Linux removes support for 386

From: Dave Brockman 
------------------------------------------------------
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

https://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commit;h=743aa456c1834f76982af44e8b71d1a0b2a82e21

Ingo submitted it, Linus "pulled" the patch into his repo.

Regards,

dtb
- -- 
"Some things in life can never be fully appreciated nor
understood unless experienced firsthand. Some things in
networking can never be fully understood by someone who neither
builds commercial networking equipment nor runs an operational
network."  RFC 1925
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.17 (MingW32)
Comment: Using GnuPG with undefined - http://www.enigmail.net/

iEYEARECAAYFAlDI8u4ACgkQABP1RO+tr2QaJwCfUHGy6NXVJR0YeQz+JA/GbNN4
SzEAoIlCLHjJmbxM25Om7CK6r4+Fh1r4
=XAG+
-----END PGP SIGNATURE-----

=============================================================== From: Stephen Kraus ------------------------------------------------------ Damn, what am I going to do with that stack of Pentiums now?!

=============================================================== From: William Wade ------------------------------------------------------ Try the 1.0 series?

=============================================================== From: Dave Brockman ------------------------------------------------------ -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 If you mean the P5 generation variety (P54C/P55C), you run either the 486 or 586 kernel. Anything Pentium Pro and above can run 686 kernel, although you'll have to adjust PAE support depending on your exact processor. Regards, dtb - -- "Some things in life can never be fully appreciated nor understood unless experienced firsthand. Some things in networking can never be fully understood by someone who neither builds commercial networking equipment nor runs an operational network." RFC 1925 -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.17 (MingW32) Comment: Using GnuPG with undefined - http://www.enigmail.net/ iEYEARECAAYFAlDI+RkACgkQABP1RO+tr2RHkACfbYMXBRxrIgfAnN+i56Dwizdn q6cAn2J649Pwrs4v0RDhVsV7slNOLI+4 =nUVR -----END PGP SIGNATURE-----

=============================================================== From: Stephen Kraus ------------------------------------------------------ I meant my stack of 33 and 66 Mhz Pentiums. Oh well.

=============================================================== From: Dave Brockman ------------------------------------------------------ -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Those would be P54C processors, 486 or 586 kernels work fine. Regards, dtb - -- "Some things in life can never be fully appreciated nor understood unless experienced firsthand. Some things in networking can never be fully understood by someone who neither builds commercial networking equipment nor runs an operational network." RFC 1925 -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.17 (MingW32) Comment: Using GnuPG with undefined - http://www.enigmail.net/ iEYEARECAAYFAlDI+2IACgkQABP1RO+tr2SrGQCfV19G/qbAW9KSybZIakrDRceR 30oAoL6wX4aMXrLvbpb/PvqafTdTu1zj =CX1L -----END PGP SIGNATURE-----

=============================================================== From: Dave Brockman ------------------------------------------------------ -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 2.4.37ish is probably the latest safe bet on an actual 386 processor. This only affects actual 386 processors, 486 and above remain supported. Regards, dtb - -- "Some things in life can never be fully appreciated nor understood unless experienced firsthand. Some things in networking can never be fully understood by someone who neither builds commercial networking equipment nor runs an operational network." RFC 1925 -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.17 (MingW32) Comment: Using GnuPG with undefined - http://www.enigmail.net/ iEYEARECAAYFAlDI/IAACgkQABP1RO+tr2Rr6ACeOXbZKD555wp2kW6Z22AmYi8w 1qwAn2mrtWHpziHsudhpFowyDKWDbbjG =xKoP -----END PGP SIGNATURE-----

=============================================================== From: DaWorm ------------------------------------------------------ Wonder how many embedded 386 boards out there will notice? Anyone else ever run one of those? I have, but purely DOS only (and not even real MS DOS at that). Jeff.

=============================================================== From: Dan Lyke ------------------------------------------------------ I found it fascinating that Intel was building 386 chips into 2007. I mean I know various embedded device chips have long lifespans, but there's got to be some interesting economics involved in whatever devices were still 386 based a decade and a half into the product's life without warranting a re-design to something lower power and... what... according to Moore's law a thousand times faster? Wow. Dan

=============================================================== From: Ed King ------------------------------------------------------ Answer: none I predict Elks linux is about to see development hit an all time high (like I am right now)

=============================================================== From: Chad Smith ------------------------------------------------------ Some things I just don't understand. Like why anyone would be using a 386 in 2013 (which it almost is). The "last ones" were made in 2007 - which is about to be 6 years ago - but in 2007, we already had Core 2 Duos, so I don't know what they were being used for even then. 386 is an 80s chip! Unless you got some missions critical data on a 30 year old computer - (I'm sure I'm looking at you US Government) - you should not be using a 386 in 2013. Or 2007. Or any year starting with a 2. If you do have missions critical stuff on a computer not made this millennia - start getting it off of there immediately. What is your problem? *- Chad W. Smith*

=============================================================== From: Stephen Kraus ------------------------------------------------------ Chad, There are plenty of 386 based PLCs and embedded systems....

=============================================================== From: Chad Smith ------------------------------------------------------ But WHY?!?!? The 486 came out in 1989. NINETEEN EIGHTY-NINE. That means before the 1990s even started, the 386 was outdated and replaced. This is over 20 years after that. Our $8.99 prepaid cell phones have more processing power than that. It would have to be more expensive to run a 386 for a month than to buy a brand new system with more power. There are $50 tablets with more power - and they have WiFi, a touchscreen, and a battery. So even saying "I already have this system" doesn't mean you are actually saving money by using it. An order-of-magnitudes more powerful system that takes less electricity to run would save you money in the long term. And by "long term" I mean a few weeks. *- Chad W. Smith*

=============================================================== From: Stephen Kraus ------------------------------------------------------ Because PLCs and many embedded systems don't need power, or even efficiency. They just need a dirt cheap processor to parse things as they come along.

=============================================================== From: Rod-Lists ------------------------------------------------------ recompile. It is open source. ----- Stephen Kraus wrote:

=============================================================== From: Sean Brewer ------------------------------------------------------ Go upgrade the command computers on the international space station. Let us know how that works out.

=============================================================== From: Rod-Lists ------------------------------------------------------ I read somewhere that 386 & 486 are used in space due to the fact that they are more resistant to cosmic radiation than modern processors. ----- Stephen Kraus wrote:

=============================================================== From: Sean Brewer ------------------------------------------------------ Radiation hardened ones, yes. Radiation hardened processors are super expensive to develop and manufacture. Pretty much no one could afford to build a current generation rad-hard processor. The RAD750, which is just a rad-hardened PowerPC 750, is used in the Curiosity rover that landed on Mars back in August. It has two of them with one acting as a backup. They cost about $200,000 each. Yes, you can land on Mars with just 200MHz of processing power.

=============================================================== From: Chad Smith ------------------------------------------------------ They work without Power? How is that possible? *- Chad W. Smith*

=============================================================== From: Chad Smith ------------------------------------------------------ Which is about 5 times more powerful than the most powerful 386. But I'm hoping that NASA isn't using a vanilla Linux kernel for its work, and if it is, they don't have to upgrade to the latest version. Only machines that are being upgraded to the latest kernel will be affected, and if you haven't changed the processor in 28 - 5 years, you probably aren't that big on being "up-to-date" on things... *- Chad W. Smith*

=============================================================== From: Aaron Welch ------------------------------------------------------ Well crap... and here I was about to update my toaster oven to run Samba 4.0= . Gotta scrap that idea. Maybe it will run on the garage door opener? -AW nd if it is, they don't have to upgrade to the latest version. d, and if you haven't changed the processor in 28 - 5 years, you probably ar= en't that big on being "up-to-date" on things... ure. Pretty much no one could afford to build a current generation rad-hard p= rocessor. osity rover that landed on Mars back in August. It has two of them with one a= cting as a backup. They cost about $200,000 each. hey are more resistant to cosmic radiation than modern processors. hey he 20 g power nth ts with ually em that And e: The but in eing used ritical any year s your

=============================================================== From: Dave Brockman ------------------------------------------------------ -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 1) Lots of embedded devices aren't being upgraded, and if they are, they damn sure aren't being upgraded to the latest anything 2) The last kernel release that is actually 386 stable was released sometime last year. And that was again 2.4 train. You know anyone shipping 2.4 kernels still, let alone as default? Not all "computers" are obvious. They aren't all designed with a 2 year obsolescence. If Fab stopped in 2007, most "new" units were sold by 2008, I hope they are starting to become a non-factor by 2018 or so. Regards, dtb - -- "Some things in life can never be fully appreciated nor understood unless experienced firsthand. Some things in networking can never be fully understood by someone who neither builds commercial networking equipment nor runs an operational network." RFC 1925 -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.17 (MingW32) Comment: Using GnuPG with undefined - http://www.enigmail.net/ iEYEARECAAYFAlDJaagACgkQABP1RO+tr2SuOQCcD1yclry1b6N0CgwK6ZbnC32c IiAAoL3jNmgQxeG9QSk8FPcz2z0UpDiE =vLAv -----END PGP SIGNATURE-----

=============================================================== From: Sean Brewer ------------------------------------------------------ They got rid of them quicker than that. They stopped taking orders in March, and the last shipment left in September: http://web.archive.org/web/20061009060120/http://developer.intel.com/design/pcn/Processors/D0106013.pdf

=============================================================== From: Dave Brockman ------------------------------------------------------ -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I'm gonna stick with my 2018 prediction :) Regards, dtb

=============================================================== From: Chad Smith ------------------------------------------------------ That's what I said. I am not sure, but I think we are in agreement. Anyone still using a 386 isn't worried about being up-to-date. So who cares about the latest version of Linux lacking 386 support? That would be like "Shocking News - your 1908 Model T will NOT run on the latest unleaded/Ethanol hybrid fuel." Which, of course, is actually not true, since a Model T would run on pretty anything that was liquid and burned. But my point is - if you are a Model T enthusiast - or if NASA needed Model Ts to haul their rockets out to the launchpad for some reason involving cosmic rays, it still would not effect the use of the latest fuel in the slightest by anyone who is worried about buying the latest fuel. Because they aren't driving a Model T. Nor would it affect Model T users because they are used to having to find sources for rare parts and whatnot, because they are Model T enthusiasts, and people who cling to old crap cling to it because it is old, and not everyone could make a Model T run - you have to baby it, or manhandle it, or both somehow - you have to take the time to make it work, to get to know it, to find the parts and know where they go.... The challenge is part of the appeal. If I could walk into any Best Buy and get a fully loaded PC Jr. brand new in the box - it wouldn't be a collector's item. If every Walmart had Newtons on sale, would anyone buy them? Even those people who search for them in working condition now? it's not just about holding onto the past - it's about putting forth effort into holding onto the past. Now, those people who "need" 386s - for space stations or defense systems or whatever - they are not relying on Linus Torvald to keep them running. At least I pray they aren't. Or stuff might just start falling from the sky, since he has let them down. *- Chad W. Smith*

=============================================================== From: Dan Lyke ------------------------------------------------------ On Wed, 12 Dec 2012 20:58:34 -0600 Chad Smith wrote: I know of a facility that runs their CNC routers on a Commodore 64. When I asked why, the boss said "I know how to repair discrete logic, and it still works." Why bother re-designing when you don't have to? Engineering can be *expensive*. Dan

=============================================================== From: Chad Smith ------------------------------------------------------ So can losing everything because your ancient hardware breaks and you can't find one at a yard sale anymore. Hope you backed that up to something other than a cassette tape.

=============================================================== From: Mike Harrison ------------------------------------------------------ Pretty cheap for a custom chip of that magnitude. I know that -ONE- of the factors is the runs and gateways inside the chip have to be much larger, to delay the inevitible cosmic rays degrading them. This affects the ability to run at higher speeds. Ram, and everything else has the same considerations. All of a sudden, programming everything in assembly and C becomes important, no room to run Ruby on Rails.. (sarcasm). If you do have a 386 based system (embedded or other), the good news is that it has plenty of operating system choices. Just not the latest Linux one.. (unless you roll your own.. still an option). No loss for the Linux world.. it's time to move on. Channelling the future: When will the bio-wetware interface driver for my cubic photon processor array be bug free? I still see the occaisional glitch (Deja Vu style) when using the two-way full sensory mode of YouPorn. If only they would drop all that crap code for backwards compatibility to silicon based octo-processor chips and beowulf clusters....

=============================================================== From: Ed King ------------------------------------------------------ how about a car-wash run by a c64 ;) http://www.youtube.com/watch?v=TmvyLCxknI8

=============================================================== From: Sean Brewer ------------------------------------------------------ Haha, if only mruby would become stable.. ;-) The Curiosity was running on something like 2.5 million lines of C code. Testing scripts written in Python. If you're interested in C language standards at the JPL: http://lars-lab.jpl.nasa.gov/JPL

=============================================================== From: Stephen Haywood ------------------------------------------------------

=============================================================== From: Benjamin Stewart ------------------------------------------------------ Here's another interesting read relating to code at JPL: http://www.flownet.com/gat/jpl-lisp.html

=============================================================== From: Ed King ------------------------------------------------------ interesting reading. makes me wonder how many of us have a coding standards document in place in our shops that is a hypothetical question, no need to answer and incriminate yourselves ;-)

=============================================================== From: William Wade ------------------------------------------------------ We created some at SupplyHog, but I have yet to write a sniffer for them. Ours are fairly loose compared to some however. When we add more coders I will be writing a sniffer (perhaps the new recruit should do it...) and running it before commits as well.

=============================================================== From: "Alex Smith (K4RNT)" ------------------------------------------------------ You can recompile processors now? Wow, and to think I was planning on spending hundreds of dollars on FPGAs, all I had to do is buy a Core i3?!?!?!? Where do you get some of this? :P ,

=============================================================== From: Ed King ------------------------------------------------------ I just compiled an hp pa-risc "processor" (via qemu source code) on my =0Ai= ntel+linux box ;-)=0A=0A=0A=0A=0A=0A

=============================================================== From: "Alex Smith (K4RNT)" ------------------------------------------------------ Emulation yes, but not on an existing piece of silicon. That was my point. :) e l,