raspberrypi

Pretty sure these are that, right? github.com/raspberrypi/pico-sdk/tree/master/cmake… 'Cos these should already be called via pico_sdk_init (I'm trying to build e.g. github.com/raspberrypi/pico-examples/tree/master/… )

Pico 不只适合点灯。官方 pico-sdk 提供 C/C++ 接口，覆盖 PIO、定时器、USB 和多核等能力。 想研究自定义外设时序，可以从 PIO 相关示例往下挖。无线功能则要配合对应的硬件型号。 raspberrypi/pico-sdk t.co

Raspberry Pi Zero 3 Wに期待♪ メモリー4G欲しいけど良くて1Gまでな予感…メモリー高騰が悔やまれるね |･ω･`)とりあえず出たら買う！ linux#7711

Woke up to a broken MAC (wtf did the menu bar go of all things?) AND a broken RaspberryPi camera in need of updates libcamera#364

Raspberry Pi 5の有線LANで時々通信が欠落。EEE（省電力Ethernet）をOFFにしたら改善し、翌朝も欠落なしでした。kernel 6.18.50での話。同様の報告もあるので、心当たりのある方は確認候補に。 linux#7511

Raspberry Zero 3 Wのデバイスツリーファイルから読み取れることの情報をかき集めた。 github.com/raspberrypi/linux/blob/rpi-6.18.y/arch… SoCはBCM2711でPi 4と同じもの(クロックは落とされるかも)。電源ボタンが付きそう。PCIe,SPI Flashの対応。eMMCの可能性。ぐらいが気になるところ。USB-Cかは読み取れない。

Raspberry Pi Zero 3W Specs Leak: Pi 4 Chip, Release Date and Why It Won't Be $15 The Raspberry Pi Zero 3W now has a device tree, and it says the board runs on the BCM2711, the same quad Cortex-A72 chip as the Pi 4. I don't see a way it costs $15 like the Zero 2W, and the reason is the LPDDR4 that chip needs, which Raspberry Pi says has gone up sevenfold in a year. I went through the new .dts files to see what they actually confirm about Zero 3W specs and what's still guesswork (clock, RAM, price, release date). I also pulled published Pi 4 benchmarks to get a rough idea of how fast it could run a small LLM, and whether you should just buy a Zero 2W now. Where the Raspberry Pi Zero 3W leak came from Phoronix covered it too, but bret.dk wrote the detailed breakdown on 27 October 2026. Two device tree source files written by Marta Momotko were merged into Raspberry Pi's own kernel tree, the rpi-6.18.y branch of raspberrypi/linux, through PR #7711 on 25 October 2026. They aren't in mainline Linux, and Raspberry Pi hasn't announced either board. The links below point at the original commits on the martamomotko/linux fork. - Zero 3W: bcm2711-rpi-zero-3-w.dts, commit 1afad193 - Pi 4 Model A: bcm2711-rpi-4-a.dts, commit c17dd5f6 The Zero 3W file is very direct about what it describes: compatible = "raspberrypi,model-zero-3-w", "brcm,bcm2711"; model = "Raspberry Pi Zero 3 W"; Zero 3W specs: confirmed vs. guesswork A device tree describes how the board is wired. It doesn't cover clocks, RAM or price, so I've split what's in the file from what isn't. | | Zero 2W (shipping) | Zero 3W (device tree) | Status | |---|---|---|---| | SoC | RP3A0, 4x Cortex-A53 @ 1GHz | BCM2711, 4x Cortex-A72 | In the DTS | | Clock speed | 1GHz | ? | Not in DTS | | RAM | 512MB LPDDR2 | ? | Device trees don't describe RAM | | Wi-Fi/BT | Yes | Yes, includes bcm283x-rpi-wifi-bt.dtsi | In the DTS | | Storage | microSD | EMMC2 with bus-width = <8>, comment says "eMMC/SD card" | In the DTS, final form unknown | | Bootloader | On SD | SPI flash lines, so an EEPROM bootloader like the Pi 4 | In the DTS | | HDMI | 1x mini | 1x (second controller disabled) | In the DTS | | PCIe | No | &pcie0 enabled | In the DTS; whether it's exposed is unknown | | Camera | 2-lane CSI | 4-lane CSI config | In the DTS | | Power button | No | gpio-keys on GPIO 46 → KEY_POWER | In the DTS | | Price | $15 | ? | Pure speculation | The 8-bit bus width caught my eye. microSD only uses 4 data lines, so an 8-bit bus points to eMMC. Bret thinks it's a leftover and expects microSD on the final board. That could be right, but I'd keep an eye out for an eMMC SKU, because a Zero with soldered storage would make sense for industrial buyers. The Pi 4 Model A file is the boring sibling. It has a disabled Ethernet controller, one HDMI, no analogue audio or composite, microSD only, and antenna-switching lines (ANT_CTRL1/ANT_CTRL2, overlay params ant1, ant2, noant) that look a lot like the CM4's internal/external antenna option. Raspberry Pi Zero 3W release date No date has been announced. A device tree merged into Raspberry Pi's kernel means bring-up work is happening, but that doesn't tell you when boards will ship. Anyone quoting a month is guessing. The Zero 2W launched about five years ago and the last Model A was the 3A+ in 2018, so a refresh is overdue. The $15 Zero 3W problem: LPDDR4 prices The Zero 2W still costs $15 and hasn't gone up during the RAM crunch. Raspberry Pi's own April 2026 pricing post explains why. The LPDDR2 used in the Zero, Zero 2W and Pi 3 comes from stock the company already holds. The LPDDR4 used in the Pi 4 and Pi 5 has seen "a seven-fold increase over the last year". That same post put another $25 on the Pi 4 4GB, $50 on the 8GB, and launched a 3GB Pi 4 at $83.75. The BCM2711 uses LPDDR4, so the Zero 3W lands on the memory whose price went up sevenfold instead of the LPDDR2 Raspberry Pi already has in a warehouse. I'd bet on a 512MB or 1GB base model at a price that makes people angry in the comments. On the HN thread, one commenter points out that an Orange Pi Zero used to cost less than $10 and now sells for a tad over $30. Can the Zero 3W run an LLM? Pi 4 numbers are the ceiling There are no Zero 3W benchmarks because there's no Zero 3W yet. But it uses the same SoC as the Pi 4, and a small, passively cooled board almost certainly won't beat a Pi 4. So published Pi 4 numbers are a reasonable upper bound. The most useful dataset I found is SimeonOnSecurity's Ollama runs on a 2GB Pi 4B. The runs used five prompts, a 140-token output cap, and cold loads: | Model (Pi 4, 2GB) | Generation tok/s | Notes | |---|---|---| | LFM2.5-230M | 12.8 | 383MB peak RSS | | LFM2.5-350M | 8.6 | 516MB peak RSS | | qwen2.5:0.5b | 8.0 | 5/5 prompts | | MiniCPM5-1B (no reasoning) | 5.9 | 883MB peak RSS | | granite4:350m | 4.6 | stable, no swap | | gemma3:1b | 4.0 | heavy swapping, board rebooted | | qwen3.5:0.8b | n/a | didn't finish, board rebooted | For comparison, the Adafruit guide measured gemma3:1b at 9.99 tok/s on an 8GB Pi 5, and SimeonOnSecurity got 11.2 tok/s on its 8GB Pi 5. That's the same model that swapped a 2GB Pi 4 into a reboot at 4.0 tok/s. On the Zero 2W, the pi-llm project got llama.cpp working with SmolLM2-135M at Q4_K_M. It managed 9.5 to 10.9 tok/s generation, needed a mandatory 2GB swapfile, and took about 24 hours to compile on the board. The README warns that larger models may not run at all. What I take from those numbers: 1. RAM matters more than the A72 cores. Gemma 3 1B was already swapping and rebooting a 2GB Pi 4. If the Zero 3W ships with 512MB or 1GB, 1B-class models will be painful or won't load. 2. Models from 230M to 500M are where it'll work. LFM2.5-230M at 12.8 tok/s, or LFM2.5-350M and Qwen2.5-0.5B at around 8 tok/s, on Pi 4 silicon is fine for intent parsing, short summaries or a voice-assistant command layer. I wouldn't chat with it. 3. Over the Zero 2W, you mostly gain model size. The Zero 2W gets about 10 tok/s from a 135M model; Pi 4 silicon with enough RAM should get similar speed from a model about 3x bigger. That's worth having, since the pi-llm README admits its 135M model's responses "can be repetitive or incoherent". If you want a 1B model on a Pi today, buy a Pi 5. Boring answer, but that's what the numbers say. Buy a Zero 2W now or wait for the Zero 3W? My take: - Buy the Zero 2W now if your project is GPIO, sensors, a Pi-hole, a USB gadget, or anything that already runs fine on it. It costs $15, it's in stock, and the price is protected by LPDDR2 stock. A Zero 3W probably won't help a temperature logger. - Wait if you're hitting the 512MB wall, need PCIe or 4-lane CSI, or want a small on-device model. Expect to pay more. - Skip both if you want local LLMs. Get a Pi 5 with at least 4GB. HN commenters also bring up Rockchip boards like the Radxa Zero 3W, though one owner there says its Cortex-A55 cores are only slightly better than the Zero 2W's A53s. If you already have a Zero 2W and want to try a tiny model, don't compile on the board. Build on a faster ARM64 machine or cross-compile, copy the binary over, and add swap first. On a Bookworm-based Raspberry Pi OS image that still uses dphys-swapfile: sudo dphys-swapfile swapoff sudo sed -i 's/^CONF_SWAPSIZE=.*/CONF_SWAPSIZE=2048/' /etc/dphys-swapfile sudo dphys-swapfile setup && sudo dphys-swapfile swapon free -h Then run a model in the SmolLM2-135M class with -t 4, and watch vcgencmd get_throttled while it generates. On a Zero, thermals will start to matter before model quality does. So what's it for? I like the idea of a Pi 4 chip in a Zero footprint, and it's overdue. With LPDDR4 at current prices, though, I doubt it'll be the cheap hacking board people on HN want. My guess is a small, pricier board for camera rigs, PCIe experiments and edge projects that have outgrown 512MB. What would you build with a Zero 3W that the Zero 2W was too slow for? And at what price do you skip it for a Pi 5 or an ESP32? Tell me in the comments. I especially want to hear from people running models on a Zero 2W today. Sources - bret.dk: Raspberry Pi Zero 3 W and Pi 4 Model A on the way? - Zero 3W device tree, commit 1afad193 - Pi 4 Model A device tree, commit c17dd5f6 - raspberrypi/linux PR #7711 - Hacker News discussion - Raspberry Pi: a new 3GB Raspberry Pi 4 for $83.75, and more memory-driven price increases - SimeonOnSecurity: AI models on Raspberry Pi 4 and 5 - Adafruit: Local LLMs on Raspberry Pi - ravijo/pi-llm: llama.cpp on Pi Zero 2W

Sources: the merged pull request linux#7711 and The Register's report: theregister.com #RaspberryPi

Raspberry Zero 3 Wが来るのか。あと4 Model Aも。 linux#7711 Zero 2 Wが買えるのを待ってる状態だったんだけど、様子見した方がいいんだろうか。

Pi 5 RAM Swaps Have Been Firmware-Locked Since 2024. Here's What That Actually Means Raspberry Pi enthusiasts have been quietly upgrading the RAM on Pi 5 boards for a while now. The process involved swapping out the LPDDR4X memory package, and for the technically adventurous, it worked. It stopped working two years ago. The restriction is not new: Pi 5 bootloader EEPROM releases dated 2024-09-23 and later refuse to bring up memory that does not match what the board was provisioned with at the factory, and a board with a swapped chip fails SDRAM initialisation and does not boot. What is new is that the behaviour has finally been documented in public. One naming note: this is Raspberry Pi Ltd, the trading company, not the Raspberry Pi Foundation. Jeff Geerling wrote it up on 21 September 2026, which is why it is being discussed now. His own summary of the timing: "The firmware update that disabled RAM swaps was actually from late 2024, two years ago." Let's break down what the firmware checks, what the changelog does and does not say about it, and what your options actually look like. 💡 Why This Matters Raspberry Pi lists the Pi 5 in 1 GB, 2 GB, 4 GB, 8 GB and 16 GB variants, all LPDDR4X-4267. For users who wanted more RAM than their board shipped with, unofficial chip swaps were an appealing path. It kept older boards in service and gave power users more headroom for workloads like running local AI models, containers, or memory-intensive self-hosted apps. It was also, according to Raspberry Pi, being abused at the other end of the market: the company's position is that resellers were buying cheap low-RAM boards, soldering larger chips of unknown provenance onto them, and selling them as factory 8 GB units. Firmware-level lockdowns change that calculus entirely. When the bootloader refuses to initialise memory that does not match the board's factory provisioning, the board does not boot at all. The reported symptom is the green ACT LED flashing eight times in a repeating pattern, which is the SDRAM failure code. This is not a soft restriction you can work around with a config file tweak. 🔧 What the Firmware Actually Does Modern SBCs rely on their firmware to train and initialize RAM at boot. The firmware contains timing parameters, voltage settings, and compatibility tables specific to the memory chips the manufacturer has validated. When Raspberry Pi updates that firmware to check against an approved list of memory configurations, any chip not on that list can fail initialization entirely or be flagged and rejected. The practical result is that a board that previously booted fine after a RAM swap stops booting. The restriction lives below the OS, so there is no Linux-side fix. One important caveat on sourcing: the rpi-eeprom changelog entry for 2024-09-23 says only "SDRAM performance tuning", "Allow BANKLOW to be configured by SDRAM_BANKLOW parameter" and "Manufacturing test updates". It never mentions a memory check, so the changelog will not tell you this happened. The clearest primary evidence is rpi-eeprom issue 761, filed 15 October 2025 by a user whose Pi 5, sold as 8 GB, failed to boot on every EEPROM release from 2024-09-23 onward and booted fine on 2024-09-10. Raspberry Pi engineers asked for vcgencmd otp_dump, concluded from the OTP provisioning that the board had been modified after manufacture, and said such a board "won't cope with any changes in firmware / otp provisioning and is stuck on a pre-2712D0 firmware release with old/test SDRAM code." A Raspberry Pi engineer has also stated the policy directly on the company forums, in this thread. 🤔 Why Would Raspberry Pi Do This? There are a few plausible reasons, and they are not mutually exclusive. - Stability and support: Unofficial RAM swaps can cause hard-to-diagnose instability. Restricting unsupported configurations reduces support burden. - Supply chain and binning: Memory chips are binned and validated for specific performance envelopes. A chip that works in one context may not meet the timing tolerances Raspberry Pi has tested against. - Revenue protection: Selling higher-RAM SKUs is a meaningful part of the business model. Unofficial upgrades cut into that, particularly now that the Pi 5 is widely available. - Per-part memory tuning: The rpi-eeprom changelog is full of entries adding timings for specific memory configurations, including "Improved SDRAM refresh timings for Pi5 16GB" and a Broadcom SDRAM firmware bump "to support additional memory variants." The bootloader has to know which part is actually fitted. That also explains the part people find hardest to accept: a same-capacity chip pulled from another Pi 5 can still be rejected, because the check is not only about size. This has nothing to do with RP1, the Pi 5's I/O controller, which does not sit on the memory interface at all. None of these reasons are inherently sinister, but the combined effect is a more closed platform than many in the community expected from Raspberry Pi. 🛡️ What This Means for Your Homelab If you are running a Pi 5 with an unofficial RAM upgrade and it is working today, the immediate risk is a firmware update breaking your setup. Here is how to think about it. If your swapped board works today, you are already running old firmware. The only bootloader releases that tolerate a swapped chip are 2024-09-10 and earlier, so a board still running a RAM mod is two years behind on bootloader fixes. There is no release channel that keeps you both current and modded. The real choice is old firmware or a factory memory configuration, and it is worth knowing which one you have picked. To check your current EEPROM version and what update is staged: sudo rpi-eeprom-update The release channel is not set with rpi-eeprom-config. That command edits the bootloader's own configuration, things like BOOT_ORDER. The channel is the FIRMWARE_RELEASE_STATUS variable in /etc/default/rpi-eeprom-update: sudo nano /etc/default/rpi-eeprom-update In the current rpi-eeprom package that variable defaults to default, and latest is the other channel; the older critical, stable and beta names are historical and no longer what the tooling uses. The same file also holds RPI_EEPROM_ALLOW_DOWNGRADE and STRICT_MIN_VER_CHECK, which matter if you ever need to go backwards. Checked against the rpi-eeprom repository on 29 September 2026. Going backwards is getting harder. The 2026-09-23 bootloader release added a manufacture-version check: "If the MFG version of the image to update is less than MFG version in OTP then recovery.bin and self update will both fail to write the eeprom update", with bootloader_allow_mfg_downgrade=1 in config.txt as the documented bypass. On newer boards whose minimum version is burned into OTP, rolling all the way back to a 2024 bootloader is no longer a simple reflash. That change landed six days before this post went up, so treat it as current rather than settled. Consider your upgrade path instead. If you need more RAM for your workload, the official 16 GB Pi 5 is a real option. It costs more than a DIY swap, but it comes with full firmware support and no risk of a future update refusing to initialise your memory. 🧰 Alternatives If You Need More Headroom If the Pi 5 RAM ceiling is genuinely limiting your homelab, here are paths that do not depend on unofficial hardware modifications. - Raspberry Pi 5 16 GB: The official high-RAM option. Full firmware support, no risk from future updates. - Swap space on fast storage: Not a substitute for real RAM, but an NVMe-backed swap partition on a Pi 5 with an M.2 HAT can extend effective memory for workloads that tolerate it. Add a swap file on your NVMe mount point rather than the SD card. - Move the workload: For memory-intensive tasks like running local LLMs, a small x86 mini PC or a used workstation server often makes more sense than pushing a Pi to its limits. - Orange Pi, Rock Pi, or other SBCs: The broader ARM SBC ecosystem has boards with more RAM and less restrictive firmware policies. They come with their own trade-offs around community support and software compatibility. 🔍 The Bigger Picture This situation highlights a tension that has been building in the SBC space. As these boards become more capable and more commercially successful, the platforms inevitably become more controlled. The original Pi was almost impossibly open at the hardware level. The Pi 5 is a far more sophisticated piece of hardware, and sophistication tends to come with tighter firmware control. That does not make the Pi 5 a bad board. It is still one of the most capable and well-supported SBCs you can buy. But it does mean that the "buy a cheap board and hack it into something bigger" model has real limits now. Plan your Pi 5 deployments around the official RAM tiers, and treat any unofficial hardware modification as something that may stop working after a firmware update. ✅ Wrap Up Raspberry Pi's firmware-level RAM lockdown on the Pi 5 closed the door on a useful, if unofficial, upgrade path, and it did so in September 2024. If you have a working unofficial RAM swap today, you are on a two-year-old bootloader and any update will end that. If you are planning a new Pi 5 build that needs more than 8 GB, buy the 16 GB model and skip the uncertainty entirely. Jeff Geerling's 21 September 2026 post is the place to start on the community side. For the primary material, rpi-eeprom issue 761 is where Raspberry Pi engineers describe the check themselves, and the Pi 5 bootloader release notes are worth watching for future changes, with the caveat that this particular behaviour never appeared in them. --- Related from the lab - Why You Shouldn't Power a Raspberry Pi Through Its GPIO Pins - Pironman 5 Max Case Setup - Self-Host Apache Guacamole on a Raspberry Pi 5: Browser Remote Access for Your Homelab --- Originally published at codewithromi.com.