba(0,0,0,.35), transparent) right / 12px 100% no-repeat;
+    background-attachment: local, local, scroll, scroll;
+  }
+  .tbl { min-width: 560px; }
+
+  .topbar { padding: 12px 14px; }
+  .brand h1 { font-size: 17px; }
+  .brand-sub { font-size: 11.5px; }
+  .brand-icon { width: 26px; height: 26px; }
+
+  .foot { padding: 10px 14px 24px; }
+  /* Three items on one line is unreadable at 375px — let them stack. */
+  .foot-meta { flex-direction: column; align-items: flex-start; gap: 6px; }
+}
+
 @media (max-width: 560px) {
   .ss-plate__link { flex-direction: column; align-items: flex-start; }
 }
 
+/* Comfortable thumb targets. The WCAG 2.2 minimum is 24px; 44px is the iOS HIG
+   figure and the one that actually feels right when you're standing up. */
+@media (pointer: coarse) {
+  .foot-docs, .ss-plate__link { min-height: 44px; }
+}
+
 @media (prefers-reduced-motion: reduce) { * { animation: none !important; transition: none !important; } }
 
 /* "approx" marker when GPSNTP_POSITION=coarse */

docs-site/.env.example

@@ -4,6 +4,6 @@
 
 COMPOSE_PROJECT_NAME=cuckoo-escapement-docs
 
-# Production:  cuckoo.warehack.ing
-# Local dev:   cuckoo.l.warehack.ing   (internal only — never reference publicly)
-DOMAIN=cuckoo.warehack.ing
+# Production:  escapement.warehack.ing
+# Local dev:   escapement.l.warehack.ing   (internal only — never reference publicly)
+DOMAIN=escapement.warehack.ing

docs-site/Makefile

@@ -56,4 +56,4 @@ deploy: ## Pull main + rebuild the docs container on the prod host
 	@echo "==> deploying $(DEPLOY_HOST):$(DEPLOY_PATH)"
 	ssh -A $(DEPLOY_HOST) "cd $(DEPLOY_PATH) && git fetch origin main && git reset --hard origin/main && cd docs-site && make prod"
 	@echo "==> sanity check"
-	@curl -s -o /dev/null -w "    HTTP %{http_code}  %{url_effective}\n" "https://cuckoo.warehack.ing/explanation/preempt-rt-made-it-worse/"
+	@curl -s -o /dev/null -w "    HTTP %{http_code}  %{url_effective}\n" "https://escapement.warehack.ing/explanation/preempt-rt-made-it-worse/"

docs-site/astro.config.mjs

@@ -10,16 +10,17 @@
 // when the dev server runs behind Caddy (TLS-terminating proxy) — without an
 // explicit host/protocol/clientPort, Vite's WebSocket drops every ~10s.
 //
-// Site URL comes from DOMAIN so one image serves both cuckoo.warehack.ing (prod)
-// and cuckoo.l.warehack.ing (local dev).
+// Site URL comes from DOMAIN so one image serves both escapement.warehack.ing (prod)
+// and escapement.l.warehack.ing (local dev).
 
 import mdx from "@astrojs/mdx";
 import sitemap from "@astrojs/sitemap";
 import starlight from "@astrojs/starlight";
 import { defineConfig } from "astro/config";
+import remarkGfm from "remark-gfm";
 import starlightLinksValidator from "starlight-links-validator";
 
-const domain = process.env.DOMAIN ?? "cuckoo.warehack.ing";
+const domain = process.env.DOMAIN ?? "escapement.warehack.ing";
 
 export default defineConfig({
   site: `https://${domain}`,
@@ -39,12 +40,20 @@ export default defineConfig({
       description:
         "What the Raspberry Pi time-server guides get wrong, and the numbers to prove it. " +
         "GPS Stratum 1 on a Pi 4: PREEMPT_RT makes PPS jitter worse, the PPS interrupt " +
-        "cannot be pinned, PTP is impossible, and your dashboard is taxing your clock.",
+        "cannot be pinned, PTP needs a CM4 not a 4B, and your dashboard is taxing your clock.",
 
       // The mark IS the word "cuckoo" — a rebus. replacesTitle stops Starlight
       // rendering the title text beside it (which would read "…escapement The
       // Cuckoo Escapement"). The SVG's aria-label carries the full name.
-      logo: { src: "./src/assets/logo.svg", replacesTitle: true },
+      //
+      // Two variants, because Starlight renders the logo as an <img> and CSS
+      // cannot reach inside it. The single cream-on-black logo was invisible in
+      // the light theme — cream wordmark on a cream page.
+      logo: {
+        light: "./src/assets/logo-light.svg",
+        dark: "./src/assets/logo.svg",
+        replacesTitle: true,
+      },
       favicon: "/favicon.svg",
       customCss: ["./src/styles/brass.css"],
 
@@ -99,7 +108,11 @@ export default defineConfig({
       pagination: true,
       lastUpdated: true,
     }),
-    mdx(),
+    // GFM tables do NOT render in .mdx without this, even though they work fine
+    // in plain .md — Astro's remark-gfm doesn't reach the MDX pipeline here. The
+    // failure is silent and ugly: the table falls through as a paragraph of raw
+    // pipe characters. It shipped that way on the landing page. Don't remove.
+    mdx({ remarkPlugins: [remarkGfm] }),
     sitemap(),
   ],
 });

docs-site/docker-compose.yml

@@ -2,7 +2,7 @@
 #
 # Default (no --profile flag):
 #   prod-style — Caddy serves the built dist/. Use for production-like
-#   deploys (the public site at cuckoo.warehack.ing runs this).
+#   deploys (the public site at escapement.warehack.ing runs this).
 #
 # --profile dev:
 #   Astro dev server with HMR. Volume mounts on src/ so edits hot-reload.
@@ -13,7 +13,7 @@
 #
 # Both services attach to the external `caddy` network and expose
 # themselves to caddy-docker-proxy via labels. Edit DOMAIN in .env to
-# switch between cuckoo.warehack.ing (prod) and cuckoo.l.warehack.ing
+# switch between escapement.warehack.ing (prod) and escapement.l.warehack.ing
 # (local-dev tier).
 
 services:
@@ -28,10 +28,16 @@ services:
     networks:
       - caddy
     labels:
-      caddy: ${DOMAIN:-cuckoo.warehack.ing}
+      caddy: ${DOMAIN:-escapement.warehack.ing}
       caddy.reverse_proxy: "{{upstreams 80}}"
       # encode + gzip already in the container; let caddy pass through.
 
+      # The site was briefly live at cuckoo.warehack.ing before the rename. Keep
+      # the old host as a permanent redirect rather than letting any link that
+      # escaped rot. {uri} preserves the path, so deep links survive too.
+      caddy_1: cuckoo.warehack.ing
+      caddy_1.redir: https://escapement.warehack.ing{uri} permanent
+
   docs-dev:
     profiles: ["dev"]
     build:
@@ -41,7 +47,7 @@ services:
     container_name: cuckoo-escapement-docs-dev
     restart: unless-stopped
     environment:
-      - DOMAIN=${DOMAIN:-cuckoo.l.warehack.ing}
+      - DOMAIN=${DOMAIN:-escapement.l.warehack.ing}
       - ASTRO_TELEMETRY_DISABLED=1
     volumes:
       # Hot-reload bind mounts. node_modules stays inside the container
@@ -53,7 +59,7 @@ services:
     networks:
       - caddy
     labels:
-      caddy: ${DOMAIN:-cuckoo.l.warehack.ing}
+      caddy: ${DOMAIN:-escapement.l.warehack.ing}
       caddy.reverse_proxy: "{{upstreams 4321}}"
       # Vite HMR over WebSocket. Caddy's defaults close "idle" WS
       # connections after ~10-15s; HMR doesn't send app-level pings, so

docs-site/package-lock.json

@@ -12,6 +12,7 @@
         "@astrojs/sitemap": "^3.7.2",
         "@astrojs/starlight": "^0.39.2",
         "astro": "^6.3.1",
+        "remark-gfm": "^4.0.1",
         "sharp": "^0.34.0",
         "starlight-links-validator": "^0.24.0"
       }

docs-site/package.json

@@ -14,6 +14,7 @@
     "@astrojs/sitemap": "^3.7.2",
     "@astrojs/starlight": "^0.39.2",
     "astro": "^6.3.1",
+    "remark-gfm": "^4.0.1",
     "sharp": "^0.34.0",
     "starlight-links-validator": "^0.24.0"
   }

docs-site/src/assets/logo-light.svg

@@ -0,0 +1,43 @@
+<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 360 78" role="img" aria-label="The Cuckoo Escapement">
+  <style>
+    .mono  { font-family: "JetBrains Mono","SF Mono",Menlo,monospace; font-weight: 700; }
+    .cream { fill: #2b2721; }
+    .gear  { fill: none; stroke: #a97c1f; stroke-width: 1.9; stroke-linejoin: round; }
+    .rim   { fill: none; stroke: #a97c1f; stroke-width: 1.3; opacity: .8; }
+    .tick  { stroke: #a97c1f; stroke-width: 1.4; opacity: .5; stroke-linecap: round; }
+    .head  { fill: #2b2721; }
+    .eye   { fill: #faf7f1; }
+  </style>
+
+  <!-- one ring, two readings: gear teeth outside (escapement), hour ticks inside (dial) -->
+  <path d="M 30.61 13.60 L 34.99 6.14 L 41.01 6.14 L 45.39 13.60 L 46.36 13.91 L 54.29 10.46 L 59.16 14.00 L 58.33 22.60 L 58.93 23.43 L 67.37 25.29 L 69.23 31.02 L 63.49 37.49 L 63.49 38.51 L 69.23 44.98 L 67.37 50.71 L 58.93 52.57 L 58.33 53.40 L 59.16 62.00 L 54.29 65.54 L 46.36 62.09 L 45.39 62.40 L 41.01 69.86 L 34.99 69.86 L 30.61 62.40 L 29.64 62.09 L 21.71 65.54 L 16.84 62.00 L 17.67 53.40 L 17.07 52.57 L 8.63 50.71 L 6.77 44.98 L 12.51 38.51 L 12.51 37.49 L 6.77 31.02 L 8.63 25.29 L 17.07 23.43 L 17.67 22.60 L 16.84 14.00 L 21.71 10.46 L 29.64 13.91 Z" class="gear"/>
+  <circle cx="38.0" cy="38.0" r="23.5" class="rim"/>
+  <g class="tick">
+    <line x1="38.00" y1="15.50" x2="38.00" y2="18.50"/>
+    <line x1="49.25" y1="18.51" x2="47.75" y2="21.11"/>
+    <line x1="57.49" y1="26.75" x2="54.89" y2="28.25"/>
+    <line x1="60.50" y1="38.00" x2="57.50" y2="38.00"/>
+    <line x1="57.49" y1="49.25" x2="54.89" y2="47.75"/>
+    <line x1="49.25" y1="57.49" x2="47.75" y2="54.89"/>
+    <line x1="38.00" y1="60.50" x2="38.00" y2="57.50"/>
+    <line x1="26.75" y1="57.49" x2="28.25" y2="54.89"/>
+    <line x1="18.51" y1="49.25" x2="21.11" y2="47.75"/>
+    <line x1="15.50" y1="38.00" x2="18.50" y2="38.00"/>
+    <line x1="18.51" y1="26.75" x2="21.11" y2="28.25"/>
+    <line x1="26.75" y1="18.51" x2="28.25" y2="21.11"/>
+  </g>
+
+  <!-- the cuckoo, head only, beak open mid-call -->
+  <g class="head">
+    <!-- crest: three feathers, swept back. this is what says "bird" -->
+    <path d="M 30 26 L 27 16 L 35 23 Z"/>
+    <path d="M 36 23 L 36 13 L 42 22 Z"/>
+    <path d="M 42 23 L 46 15 L 47 25 Z"/>
+    <circle cx="37" cy="39" r="13.5"/>                 <!-- big round head -->
+    <path d="M 47 33.5 L 62 34.5 L 48.5 40 Z"/>        <!-- upper beak: SHORT + chunky -->
+    <path d="M 47.5 42 L 60 46.5 L 46 44 Z"/>          <!-- lower beak, gape open -->
+  </g>
+  <circle cx="40.5" cy="35.5" r="2.2" class="eye"/>
+
+  <text x="92" y="52" class="mono cream" font-size="33">escapement</text>
+</svg>

docs-site/src/content/docs/explanation/no-ptp-on-a-pi-4.md

@@ -1,14 +1,33 @@
 ---
-title: Why PTP is off the table on a Pi 4
-description: PTP's entire value is hardware timestamping. The Pi 4's NIC has no PTP hardware clock. Software PTP is a worse NTP.
+title: No PTP on a Pi 4 Model B (but yes on a CM4)
+description: The CM4 does hardware PTP. The Pi 4 Model B does not — and both report the identical PHY ID, so the kernel tells them apart by MDIO address and declines ours by name.
 sidebar:
   order: 3
 ---
 
-The reference builds all reach for **PTP** (IEEE 1588), and they're right to: on
-the right hardware it's dramatically better than NTP.
+:::caution[Corrected — 2026-07-14]
+This page used to be called *"Why PTP is off the table on a Pi 4"* and blamed the
+`bcmgenet` MAC driver for not exposing a PHC. **The headline overclaimed and the
+mechanism was wrong.** A reader sent us
+[Jeff Geerling's CM4 PTP write-up](https://www.jeffgeerling.com/blog/2022/ptp-and-ieee-1588-hardware-timestamping-on-raspberry-pi-cm4/),
+which shows the **Compute Module 4 doing hardware PTP** — two of them synced to
+within 10–15 nanoseconds.
 
-The Pi 4 is not the right hardware. One command settles it:
+Our measurement was real, but it is specific to the **Pi 4 Model B**, and the
+actual reason is far better than "unsupported." Corrected version below. This is
+the site working as intended.
+:::
+
+PTP's entire value is **hardware timestamping**: the network silicon stamps the
+packet as it crosses the wire, outside the operating system. That's what takes
+scheduling, driver latency and queueing out of the measurement, and it's why PTP
+reaches nanoseconds where NTP reaches microseconds.
+
+With no PHC (PTP Hardware Clock), PTP is just a protocol — packets stamped by the
+*kernel*, on the *CPU*, subject to precisely the jitter you were trying to escape.
+A more complicated NTP with worse tooling.
+
+On our board:
 
 ```console
 $ ethtool -T eth0
@@ -17,36 +36,106 @@ Capabilities:
 	software-receive
 	software-system-clock
 PTP Hardware Clock: none
-Hardware Transmit Timestamp Modes: none
-Hardware Receive Filter Modes: none
+
+$ ls /dev/ptp*
+ls: cannot access '/dev/ptp*': No such file or directory
 ```
 
-**`PTP Hardware Clock: none`.** There isn't one. There's no `/dev/ptp0` to open.
+On a CM4 that same command reports `PTP Hardware Clock: 0`, and
+`/sys/class/ptp/ptp0/clock_name` says `bcm_phy_ptp`. Same family, same era,
+different answer.
 
-## Why that's fatal rather than inconvenient
+## Two different chips — that report the same ID
 
-PTP's whole advantage is **hardware timestamping**: the network card itself
-stamps the packet as it crosses the wire, in silicon, outside the operating
-system. That's what removes kernel scheduling, driver latency, and queueing from
-the measurement, and it's why PTP reaches nanoseconds where NTP reaches
-microseconds.
+| board | PHY | hardware PTP |
+|---|---|---|
+| Compute Module 4 | BCM54210PE | **yes** |
+| Compute Module 5 | — | **yes**, out of the box |
+| Pi 4 Model B | BCM54213PE | no |
 
-Take the hardware clock away and PTP is just... a protocol. Software-timestamped
-PTP has the packets stamped by the *kernel*, on the *CPU*, subject to exactly the
-scheduling jitter you were trying to escape. It is a more complicated NTP with
-worse tooling.
+Here's the part that makes this worth a page. **Both PHYs report an identical ID:**
 
-## But the guides say the Pi 4's PHY supports PTP
+```console
+$ cat /sys/class/mdio_bus/unimac-mdio--19/unimac-mdio--19:01/phy_id
+0x600d84a2
+```
 
-They do, and the *chip* does — the BCM54213PE PHY has PTP capability on paper.
-It doesn't matter. The Pi 4's `bcmgenet` MAC driver doesn't expose a PHC, so
-Linux has nothing to give you. And the reference builds that make PTP work feed
-the PPS into the NIC through a **SYNC pin that only the CM4/CM5 break out** — a
-regular Pi 4 board doesn't route it anywhere you can reach.
+That value is `PHY_ID_BCM54213PE` in `include/linux/brcmphy.h` — and a CM4
+reports it too. The silicon genuinely differs; the identifier does not. So the
+kernel *cannot* tell these boards apart by asking the chip who it is.
 
-## So don't chase it
+## So it tells them apart by where they're plugged in
 
-We spent real time on this before running `ethtool -T`, which we should have run
-first. If your board reports `PTP Hardware Clock: none`, close the tab. Put the
-effort into the PPS path instead — that's where the nanoseconds actually are, and
+From the Raspberry Pi kernel, `drivers/net/phy/bcm-phy-ptp.c`, comments and all:
+
+```c
+struct bcm_ptp_private *bcm_ptp_probe(struct phy_device *phydev)
+{
+	switch (BRCM_PHY_MODEL(phydev)) {
+	case PHY_ID_BCM54210E:
+		break;
+#ifdef PHY_ID_BCM54213PE
+	case PHY_ID_BCM54213PE:
+		switch (phydev->mdio.addr) {
+		case 0: // CM4 - this is a BCM54210PE which supports PTP
+			break;
+		case 1: //  4B - this is a BCM54213PE which doesn't
+			return NULL;
+		default: // Unknown - assume it's BCM54210PE
+			break;
+		}
+		break;
+```
+
+The disambiguator is the **MDIO bus address**. A CM4's PHY sits at address 0; a
+Pi 4 Model B's sits at address 1. Ours:
+
+```console
+$ ls -d /sys/class/mdio_bus/unimac-mdio--19/unimac-mdio--19:*
+/sys/class/mdio_bus/unimac-mdio--19/unimac-mdio--19:01
+                                                    ^^ address 1 — the 4B branch
+```
+
+Address 1 → `return NULL` → no PHC is ever registered. That is the whole story.
+
+## It is not your kernel, and there is no option to flip
+
+This is the part worth internalising, because it's a trap that looks exactly like
+a misconfiguration. All the support is *present*:
+
+```console
+$ grep -E "CONFIG_NETWORK_PHY_TIMESTAMPING|CONFIG_BCM_NET_PHYPTP" .config
+CONFIG_NETWORK_PHY_TIMESTAMPING=y
+CONFIG_BCM_NET_PHYPTP=y
+```
+
+The driver is compiled in. The PHY ID is a recognised constant. The code path
+exists, and it runs. **It runs, looks at your board, and declines.** No kernel
+option changes that, no patch, no `rpi-update` — the refusal is a deliberate and
+correct statement about the silicon in front of it.
+
+It also means our [`IRQF_NO_THREAD` kernel](/reference/the-patch/) is not the
+culprit. That was the obvious suspicion when we re-opened this, so we checked it
+first. It isn't.
+
+## What to actually do
+
+**On a Pi 4 Model B:** stop. Spend the effort on the PPS path, which is where the
+nanoseconds actually live — and
 [it needs the help](/explanation/preempt-rt-made-it-worse/).
+
+**If you want PTP on a Pi:** buy a **CM4 or CM5**, not a 4B. You get `/dev/ptp0`,
+`ethtool -T` reports hardware transmit and receive timestamping, and on the CM4 IO
+board you can drive a PPS out of pin 9. Geerling's post is the guide; we're not
+going to rewrite it.
+
+**Either way, run `ethtool -T` first.** We didn't, and burned real time on a
+question that one command answers.
+
+:::note[Credit]
+The correction, and every CM4 figure here, comes from
+[Jeff Geerling's write-up](https://www.jeffgeerling.com/blog/2022/ptp-and-ieee-1588-hardware-timestamping-on-raspberry-pi-cm4/)
+and the kernel patches it links to. We had the right conclusion for the wrong
+reason, and stated it too broadly. Checking beats believing — including when the
+thing being checked is us.
+:::

docs-site/src/content/docs/findings.md

@@ -12,11 +12,18 @@ A GPS-disciplined Stratum 1 NTP server: **Raspberry Pi 4** + **BerryGPS-IMU v4**
 199 ns**, root delay ~1 ns, survives a cold power cut unattended. About $130 of
 parts, replacing an appliance that costs $1,500–$10,000.
 
-## What the guides get wrong on a Pi 4
+## What the guides get wrong on a Pi 4 Model B
+
+:::note[Model B, specifically]
+These are properties of the **Pi 4 Model B**, not the whole Pi 4 family. The
+**CM4 and CM5 do hardware PTP** — [we got that one wrong at first, and corrected
+it](/explanation/no-ptp-on-a-pi-4/). If a claim here matters to you, check it on
+*your* board. That's the entire point of this site.
+:::
 
 | Claim | Reality |
 |---|---|
-| "Use PTP for real precision" | **Impossible.** `ethtool -T eth0` → `PTP Hardware Clock: none`. No hardware timestamping exists on this NIC. |
+| "Use PTP for real precision" | **Not on a 4B — but yes on a CM4.** Both report the *same* PHY ID, so the kernel disambiguates by MDIO address and hands the 4B a `return NULL`. `ethtool -T` → `PTP Hardware Clock: none`. |
 | "Isolate the PPS IRQ on a dedicated core" | **Not permitted.** GPIO IRQs demux through `pinctrl-bcm2835` and reject `smp_affinity`. |
 | "Install PREEMPT_RT" | **Made jitter 3× worse** until patched — it threads the handler that takes the timestamp. |
 | "Raise the GPS baud rate" | **Irrelevant.** PPS offset measured −1 ns at 9600 vs 115200. Identical. NMEA only *labels* the second. |

docs-site/src/content/docs/index.mdx

@@ -76,9 +76,10 @@ board rather than take our word for it. That's the whole point.
     costs you the accuracy.
     [→ Why](/explanation/the-interrupt-you-cannot-move/)
   </Card>
-  <Card title="PTP is impossible on a Pi 4" icon="error">
-    `ethtool -T eth0` → `PTP Hardware Clock: none`. There is no hardware
-    timestamping. Software PTP is just a worse NTP. Don't chase it.
+  <Card title="No PTP on a 4B — but yes on a CM4" icon="error">
+    The CM4 and the 4B report the **same PHY ID**, so the kernel tells them apart
+    by MDIO address and hands the 4B a `return NULL`. The driver is compiled in;
+    it looks at your board and declines. Buy a CM4 if you want PTP.
     [→ Why](/explanation/no-ptp-on-a-pi-4/)
   </Card>
   <Card title="Your dashboard is taxing your clock" icon="rocket">