Linux Drops Support For Imaginary CPU Platform

Linux is known for having a wide range of hardware support, but that may even include devices that don't actually exist like the Carillo Ranch CPU platform from Intel that doesn't exist at all online
==========Support The Channel==========
► Patreon: brodierobertso...
► Paypal: brodierobertso...
► Liberapay: brodierobertso...
► Amazon USA: brodierobertso...
==========Resources==========
Kernel Update: lore.kernel.or...
Tungsten Graphics: tungstengraphi...
Montavista: www.mvista.com/
Toms Hardware: www.tomshardwa...
Phoronix Article: www.phoronix.c...
HP Docs: www.hp.com/hpi...
=========Video Platforms==========
🎥 Odysee: brodierobertso...
🎥 Podcast: techovertea.xy...
🎮 Gaming: brodierobertso...
==========Social Media==========
🎤 Discord: brodierobertso...
🎤 Matrix Space: brodierobertso...
🐦 Twitter: brodierobertso...
🌐 Mastodon: brodierobertso...
🖥️ GitHub: brodierobertso...
==========Credits==========
🎨 Channel Art:
Profile Picture:
/ supercozman_draws
#Linux #FOSS #OpenSource #CPU
🎵 Ending music
Track: Debris & Jonth - Game Time [NCS Release]
Music provided by NoCopyrightSounds.
Watch: • Debris & Jonth - Game ...
Free Download / Stream: ncs.io/GameTime
DISCLOSURE: Wherever possible I use referral links, which means if you click one of the links in this video or description and make a purchase I may receive a small commission or other compensation.

[@shApYT]

Brodie, please stop violating Tux in the thumbnail. We talked about this. It is not appropriate or nice to do that to Tux. Just because he does not verbally object to it doesn't make it okay.

[@tablettablete186]

😅😅😅😅😅

[@o_q]

LMAO

[@LoveLaughlardlivelaund]

#brodie_deserves_to_be_cancelled

[@Megalomaniakaal]

#NoKinkShaming

[@liforra]

#deArrow

[@Sollace]

Props to that one guy out there mainlining KDE Plasma on their 1.2GHz single-core intel printer CPU.

[@seedney]

actually - good os without bloatware, and it's still usable today - but no to plasma - it's bloatware ;)

[@GerardLementec]

​@@seedneyhow is plasma bloat

[@applecastaway4256]

@@GerardLementecit got a gui it’s bloat /s

[@shApYT]

I call it the granny test. If my grandma can use it, then it is clearly bloated.@@GerardLementec

[@GerardLementec]

@@applecastaway4256 there are people that actually think this which is really weird

[@teslainvestah5003]

Support for Carillo Ranch should be continued in case of sidewinding time travelers. If a time traveler from a timeline that diverged from ours just before the successful launch of carillo ranch in his timeline sidewinds into our timeline while carrying a Carillo Ranch Linux machine, he should be able to downgrade his kernel to the latest common ancestor of his kernel and our kernel and then upgrade his kernel to the latest version of our kernel to get the security patches that are relevant to the malware biosphere of our internet, without losing support for his own processor and stranding himself here because he can't contact his guild leader anymore. Even if you hate time travelers, you have to be concerned that his body might flash-rot a few years after his divorce from the voice of Miandir. That could take out a city, especially if he doesn't self-isolate because he doesn't _know_ he's divorced from the voice of Miandir, because without a computer to receive such a rare warning, why would a time traveler ever guess that they are the only time traveler on earth? Sean, are you out there in this comments section, researching Linux support for Carillo Ranch? It's me, David. GET OUT OF THE CITY! You are divorced! I have done almost all I can to prevent catastrophe. The Linux foundation, it seems, has not, although I never knew them to be anarchists. Is it possible that they don't know about sidewinder resource fuzzing at all? Usagi's second pillar of universal hospitality? They absolutely must have an ambassador singleton, nobody achieves this level of coherency without one. So who is in charge of hospitality?

[@HappyBeezerStudios]

We also should offer support for time travelers using their iAPX 432 and IA-64 laptops. After all the 8086 was just a rushed stopgap until the real successor to the 8008 and 8080 was out: the 8800 And a proper IA-32 successor was only possible thank's to engineers at HP and Merced launched in 1998 to immense success.

[@JamesTDG]

That's not time travel, that's multiverse travel. If they want an updated Linux device, they better find a way to phone gone with something functional in our universe

[@neoqwerty]

@@JamesTDG this is how you wind up accidentally inventing SHODAN I think

[@SeralyneYT]

I've found exactly two references to Carillo Ranch that isn't about the driver support being dropped. One from 2008 and one from 2013. One is an error log on the Gentoo forums that mentions a Carillo Ranch MCH Device not being found. The other is the same but on the Lenovo forums.

[@JessicaFEREM]

lol

[@Octahedran]

So either way the references still are about it not existing lmao

[@AidanMacgregor-Personal]

​@@Octahedranhaha fair!

[@daniellapain1576]

and it wouldn't run that damn laptop unless the driver was present. I have the feeling someone programming the drivers was using this driver as a workaround to bypass a security measure on the machine for linux. I ran into that issue on an emachine laptop as well. The processor was dual core on both laptops.

[@sierra991]

​@@JessicaFEREMwhy are you everywhere

[@jbird4478]

Tech people have such a warped view of hardware specs now. I remember a time when all of the world ran on lesser hardware. What's now discarded as a useless 1,2 ghz once ran the international banking system and the leading armies of the world.

[@xaviervanzyl9299]

not to be contrarian because I also somewhat agree, but I've heard a couple times about people bringing up how we first got to the moon with hardware as powerful as our current smartphones

[@birddaughter]

@@xaviervanzyl9299 *less* powerful. like they had 4 Kb of RAM.

[@toxicpsion]

@@xaviervanzyl9299 we got to the moon with hardware as powerful as a 1990's graphing calculator. the scale of current processing power is almost incomprehensible.

[@brettfo]

Try browsing the web on a low-teir laptop from 2008 for a month and then bemoan the warped perspective.

[@ukyoize]

​@@brettfo Browsing is warped and javascript is punishment for our sins.

[@DDracee]

this will probly clear things up a bit; vermillion range is the codename for the cpu, carillo ranch is the codename for the motherboard that implements the vermillion range chipset so... EP80579 goes into the LE80578

[@HappyBeezerStudios]

Who cares about 80586 when you can have 80579

[@dorukayhanwastaken]

2:39 A lot of things can get done with a 0.048 GHz ARM SoC that has less non-volatile storage than this thing has cache. 1.2 GHz is _plenty._

[@CptJistuce]

640 megahertz should be enough for anybody.

[@annieworroll4373]

The Z80, even in its original form, drop in compatible with classic hardware, still has wide deployment in embedded applications, and is still being manufactured. The CPU on my TS1000 dies, I just hop on Digikey, spend a few bucks, a couple days later it will be up and running again. In that particular system, you're looking at 3.5mhz or .0035ghz. There's a couple 6502 variants still being made for similar use cases, though I'm not sure if they're drop in compatible with the classic form the way currently available Z80s are. Most hobbyist microcontroller projects use an Atmega 328, another slow 8 bit chip(16mhz in the Arduino Uno)

[@helidrones]

I still own an Epson HX-20 driven by two Hitachi 6301 (a CMOS variant of 6800 from which the 6502 has been derived) running at 614kHz. Back then it was a very capable device, could run for 24 hrs on 4 C-sized NiCd batteries, had a built in dot matrix printer, a true notebook sized keyboard, an relay interface for a cash drawer, two UARTS, a barcode reader interface, swappable ROM cartridges and an optional tape drive. One could argue that it was the great-grandpa of modern notebooks.

[@damouze]

@@annieworroll4373 Believe it or not, but the MCU core on some PS/2 to USB converters are 6502 based.

[@annieworroll4373]

@@damouze I'm not surprised. I would doubt that would be all that complicated, since it doesn't need remotely the whole USB spec, so ancient cores could do it.

[@C1rnobyl]

Why install the printer in Linux, when you can install Linux right into the Printer? XD

[@HappyBeezerStudios]

Where is the Doom port? Performance should be decent.

[@knightmarSPZM]

It was a driver for audio devices handling audio on weddings on Leo Carrillo Ranch. No more weddings in linux world. I suppose it wasn't really used that often anyway.

[@Vicorcivius]

Nooooooooooooooooooooo!!!!!!!!!!!1111111111

[@FantomMisfit]

Well yeah getting married in this day and age as a Male is just a bad idea Marriage is a contract and any contract the rewards someone for breaking it is a bad contract

[@mx248]

​@@FantomMisfit So then just get a bespoke contract that's to your mutual satisfaction, yeah? 🤔

[@FantomMisfit]

@mx248 In my experience prenups aren't worth the paper they're printed on...its really down to the Judge's whim whether it'll be honored or not. Judge's can literally just decide nah and screw you anyway (and they do) idk what country you're from but in the US the marriage contract is a really bad deal for males and no guy with something to lose should ever agree to it. Edit: I'm not saying don't date or look for love but maybe just skip the whole getting married part

[@PadaV4]

@@mx248 High chance a judge would just throw it out because its not "fair" to the women. In this age any marriage contract is of dubious use.

[@superpou184]

intel generations 1-3, when the mobile cpus were marked with "M". instead of being soldered to the board, they were actually removable.

[@Cmdrbzrd]

Haswell (Core I gen 4) had M-marked mobile CPUs in sockets too, not to mention all the stuff prior.

[@klontjespap]

i shudder at teh thought of an embedded/baked-on nehalem chip :')

[@Padgriffin]

It depended on the implementation- soldered mobile CPUs weren’t uncommon (the i5-3320M in an ThinkPad X230 would be soldered while the same CPU in a ThinkPad T430 would be socketed, for example). IIRC Haswell was the last generation to have socketed options, with the most commonly found example being the ThinkPad T440p, but only for the -M and -MQs and not the ULV chips.

[@p3chv0gel22]

​@@Padgriffinit was the last Generation with official socketed CPUs You can get even modern CPUs socketed (but only through sketchy channels)

[@HappyBeezerStudios]

That chip is much older. 90nm and from 2008? That sounds like a Pentium M derivate.

[@Vatharian]

That's Intel's First SoC since 80386SX. Also that's Pentium M, so basically Pentium 3 with PCIe 1.0 root, embedded DDR2 memory controller and MCH glued UNDERNEATH the CPU. It had astonishingly high failure rate. I knew of this platform from marine gas cargo control systems but never thought it was in laptops.

[@HappyBeezerStudios]

Hmm, right, that was before the Quark X1000, that only came in 2013 Could be something stemming from Dothan (80536) or it's low power cousin Stealey, but with reduced cache. Those were produced in the right node, period, and clocks.

[@ChrispyNut]

Thank you for bringing this to my attention, Brodie. I will immediately get on to the Kernel team to let them know I need to this to remain as I often use my printer as my primary computer and as you may imagine, operating at 1 frame every 4 seconds (once the drum's warmed up) means I'm prone to clicking on the wrong button and with ads so often being malware and the CPU being too weak to run an ad-block, it's important I have the latest security patches.

[@ME0WMERE]

[@ChrispyNut]

@@ME0WMERE If only you were always around when I'm crafting code 😆

If you're lucky you may be able to squeeze out an extra frame of you install libreboot

[@FireStormOOO_]

You joke, but Ricoh for example ships their copiers with some Android variant, they include the stock browser, and I have found rare cause to use it for testing...

[@bryede]

Whew. I'm glad I'm not the only one. Do you guys chat somewhere?

[@odorikakeru]

Movidius is still a bit of a sore point. A project I was involved with at the time got scuppered partly because Intel bought Movidius and refused to sell or license their chips. (That wasn’t the only reason the project died, but it was one of the nails in the coffin)

[@gregor3099]

Would you mind expanding a little more?

[@odorikakeru]

@@gregor3099 I can’t really go into specifics because that wouldn’t be fair to the customer (even though I think the NDA has probably expired by now). It involved using a neural net for obstacle detection and indoor navigation. Fairly commonplace now, but it was exciting at the time.

[@obake6290]

Seen this a few times now. Intel is so good at upstreaming that they put in work on in-development products. Sometimes these never make it to market, and Intel then removes the drivers for them. I recall this happening with some AI driver or another a while back. Of course, usually they don't look at it like "What the hell is this?"

[@androidmirage335]

Laser printers can last a while, and may need at least a new LTS kernel for security updates to prevent various buffer overflow BadUSB and network exploits (there was a buffer overflow on the Linux driver for Intel Gigabit NIC just several weeks ago). Maybe that printer is old enough not to do USB Host or Ethernet. But nowadays, we need to perpetually update any hardware with new software, or replace it, or put it behind an FPGA stateless firewall between a USB to USB data transfer cable (hint: that third option isn't really much available to the public yet).

[@orbatos]

Or just use any router/routing software or a dedicated print server. A Pi3 would work fine.

[@mskiptr]

Well, if anyone decides to try and update any such system, the driver can be just brought back. It's not like it's being erased from the git history

[@brettfo]

It's fairly common to patch/maintain old kernels in legacy hw rather than upgrading the kernel.

[@orbatos]

@@brettfo of course this only goes so far. At a certain point mitigation may require patches that cripple performance to hack in backported security features or in some cases it may not be feasible with a given hardware configuration. In stand-alone situations the best next step is a gateway providing the missing abstractions and often containerization of whatever legacy endpoints were used. This isn't a new approach, and it's even been done with contemporary systems that are still being updated where security is paramount; big iron, telemetry, automated production, distributed clusters, etc. Sometimes even with sneakernet airgap, I worked with one of these in the form of a payroll printing system in a secure facility where the printer interface had 3 layers of emulation and still wasn't allowed on the local network.

[@brettfo]

@@orbatos if the customer is paying enough then anything is possible. I suspect when it comes to printers, the majority of customers would rather buy a new printer.

[@johnsavard7583]

And here I thought this video would be about Linux dropping support for System/380, the imaginary CPU that MVS/380 runs on (it's a trick to allow a slightly modified version of MVS 3.8 to run MVS/XA software).

[@Ezukah]

What about if we have Ram Ranch hardware. Do we need to send pics of that?

[@EwanMarshall]

Looked up support page for one of those printers, most recent firmware is July, and those printers are big photocopier type ones. HP might still need support, then again, they might not.

[@stevethepocket]

HP is probably thrilled that they have an excuse to drop support for some of their products.

[@BrodieRobertson]

I really doubt they're running anything close to modern

[@EwanMarshall]

@@stevethepocketConsidering the support of some of their new products... seriously, which can't they just have ACPI _DSD data in the UEFI firmware images for the amplifiers they are add between the DAC and the speakers in their laptops? Seriously this is an issue, ASUS has this problem to. There is long lists in the Linux kernel drivers for different models to turn the amplifiers on.

[@martenkahr3365]

@@BrodieRobertson Not to mention HP probably isn't interested in running Linux on one of their outdated printer CPUs. After all, removal of this driver doesn't mean the printers using this CPU suddenly stop having driver support, or that you can't write new firmware or drivers for the printer itself. It just means that the CPU in them can no longer be used to run the Linux kernel.

[@HappyBeezerStudios]

If they want support on their hardware, they could just revive HP-UX

[@marcelomafra]

Wow! That HP printer. No good memories from that one. A blast from the past form one the most expensive and problematic (not counting some deskjet) printers I had to work with and with an awful support in warranty, from HP.

[@AlpineTheHusky]

Intel Movidius does exist. It is a Vision Processing Unit that is added into select CPUs

[@alexanderstohr4198]

Its still a big achievement to realize a working driver, even if product management of the chip maker company finally decides that it wont see sufficient business cases to start production.

[@valenrn8657]

From Tomshardware _As far as we can tell, Carillo Ranch was supposed to be a platform/motherboard for EP80579 embedded CPUs, which combined a Pentium M core known as Tolapai with the Vermillion Range chipset and a memory controller._ Intel Carrillo Ranch Pentium M compatible Vermillion Range.

[@AceMcCrank]

Looks like the Intel EP80579 processor is used in some NAS and network devices. Axiomtek's NA-200A, and Netgear's ReadyNAS NVX both seem to use it.

[@computer_toucher]

So even more crap than the little Atom in the Ultra 2 duo from around the same time that I had? Amazing.

[@ABaumstumpf]

@@computer_toucher For a NAS they are more than fast enough.

[@QualityDoggo]

The beauty of open source: even if it's been abandoned publicly, if it's being used somewhere it's theoretically possible for someone to make changes and continue maintaining their software. There's also a large number of communities and businesses watching. This way, abandoned software can be removed without massive risks

[@JoeStuffzAlt]

The funny thing is that we'll find that one guy right after it's removed

[@chainq68k]

Wasn't the whole FOSS movement started by Stallman not being able to upgrade his printer or something? So, careful with removing FOSS support for printers... :)

[@LilithNephilim666]

maybe somewhere in a parallel universe, this cpu was actually used, and some random kernel pr got merged to the wrong universe, ours

[@klontjespap]

that same universe still shits the bed on rockchip 6 cores lol

[@NorthernChimp]

Git use in metaphysics

[@mnoxman]

*BSD is well known for having a branch that will compile on any platform except a Analog systems. Linux is not quite the same. The split between Kernel 2 and 3 dropped a lot of legacy drivers (ESDI, Token ring, etc).

[@Blue-Maned_Hawk]

2:14 I used a machine that ran at 0.9GHz well into the midst of the pandemic. 1.20 GHz _should_ be enough for many people.

[@Roxor128]

Favourite Intel CPU generation? 486. First PC we owned was a 486DX/33 with 8MB of RAM, so I've got lots of good memories of it. Also the point where x86 was first delivered in a "complete" form, without the need for an external FPU. Only other ones I've had were a second-hand Pentium MMX, a Pentium III 500MHz, and a Core i7 3930k. Everything else I've bought has been from AMD, due to being better value for money.

[@adamrak7560]

when you program an ESP32 with 4Mbyte RAM and dual core 240MHz, you realize that these old CPUs, like 486DX/100 (which I have had) were very very capable already.

[@justenkelley7158]

Like waiting hours for song to download from limewire LOL

[@Roxor128]

@@justenkelley7158 486 was before the days of MP3 files. You weren't going to waste several megabytes of disk on one song back then. Wasn't even fast enough to do the playback without breaking up, anyway. If you listened to music on your computer back then, it was MIDI or, if you wanted quality, tracker modules. I have lots of both that have been carried over through many machines since then.

[@kittenisageek]

@@Roxor128 My first PC was an 8088XT with 640kb of RAM and a 10MB hard drive. I remember playing the first "Warcraft" on a 80386 that I quickly upgraded to a 486. I'm pretty sure I had a few MP3s at that time. Hard drives by then were around 400mb and MP3s were amazing when compared to WAV because they took up 1/10th of the space. By the mid 90s, the first Pentiums were hitting the market, and I'm pretty sure that's when WinAMP and Napster started to grow in popularity. So, perhaps the 486 came out before MP3 was developed, but the 486 was still a popular platform at the time and it played MP3s just fine. MIDI and MOD (Tracker modules) were awesome. I could fit a dozen MODs on a floppy and they sounded great. I still have a folder with some MODs I collected back in the 80s off an Amiga. Those machines were contemporary with the 286, but absolutely blew them out of the water when it came to audio and graphics.

[@Radi0he4d1]

Sandy Bridge because that's what I'm still suing

[@Amaqse]

One question, u do realize literally all 'combo' printers and office printers use linux under the hood? Even 4 thousand dollar massive office devices like the Xerox ones use linux combined with some slow stable cpu like a pentium 3. Yes there are brand new office printers with pentium 3 era chips inside of them running linux because they are performant enough, and nowadays incredibly cheap. using platforms so old kindergarden childred draw schematics for as a leisure activity. Removing linux support for embeded chips means making the life of device manufacturers much harder.

[@goaserer]

It was not fully clear from the video but the drivers in question were not for the CPU itself but extended features of the chipset / CPU platform. One of the drivers would be specifically used to address integrated NOR Flash memory. The integrated Pentium M CPU in question most likely will still work as usual even after those changes. Since Intel, as well as most bigger hardware manufacturers, is part of the Linux Foundation it is somewhat unlikely that anything would be removed that still had any value for manufacturers

[@HappyBeezerStudios]

And I doubt any manufacturer of these devices still offers any support for it anyway.

[@GSBarlev]

When I clicked and when you said Intel, I thought this was going to be about Pathfinder, Intel's RISC-V project that only ever existed inside their "Simics" simulator (is that somehow distinct from an emulator?)

[@TurtleKwitty]

Emultaors emulate real things simulators simulate fake scenarios

[@TheExileFox]

An emulator generally doesn't mainly run synthetic tests

[@CptJistuce]

Emulator and simulator have very similar meanings, and which term is used is more down to stylistic preference.

[@Roxor128]

Generally, I'd say that "emulator" gets used for hardware that's been released, while "simulator" gets used for hardware that's under development.

[@NorthernChimp]

Half-way in your sentence, I thought oh * did they drop support for the Mars probe

[@janimakinen1455]

Interesting! Maybe my HP monster, the lovelily named MFP521dn, has the same Intel insideⓇ, as they mention a "800 MHz processor, 256 MB RAM" in the specs. I'm not planning to upgrade the kernel anytime soon, but I did the recommended software update a while ago, it was to fix a bug in the webserver iirc. :)

[@ChrisXPZ]

the Portwell CATO-3000 is some networking equipment that uses this processor too!

[@ai-spacedestructor]

all this talk about Carrilo Ranch makes me imagine someone preparing food on computer hardware thats openly laying on a table somwhere.

[@JohnVance]

Never would have guessed HP would have an x86 SBC instead of Arm.

[@HappyBeezerStudios]

That chip is from 2008, so they could've used a Cortex-M1 or Cortex-A9, or perhaps some ARMv6 solution. Which wouldn't even be out of place, Nintendo DS, original Raspberry Pi B+, the first generation nvidia TEgra, Qualcomm MSM7200, a buunch of late pre-Android phones.

[@ShadowDrakken]

2014 Lenovo Thinkcentre Edge 73, a secure mini tower may have also been connected to the Carillo Ranch mainboard

[@imtiredtoday]

Favorite intel cpu generation... Yorkfield Those things just didn't stop working like at all (yes I'm talking about core 2 quad, specifically a q8200)

[@HappyBeezerStudios]

My Harpertown smiles at me from under the desk. Which is just Yorkfield, but dualprocessor capable.

[@BenjaminWheeler0510]

Please bring this back. I use the eepy ceepyu (emulated on qemu of course) and this update breaks my flow. Now I can’t update 😿

[@evildragon1774]

Wtf?

[@Pentium100MHz]

"Nobody is going to run mainline Linux on a 32bit 1.2GHz CPU". Well, I do not have that specific CPU, but I like to use hardware that is not too much overkill for the task I want to use it for. For example, I use Debian 8 (later ones do not boot, it's either not enough RAM or because the CPU does not have some feature) on a 233Mz Cyrix MediaGX CPU with 128MB of RAM. I use one such PC to configure switches (it's small and has a serial port), run VPN (it manages a few mbps, but that is enough for me in that case) and use it as a router for testing/configuring some systems. It works great for all of that. I plan to use another one of these to get the data from various devices (power meter, solar inverter, generator) over modbus. I think 233MHz and 128MB would be enough for that too. I'll run Debian 8 on it, but I considered DOS and Windows NT - but I'll try Linux first because I can boot it over network. DOS would be good because I could use a small DiskOnChip2000 SSD, but I would need to do more stuff to program the interface compared to Linux and libmodbus.

[@autohmae]

Man, I guess we won't be upgrading the kernel on the old printer...

[@ABaumstumpf]

There are so many obscure CPUs and platforms out there - many of which you will not find any public information on. We had tested some 10nm Intel CPUs in 2017 (when intel still said they would give us the products in Q4 that year), we have some Epyc CPUs with different memory support and some that you can only find some obscure references from IBM, Oracle and Amazon. 0.9-1.2GHz singlecore low-power CPU in 2008? there might be many millions that were actually sold and are still running. For printers the kernel is likely no concern, but would not be surprising to see the code being revived in 1-2 years when a manufacturer notices he still needs those systems. People tend to forget that private Desktop-Systems mostly used for gaming are just a very small section of the CPU market. In early 2020 Intel already had about half their production-capacity in 10nm but only in late 2021 did they release desktop-CPUs on that node. Those facilities were up and running producing things like 20core CPUs for communication-infrastructure.

[@butre.]

these had tolapai SOCs. it's mainly in printers as you've found, routers, etc

[@namvet_13e]

It is hard to prove a thing doesn't exist.

[@numbertumbers529]

New Linux ARG dropped?? Is this some kind of SCP type nonsense???

[@nosotrosloslobosestamosreg4115]

SCP-Gentoo > Drains life joy from n00bs.

[@orbatos]

Shh😊

[@IIGrayfoxII]

Remember hearing that a CPU support was removed from the linux kernel, the CPU that was removed never got released. So I would think Intel added support for this CPU(either so it worked out of box when it got released or so they could test it in house) So it made sense for support for a never released CPU to be removed from the kerenel. Heck I am sure the linux kerenel can be heavily stream lined if more ancient hardware support was removed.

[@peterjohnson9438]

No need. The kernel is written in a modular manner, you can compile everything you need right into the kernel while leaving everything unnecessary out. The end result is very fast and efficient.

[@HappyBeezerStudios]

@@peterjohnson9438pretty much the best solution. Offer a full package with all the legacy support, but allow for slimming it to make a super streamlined individualised system.

[@TheObsesedAnimeFreaks]

it seems that the systems that would be using this driver are older systems that would not be running modern software to begin with, it also seems like if for whatever reason these systems would need to be updated it they will have a headache on their hand. from what some of the comments have said, it seems that it was used in a lot of critical operations equipment and other business applications. things that might not be updated or upgraded for a variety of reasons. someone also said that they have a high failure rate so it might also be a non issue over time because eventually all these chips will be gone. but it's not quite an "imaginary" chip, it's quite like some of those other extremely obscure electronics hardware that if you know what it is, you needed it. rather then, you could just happen across the chip or device randomly.

[@HappyBeezerStudios]

A 1.2 GHz 90 nm single core Intel CPU from 2008 makes me think it's some kind of Pentium M offshoot. Good old 686 IA-32

[@cmd.ada.]

hey, at least we got 64 bits in emdeded devices out of the ep80579

[@ruediix]

I bet, if anything, said chipset never left the lab. There are always a good number of such chipsets, but typically the patches don't get merged to support them, and are only kept in house, or at most on the company public github. Also, if it did leave the lab and found it's way into embedded devices, most embedded devices ship custom kernels. So, any manufacturer still supporting this hardware can revert the patch on their kernel.

[@supernovaw39]

I thought that this is going to be about Itanium-64 (IA64)

[@1DwtEaUn]

to be fair half of that blame for that one also goes to HP

[@MarcusHuang-nh9vc]

that's been a thing for a good few weeks now.

[@progste]

Where will I farm my carillos now???

[@Solution-Smith]

Are they going to end support for my teapot with a 418 processor?

[@katanasteel]

Favorite Intel ISA CPU... hmm... I'd say 2nd gen Opteron for its amd64 extensions and amd-v tech... best dual core CPU I've had

[@rfvtgbzhn]

May favourite Intel CPU Generation is the original Pentium (P95). It was really ahead of it's time.

[@nicholaswjamrock]

you are being a bit dramatic saying its unreleased hardware, alot of hardware still use that chip, you should know that many devices uses linux, i was looking at an ATM seval years ago and it had that chip, i have even seen some that are still in use with Pentium 3. It man not be a consumer product, that dont mean it dont exist, it mean its a niche product

[@ManuFortis]

Hmmm... seems youtube doesn't like the product name string. You want to look for the EP Eighty Five Seventy Eight, not Nine. It seems from looking at a product page and a recast page intel released, that perhaps it was canned for the new version with no lead in it. Perhaps.

[@DLexEdition]

The Intel EP80579 Integrated Processor and associated chipset was known as the ill-fated Carillo Ranch chipset. It was one of Intel's biggest flops since their foray into 32-bit a long time ago with the 80i432 CPU, the i432 is not compatible with x86_32 or x86_64 designs.

[@rogercruz1547]

18 naked circuits in the showers...

[@JohnGotts]

Any code removed from the kernel can be maintained as a patch. I maintained two kernel patches for many years. No big deal. If it needs to go back in, it can be done.

[@robertheinrich2994]

my guess: soon there will be an outcry from the NSA, because they want to keep their linux support for their hardware platform ;-)

[@marcogenovesi8570]

28 US Marines, pulling up in Ford Raptor Trucks

[@alphaomega154]

just to add, since intel is on topic, i notice that linux based on OLDER devices relying HEAVILY on intel drivers. which known to be UNIVERSAL. which many old AMD devices are USING in LINUX. if i never use linux, i will never know that. it kind of makes me respect the company more. despite many of their bad business decisions. and that practice apparently have been long before AMD make their "universal stuff like the FSR feature. inf act my 12 year old AMD device running on linux right now uses some of those INTEL drivers. and it runs great.

[@electricindigoball1244]

The thing to remember about Intel is that it's a massive company and this means that you can't treat it as a monolith. Even though the leadership was engaged in anti-competitive practices or making anti-consumer decisions that didn't prevent the Intel engineers from doing good work that benefits everyone. This applies to much smaller companies as well since at the end of the day every company is only as good/bad as the people working at it and leading it.

[@goaserer]

AMD only started building their own PC processor design 1996 and made it compatible to the already existing Intel x86 instruction set. Since Intel dominated the PC market at this point (and it took AMD a few years to catch up) it would make sense that most drivers were originally written or commissioned by Intel. I'm not sure whether Intel ever wanted to produce universal drivers - it's more likely AMD copied heavily to make sure all existing PC applications would run on their chips as well as on Intel ones. The outcome is nonetheless a positive one in the way that for the OS it makes little to no difference whether your machine runs AMD or Intel CPUs nowadays

[@HappyBeezerStudios]

I was so pissed when Intel cut down on their driver archive. Where should I get DOS drivers for that network card.

[@HappyBeezerStudios]

@@goaserer A bit more complicated than that. IBM wanted a second supplier, and AMD took the job. The first x86 they properly designed was their 486DX4 (because Intel kept that locked) and their 586, later named K5. The K6 was designed externally and is based on NexGen's Nx686 (successor of their Nx586 and Nx486) on an architechture called RISC86, because it just translates x86 instructions into RISC instructions. (Intel did the same with P6) And before they made x86 compatible chips, they had their own chips. Like the Am9080 in 1975, followed by Z8000 chips and their 29000 RISC chips

[@tech34756]

If you want some fun, try playing the driver lottery with old but once fairly common Bay Trail hardware. So far I've got a Linx 1010b running Mint fine/good enough ootb (after getting grub 32bit efi on it) but my Archos wont boot Mint (black screen) and Debian 32bit and Fedora lack touch screen support among other things.

[@justf0dd3r]

I was watching the video and wanted to say that you missed it, but you sped right by the part that said that "Carillo Ranch" was a motherboard not a chip. So that is what got you confused about that chip number.

[@atomicskull6405]

I wonder if it's some military contract thing that was never a commercial product?

[@ayrendraganas8686]

the ram ranch written on the whiteboard caught me off guard x)

[@obake6290]

Favorite Intel generation? Well, the only Intel CPU I ever owned was a Skylake, so I guess that wins by default.

[@Poldovico]

Could it be that Carillo Ranch was deleted from the archives?

[@BrodieRobertson]

Possibly but surely there would be a forum post

[@Poldovico]

@@BrodieRobertson Perhaps the archives are incomplete, is what I meant to say :P

[@imark7777777]

and It's for an HP printer so we really don't need it at all!

[@caparazo3488]

It's not imaginary, it's real. Real shit.

[@tylerdean980]

Still waiting on VAX support

[@SviatoslavDamaschin]

Next EAR you'll have it /badjoke

[@zzco]

The datasheet is 100MHz off from a Back to the Future reference. Dammit!

[@Roxor128]

10MHz, you mean. 1.21G-1.2G=10M

[@blackdew2]

90nm wasn't that shabby in 2008. That's like Pentium 4 era node size, somewhat dated by then but plenty fine for embedded applications.

[@mckendrick7672]

This kinda goes to show how ridiculous the monolithic model of the kernel is - drivers for non existent and outdated products present on your machine even though you would never feasibly use those products yourself. It's too bad that Linux is so entrenched that a microkernel probably won't ever take over.

[@CFWhitman]

The driver for this almost certainly won't be on your machine. Even if the module was compiled, it won't be loaded, so it won't take up any of your resources other than a few kilobytes of space (if it was compiled in at all). Most current operating systems use what they call a hybrid kernel, because microkernels are notoriously slow. So Windows and MacOS, for example, both use a hybrid kernel rather than a microkernel. Linux is a monolithic kernel that uses modules for flexibility. Most don't consider that enough to make it qualify as a hybrid kernel. Linus Torvalds, though, claims that "hybrid kernel" is mostly just a marketing term, and considers the Windows and MacOS kernels essentially monolithic. No popular operating system uses a true microkernel.

[@marcogenovesi8570]

this only goes to show how ignorant about Linux some people are. This driver is not built by default. A good 60% of Linux drivers are for random embedded stuff or specialized server stuff and not built at all for a PC, or if they are they are built as a module and loaded only if needed. Most embedded devices build only the drivers needed for their own SoC, 99.99% of the other drivers are disabled and not even built

[@getzybaggins]

Intel Carillo Ranch was a motherboard and platform for embedded EP80579 CPUs that were released in 2006. As of December 2023, Linux has stopped supporting the Intel Carillo Ranch motherboard because it doesn't seem to exist

[@getzybaggins]

The Intel EP80579 is a family of 32-bit x86 system on chips (SoC). It is designed for security, storage, and communications. The EP80579 has a Pentium M CPU with a clock speed of 600-1.2 GHz and a 256 KB cache. It also has the following specifications

[@getzybaggins]

simple google search brought up a lot of info

[@PeterBrand96]

My favourite Intel CPU generation? Bloomfield. Had an i7 920 long ago in the first pc I build myself.

[@bleack8701]

if there are products that use it and it was launched then it exists? I don't see the issue?

[@MatthewSuffidy]

If Linux dropped something it is REALLY small pickings. I used to use QIC-02 in the 90s. This sounds like a simple vga device or maybe video output modulator.

[@watercat1248]

I know that is probably very apchour protect But it's the reason to get the length to stop support i driver ?

[@DexieTheSheep]

LMFAO I JUST FOUND OUT ABOUT THAT SONG BECAUSE OF THIS VIDEO, wouldn't have even bothered looking it up if you didn't point it out

[@emperorarasaka]

By removing the Corillo driver, they’ve killed God. Long live the holy 1 Core 1 GHz chip!

[@qwertykeyboard5901]

NSA backdoor?

[@kienhwengtai8113]

I found that CPU on eBAY for $US 14.90 ... it does exist! It's a Pentium M. You have solder the thing to a board to use by the looks of it.

[@dokt9304]

and what if this was actually never released and this was an hidden backdoor driver?

[@Hyperboid]

i didn't see what you wrote on the board until you addressed it. i've won, but at what cost?

[@cem_kaya]

what is the drawback of having that driver in the kernel ?

[@legendboyAni]

Extra work for no reasion

[@cem_kaya]

@@legendboyAni what work would it require tho ? isnt it already there.

[@Raletia]

My guess is Maintenance and Bloat. If no one has access and understanding to test it, they can't fix any bugs or unexpected interactions. Also, every single driver would add extra size and complexity to the kernel, maybe small but it could add up.

[@goaserer]

I think around the discussion of dropping the i386 and i486 support one of the main arguments was that if no one of the kernel devs still have the option to test the specific environment no one can tell for sure if a supported infrastructure actually still works. And there is always the possibility that there are flags or options specific to this driver that will simplify maintenance and docs when removed for good. In the end each line of existing code can cause additional work as other parts of the infrastructure are updated, like compiler versions. Therefore if functionality has no more value it is a good practice to remove it to avoid situations where it might make unnecessary work in the future. Especially true if you consider that while it is dropped now from latest releases, the code will still have to be maintained for the next 6 years in LTS versions

[@jbird4478]

If it were up to me, I'd choose to drop support without removing it though. You never know. Someone might need it and it might just keep working for quite some time. You can always remove it later.

[@ironfist7789]

Great, now my dream computer has stopped working...

[@DarthWampa_]

"Linux is generally known to have an incredible range of hardware support" (queue laughtrack)

[@ArveEriksson]

Oh hey! You could probably play Wing Commander: Prophecy on that thing.

[@SprocketWatchclock]

Why is this news? Removing old crap is no big deal. No need to bloat the kernel with old stuff nobody uses anymore.

[@BrodieRobertson]

It's fun

[@mearetom]

It's nice to know what old stuff the kernel had.

[@capability-snob]

Because usually, "nobody" is like 20 rabid fanatics who use the hardware on the weekend. In this case, nobody _really is_ nobody.

[@ZipplyZane]

The fun part was tracking down exactly what this was, since almost no one had ever heard of it.

[@ziemlichundead]

Boy I hope my nvidia drivers would magically pop into existence

[@t01]

skylake 14nm was such a chad soc it lasted 7 years +++

[@crapmalls]

My wifes computer wants a pumpkin spice driver

[@JonitoFischer]

Twice the speed of the processor of the raspberry pi 1, more or less the same process node.

[@theepicslayer7sss101]

well in the Intel side, i must praise the Q6600, i was gaming on it still in 2017 and the only reason i stopped is that my mother board was eating my CMOS batteries in 6 months and my buddy gave me my current PC (i3 8100 with 32 GB or RAM and a 1050ti i replaced the GPU with a 3080 12 GB since then.) back on the Q6600 while it was idling at 70C it never blue screened or anything, it played games at a decent 60 FPS, paired with my 550Ti (previously a pair of 8800 GT 512 MB in SLI and clone memory setting on) i had no problem what so ever playing anything! and it only had DDR2 memory at 800 MHz! i received 3 PCs from family during the years, 2 Core Duo, one old from 2006 i think and an other later version (a 6000 and 8000 series of Core Duo i think) and a AMD A4 3400 and all 3 of them barely runs minecraft and can't even be used as a Minecraft server with the upgraded RAM i put in... it barely runs it locally... i can't compare the Q6600 to them considering i have no idea where i put it but i get the feeling it was more powerful even tho it was generations older! (tho i wonder if i should spend 50$ on a AMD processor to replace the A4 3400, the top of the line quad core should run a MC server i hope!) still the Q6600 kept my room warm in the winter and got me to play Crysis at 60 FPS (minus the map ascension lol) till 2017, name another processor that kept up for 10 years!

[@eduardoroth8207]

based old hardware

[@eDoc2020]

I think I have a Q6600 but I've never used it. My basement PC has a newer gen C2Q and I have it overclocked probably around 3.8GHz.

[@theepicslayer7sss101]

@@eDoc2020 nice in my case, over clocking the q6600 would have been impossible considering it was idling at 70 C, i think the revision after of the q6600 was idling at much lower temp so that would have been overclockable!

[@eDoc2020]

@@theepicslayer7sss101 If any half-modern CPU is idling at 70C then either it's not idling or your cooling is crap. "Crap" cooling can also be a feature by only ramping up the fan as much as needed.

[@theepicslayer7sss101]

@@eDoc2020 that is just how bad the q6600 was, the revision was way better since they had to fix that, but even so my processor never died so... but it was idling at 70 and ran at 85 full load.

[@foobar476]

There was me hoping there would be a MIX implementation.

[@Flopster101]

Someone better show up running mainline Linux on Carillo Ranch or I'm gonna be mad at this community

[@BlackRedDead1943]

5:29 "If unsure, say Y." ehhhm..., N.

[@immoloism]

You can't make videos like these as now I want one just to install Gentoo on.