l1k

I'm just a kernel dev, not a product manager, but I think a sub $100 product is very unlikely I'm afraid. There are volume discounts though for customers ordering large quantities. Otherwise Accujack's answer is spot on.

We use the PREEMPT_RT patch set available at:

https://git.kernel.org/pub/scm/linux/kernel/git/rt/linux-sta...

That patch set is slowly being upstreamed into the mainline kernel. We regularly encounter incompatibilities with drivers for the Raspberry Pi SoC and upstream our fixes and feature work as well, e.g.:

https://git.kernel.org/linus/f7da7782aba9

https://git.kernel.org/linus/3c7b30f704b6

I must admit I wasn't aware of LinuxCNC, thanks for the pointer. This would seem to be a really nice use case for Revolution Pi.

We use the Raspberry Pi Compute Module, not the regular Raspberry Pi. The Compute Module is basically a SO-DIMM slot with the BCM2835 or BCM2837 SoC plus eMMC and LP-DDR2 RAM. The Compute Module is spec'ed at -25 to 80°C, versus 0 to 50°C on the regular Raspberry Pi.

We did extensive testing in the climate chamber and felt that going down to -40°C is safe.

The BCM2837 normally runs at 1200 MHz and the firmware starts downclocking when the core temperature exceeds 80°C. Once it reaches 85°C, the frequency is reduced to 600 MHz. It will be further reduced to 300 MHz when going beyond that. In our testing we found that the CPU is eventually halted when it becomes too hot, but can be rebooted without any issues once it has cooled down.

If you know that ambient temperature regularly exceeds 55°C, it may be necessary to install cooling in the cabinet together with the RevPi.

Since you're mentioning street cabinets, our customer Nicolai Buchwitz of Enda KG regularly installs our products in street cabinets for monitoring of hydrogen fuelling stations:

https://twitter.com/NicolaiBuchwitz/status/11127278200427274...

https://youtu.be/4iaInDiEiQg?t=338

True. However these modular gateways are ODVA-certified and have gone through all the mandatory testing (surge/burst, shaker, climate chamber), so they might be the preferred solution in environments where certified devices are required, be it only for compliance reasons.

There's a kernel module called piControl (source code is on GitHub) which periodically polls the I/O modules. The values written to the outputs are taken from a memory area called the "process image". Values read from inputs are stored in that same memory area. User space accesses the memory area via regular open/read/write/close system calls. Plus there's a bunch of ioctls to perform special actions such as firmware updates on the I/O modules.

You may use whatever language you prefer. There's a Python module called RevPiModIO, authored by a member of the Revolution Pi community, Sven Sager: https://revpimodio.org/en/homepage/

A NodeRED module was contributed by another customer, Erminas: https://flows.nodered.org/node/node-red-contrib-revpi-nodes

If you're into Structured Text, there's a software called logi.RTS included in our image (needs a license for unlimited use, 1 hour is free).

There's a Revolution Pi product coming this year which has WiFi built in (called "Connect Flat", was on display at the SPS fair in November 2019).

We used a chip based on Cypress CYW43455, basically the same as on the Raspberry Pi 3B+, to benefit from the existing decent driver support.

The Raspberry Pi Compute Module built into the Revolution Pi products has an eMMC of 4, 8, 16 or 32 GByte. Each product has a micro USB receptacle via which it can be flashed or mounted from an attached computer. Users are free to use whatever image they want.

Not anymore, no. We use the "Raspbian with desktop" image as a basis and delete most of the games to make the image fit on the 4 GByte eMMC variants of our products (with about 1 GByte to spare).

Customers who don't need a GUI may create a custom Revolution Pi image based on "Raspbian Lite" using our imagebakery scripts: https://github.com/RevolutionPi/imagebakery

Hi, I'm a kernel dev on the Revolution Pi team. AMA.

FWIW, having a TODO file is mandatory for anything in staging/, see Documentation/process/2.Process.rst.

https://www.kernel.org/doc/html/latest/process/2.Process.htm...

Moreover TODOs are added for unmaintained drivers outside of staging/ in order to prepare for their relegation to staging, e.g. see commit a0d58937404f ("PCI: hotplug: Document TODOs").

https://git.kernel.org/linus/a0d58937404f

Thanks for this blast from the past. I bought the book in 1997 as a sophomore student at university. My favorite book at the time. Leafing through the PDF after all these years I came across some gems I hadn't appreciated back then:

On page 161 the author mentions in passing that Eric Schmidt had been working on BerkNet. THE Eric Schmidt. The Google Eric Schmidt.

Earlier the book mentions that Bill Joy was able to hire a developer with money they got from DARPA. That developer was Michael Toy and later Sam Leffler. There's a movie called Code Rush about the open-sourcing of Netscape's source code in 1998 (which begat Mozilla and Firefox) and Michael Toy was the project lead for the open source effort back then. Sam Leffler later developed TIFF.

Zuffenhausen is now also a mixed ICE/EV factory and EV production has already started there last month. (Porsche is no longer an independent company but merely a VW brand.)

The screen on the frontpage photo says "armv7l". I think that's only printed on the Raspberry Pi 4. The specs say it's a Compute Module 3+, which I'd expect to print "armv7".

So either the photo is not genuine or the pictured device is using a Raspberry Pi 4 (possibly a pre-release Compute Module thereof).

Read the full interview. Netscape was owned by AOL back then and the motivation for Gecko was to have an HTML engine that could be embedded in AOL's software. Firefox was just one browser based on that engine, but became the primary focus once its popularity rose.

Camino was another browser based on Gecko, but died in 2011 when Mozilla killed off support for embedding: http://caminobrowser.org/blog/#mozembedding

It's also worth mentioning that the Firefox team did the same kind of "oopses" to browsers embedding Gecko. See this interview with Mike Pinkerton, who later went to Google to work on the Chrome Mac team:

"People will make changes to do something good for Firefox and because Firefox is the only really blessed project by the Mozilla Corporation and the Foundation, that’s their only focus. So they’ll make changes that work great in Firefox and then they’ll do—they’ll either like, break Camino’s build just entirely, or things will stop working, or things will slow down, and nobody will really understand why. Because, you know, there was no communication about, you know, this change might have this effect. There’s also a lot of strife back and forth between features that are implemented in the core Gecko in a way that they will only work with Firefox. [...] So we’d run into situations where we try and implement a feature and we discover we just can’t do it. And we kind of raise our hands and say, “Ah, can we get this fixed?” And the answer would invariably be, “Well, it works in Firefox. Who cares?” [...] And that’s something that eventually made me sour greatly on the Foundation and the Corporation and why we kind of took our ball and went elsewhere, and stopped trying to work directly with the Foundation for the majority of problems."

http://mozillamemory.org/detailview.php?id=7277

Only 232 patches were submitted under the Facebook e-mail address (since 2015). He worked at Fusion I/O and Intel before that, encompassing a total of 576 commits:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/lin...

Additionally there were 47 commits with his involvement in the pre-git era (starting Sep 2003):

https://git.kernel.org/pub/scm/linux/kernel/git/tglx/history...

Sad to see a prolific longtime contributor pass away.

I'm wondering if this is the technology enabling Apple's recycled Aluminium products.

The airjet sorter with X-ray detection at 5:27 is particularly crazy.

The Raspberry Pi Compute Module 3 is just that, the SoC of the RPi3 with 1 GByte memory and 4 GByte eMMC on a SO-DIMM module. You need a baseboard though to plug it in. Either the Eval Board offered by the Foundation, or a Revolution Pi if you aim for industrial use, or some other baseboard.

The desktop in this screenshot is localized to German, hence "Konsole" instead of "console". Also, "Okt" instead of "Oct" as calendar month.

A Thunderbolt controller contains a Converged I/O switch (CIO) plus a PCIe switch. The Downstream Ports on the PCIe switch are attached to the CIO switch and represent so-called PCIe adapters which encapsulate PCIe datagrams into Thunderbolt datagrams.

The PCIe throughput is thus the minimum of the CIO switch's capacity, the PCIe switch's capacity and the PCIe adapter's capacity. The PCIe 4 lane interface is only how the Upstream Port on the PCIe switch is attached to the CPU or PCH.

It is very unlikely that Intel will answer this question satisfactorily because they'd reveal details of the controller's internal architecture.

Note that a PCIe tunnel may pass through a Thunderbolt controller without being terminated on it. The CIO switch must then handle that tunnel's traffic in addition to any tunnels terminated on it.

(Disclosure, I've authored a couple of commits in the Linux kernel's Thunderbolt driver.)

The Total Perspective Vortex is allegedly the most horrible torture device to which a sentient being can be subjected.

When you are put into the Vortex you are given just one momentary glimpse of the entire unimaginable infinity of creation, and somewhere in it there's a tiny little speck, a microscopic dot on a microscopic dot, which says, "You are here."

https://en.wikipedia.org/wiki/Technology_in_The_Hitchhiker%2...

--user-agent "" --no-mtime

Ok, done.

Their competitor, Libre Computer, has mused that given their investment in the graphics stack, they might be headed in the GPGPU / AI direction. Think Raspberry Pis that you can cluster together for deep learning, or individual boards that kids can tinker with to learn about AI. The blog post also predicted the 3B+ fairly exactly:

https://libre.computer/2017/12/06/raspberry-pi-3x-raspberry-...

Money quote:

> We expect the Raspberry Pi 4 to be based on a new Broadcom SoC based on Cortex-A55 built on < 28nm fabrication process. We can call this BCM2839 and it will feature the much alluded to VC5 GPU and a new form factor. Since that hardware is more than two years away, it is difficult to predict the direction of the industry and thus the exact specifications. We expect such a Raspberry Pi 4 board to utilize dual 32-bit LPDDR3 or LPDDR4 to cope with the bandwidth requirements of OpenCL and neural network acceleration. The exact release date for such a specimen can vary dramatically depending on Broadcom’s tapeout schedule. It should be relatively soon if the Raspberry Pi Foundation expects to release hardware in 2019.

It's worse, Raspbian is compiled for armv6 to also support the BCM2835 (Raspberry Pi 1). Here's a performance comparison between a modern 64-bit armv8 Linux on a Le Potato SBC versus armv6 Raspbian on the Raspberry Pi 3:

https://libre.computer/2017/12/03/performance-and-power-cons...

Please note it's "Apple Inc." without a comma.

Updates to Raspbian packages can be installed via apt-get as usual. The same goes for packages specific to the Revolution Pi, those are made available through a separate apt repo: https://packages.revolutionpi.de/dists/

Upgrades to new releases can be performed either with an in-place "apt-get dist-upgrade" as is customary on Debian-derived distributions, or by creating a backup image of the eMMC, flashing the new image, then copying over any customized config files from the backup.

I'm not sure if the question relates to motion control of robots, human-robot collaboration safety or something else entirely.

If the question is about motion control of robots, our cycle time of 10-20 ms is probably too long as motion control usually requires < 1 ms.

If on the other hand the question is about human-robot collaboration safety, the product is likely unsuitable as this requires specially certified appliances.

If it's none of these two, it would be necessary to know more specifically what the setup is (fieldbus used, number of axes, etc) to come up with a realistic figure.

An AIO module is coming next month, but I'm not sure about the exact specs. I've just called the project manager, he promised to come here in a bit and answer open questions such as this one.

Anything that boots on a Raspberry Pi should also work on the Revolution Pi, e.g. FreeBSD:

https://wiki.freebsd.org/FreeBSD/arm/Raspberry%20Pi

It's just that we haven't tested it yet ourselves. Minimal adjustments may be necessary due to a few small differences between the standard RasPi and the RasPi Compute Module used in the Revolution Pi, e.g. the Compute Module lacks the on-board BCM43438 WiFi that's present on the standard RasPi 3.