Slashdot Mirror

This is not an anti-Linux remark; there's a similar hack for NT that puts a mini-OS under NT, and it has the same problems. I can understand why Linus isn't that thrilled with RTLinux. The right way to do this is to start with a message-passing microkernel and build upward from there, not start with a UNIX-type kernel and build down. I'd like to see an open-source microkernel catch on. The GNU Hurd people were on the right track, but that project seems to be moribund.

The key to doing this right is to have a very, very limited microkernel with fast message-passing. If you do it right, drivers, file systems, and almost everything but message passing and scheduling is outside the kernel and can't crash it. This is how you build systems that work when it matters.

A limitation of such systems is supporting hardware boards that do complicated DMA. Devices that look like channels (SCSI, FireWire, USB, etc.) are fine, but graphics accelerators create problems, as do random ISA and PCI boards that do DMA. Safely managing a device that can store anywhere in physical memory from outside the kernel is tough. QNX generally supports display devices in plain VGA mode for this reason.

Then now would be a good time to go to http://www.qnx.com/~glen/deadbeef/1897. html or http://www.hmrs.org.uk/fourfoot.htm or http://csf.colorado.edu/mail/pkt/f eb98/0556.html.

As for the Space Shuttle size, that one's rarer, but basically it goes on to say that because the SRBs (they're called something else, now, I think, but they're the long widgets that get it off the ground. not cranes) are transported by rail (they are!) through rail tunnels (probably) they are therefore only slightly wider than the rails (inasmuch as 12' is slightly wider than 5', yeah, I'll buy that) and therefore mankinds first method of transportation (horse's ass) determined important details of his most advanced (space shuttle). All very stupid.

IMHO assembly is the right choice when it comes to speed, as i'm involved in game development and most of the time the extra speed is mandatory for complex algorithms and hardware tricks, but games are developed for common hardware and most of the time they don't really need to be very flexible.

But in the OS side, speed is definitely not as important as a clever design. If these guys want their OS to lift off they better be very thoughtful, because it doesn't make any sense to code the fastest hack if you can't really extend its capabilities or program applications easily. I think a good example of what does a well-built, small, fast AND portable OS look like is QNX. OS, GUI, Web server, browser and some more stuff fits in a floppy, and most of it is written in C/C++.

Personally, I hate coding the same thing twice for two different platforms, I really like the idea of using my code for more than one thing, and that's what standards like ANSI C and POSIX are all about.

These guys might as well win some Assembly '00 award with amazing code like the V2, but let's be serious, assembly should be used only when needed.

That said, coding in assembly is really fun (except when you have to explain your code either to other person or to yourself after some weeks of working at something else :)

--------------------

I think I'll download it to see what neat new stuff they've packed into it.

--

Java is on the way... Java & QNX

Probably wont fit on a floppy. Then again, KVM on the pilot is ~160K expanded.

You can
take a look at what QNX has to say about the Amiga at http://www.qnx.com/amiga.html.

As an Amiga user (does this mark me for death?) I was excited to hear
the new OS would be based on something so solid and seemingly
revolutionary. Gateway/Amiga really stomped on the Amiga
community IMO...

Conventional operating systems use a single flat memory architecture where hard-to-detect programming errors like corrupt C pointers can cause programs to overwrite each other (bad) or the kernel (worse). The inevitable result: system failure. A Neutrino-based system, however, can intelligently recover from software faults, even in drivers and other critical programs - without rebooting - because every OS component runs in its own MMU-protected address space. Neutrino's full MMU support also simplifies testing since it identifies which module tried to perform an invalid memory access - at the exact instruction. What can often take weeks or months to identify in a conventional RTOS takes virtually no time with Neutrino.

QNX's unique approach to transparent distributed processing lets you launch processes across the QNX network, allowing for full inheritance of the environment, including open files, current directory, file descriptors, and user ID. The microkernel also includes POSIX.1 (certified) and many POSIX.1b realtime services, as well as high-speed diagnostic event tracing.

QNX's unique approach to transparent distributed processing lets you launch processes across the QNX network, allowing for full inheritance of the environment, including open files, current directory, file descriptors, and user ID. The microkernel also includes POSIX.1 (certified) and many POSIX.1b realtime services, as well as high-speed diagnostic event tracing.

QNX runs on PPC. It probably doesn't support pmacs, though, but i'm guessing you'll probably be able to find embedded altivec g4 based SBCs soon.

Even a cursory look at the phase5 and qnx sites tells a lot about the future of the amiga aftermarket.

G4 accelerators, QNX neutrino, possibility of a G4 ATX mobo, replete with the best in german engineering.

I suspect the "silence" is Amiga's way of keeping their options open after the sloppy way they dumped QNX for Linux a few months back.

QNX site
Phase5 site

no, he said he was dropping qnx and going with linux.

but that statement has since been pulled from the amiga website, which is the point of this "silence" article.

qnx still has their statement posted, ie, they are going full speed to support qnx on the "new amiga".

the only german company mentioned on the qnx site is force5, which makes accelerators for mac and amiga. they will have a G3/G4 accelerator out for existing amigas in september. it will ship with QNX.

some are speculating force5 will make a ATX G3/G4 motherboard, and ship it with QNX.

having a decent quality realtime system for free could be a bonus for the new line of amigas. the various real time packages for linux are kind of a hack, since they are done by third parties outside the kernel development tree.

The screen shots of neutrino, upcoming qnx release, look pretty nice. the company is known for small footprint/high performance/high relibility software.

but pc users get squat? what the dilly yo on dat?

QNX oughta give away a student or non-commercial version (at the least) for pc people. last time I saw a price, QNX was over $1000 USD for PC's

free qnx for amiga users

He may have pissed off QNX greatly.

QNX has news about the free QNX for amigans who use the phase5 accelerators.

Not only is QNX going to ship with Phase5's accelerators, but it is interesting to note that part of the agreement involves allowing Phase5 to ship "standalone G3/4 based products" -- could this be a rockin' G4 ATX motherboard?

Here are the links:

Phase5 G3/G4 accelerator news

qnx + phase5 announcement

Although it is off-topic, QNX also has news regarding Real Time Java! Which means the garbage collector has been reigned in. Good news.

Real Time Java/QNX announcement

He may have pissed off QNX greatly.

QNX has news about the free QNX for amigans who use the phase5 accelerators.

Not only is QNX going to ship with Phase5's accelerators, but it is interesting to note that part of the agreement involves allowing Phase5 to ship "standalone G3/4 based products" -- could this be a rockin' G4 ATX motherboard?

Here are the links:

Phase5 G3/G4 accelerator news

qnx + phase5 announcement

Although it is off-topic, QNX also has news regarding Real Time Java! Which means the garbage collector has been reigned in. Good news.

Real Time Java/QNX announcement

The QNX demo disk is a web-browser-onna-floppy. Although it's min 386+8MB, and I don't know about legality in this situation. Probably this is not what you want.
(OTOH, it could be a nice way of preventing kids from installing any software you don't want :-) )

I also recall from my DOS days a web-browser / newsreader integrated, that would run on a 286 / XT (although, from memory, not as nicely as it would on a 386).

Called Net-Tamer (aka Internet in a Thimble), it's on Simtel, or there is the home page.
(looks like they're still developing it too!)

--
Repton.

The QNX demo disk is a web-browser-onna-floppy. Although it's min 386+8MB, and I don't know about legality in this situation. Probably this is not what you want.
(OTOH, it could be a nice way of preventing kids from installing any software you don't want :-) )

I also recall from my DOS days a web-browser / newsreader integrated, that would run on a 286 / XT (although, from memory, not as nicely as it would on a 386).

Called Net-Tamer (aka Internet in a Thimble), it's on Simtel, or there is the home page.
(looks like they're still developing it too!)

--
Repton.

Does QNX even support SMP by the way?

Yes, Neutrino supports SMP.

First? I think not. http://www.qnx.com/hotnews/pr/teligent.html

Dan Dodge, CTO QNX Software Systems, officially says there WILL be a QNX-AmigaOS REGARDLESS of what Amiga Inc. says about Linux.

http://www.qnx.com/amiga/wherewestand.ht ml

On the one hand, Linux is adapting to Desktop computers in terms of integrated features and this is a good thing, nobody questions that, but on the other hand, Linux's size means its suitability for embeded applications becomes questionable. Not as questionable as using NT but still... (correct me if I'm wrong, and I wish I were, but I do have heard micros~1 is pushing its NT kernel (if there is such a thing...) for embeded applications)
Anybody remember this /. article about a cash machine BSODing?

Now... there is always a way to compile a custom kernel with minimal functionality, thanks god! the size of the compressed kernel source means for a great part that there are now lots of features / drivers / functionality that were just not available a couple of years ago.
However, how does a minimal 2.2.10 compares with a 1.2.13? (both considered stable if I recall correctly)

This is important, because projects like linuxrouter depend on being embedable, either on a floppy or much much more important, on devices like the diskonchip thingy. Do you see the market? routers, dns, xterms, you name it! Any of those devices could work on an embeded 486 with 8Mb or more of RAM. Is this 486/8Mb under my desk, a masquerading router/caching DNS? I thought so too.

Now... you can spit on Linux, and use QNX ( check it, QNX is the king!).
OR, because we're dealing here with a minimal Linux, actually designed for embeded applications, there is a lesser chance to get it wrong, therefore adding to security /stability / once again, you name it. Also, but some people might argue this is not the point (then please tell me what is the point about Linux), make it more attractive to hardware hackers.
Hackable alternatives to QNX might include katix, ELKS or even FreeDOS... and of course, uClinux.

Well well well,those are just some thoughts of mine, I'm nowhere involved with The Kernel, therefore I'm not in a very good position to critisize it, I just hope that the right people have reached the same conclusions or are aware of the fact that we may loose an important market if we only concentrate on desktops and above.

---

On the one hand, Linux is adapting to Desktop computers in terms of integrated features and this is a good thing, nobody questions that, but on the other hand, Linux's size means its suitability for embeded applications becomes questionable. Not as questionable as using NT but still... (correct me if I'm wrong, and I wish I were, but I do have heard micros~1 is pushing its NT kernel (if there is such a thing...) for embeded applications)
Anybody remember this /. article about a cash machine BSODing?

Now... there is always a way to compile a custom kernel with minimal functionality, thanks god! the size of the compressed kernel source means for a great part that there are now lots of features / drivers / functionality that were just not available a couple of years ago.
However, how does a minimal 2.2.10 compares with a 1.2.13? (both considered stable if I recall correctly)

This is important, because projects like linuxrouter depend on being embedable, either on a floppy or much much more important, on devices like the diskonchip thingy. Do you see the market? routers, dns, xterms, you name it! Any of those devices could work on an embeded 486 with 8Mb or more of RAM. Is this 486/8Mb under my desk, a masquerading router/caching DNS? I thought so too.

Now... you can spit on Linux, and use QNX ( check it, QNX is the king!).
OR, because we're dealing here with a minimal Linux, actually designed for embeded applications, there is a lesser chance to get it wrong, therefore adding to security /stability / once again, you name it. Also, but some people might argue this is not the point (then please tell me what is the point about Linux), make it more attractive to hardware hackers.
Hackable alternatives to QNX might include katix, ELKS or even FreeDOS... and of course, uClinux.

Well well well,those are just some thoughts of mine, I'm nowhere involved with The Kernel, therefore I'm not in a very good position to critisize it, I just hope that the right people have reached the same conclusions or are aware of the fact that we may loose an important market if we only concentrate on desktops and above.

---

First of all, there's more than a single embedded systems market. There are truly low-end systems (think toaster thermocontrols) where a Unix-like OS is certainly overkill. On t he other hand, there's a very large market segment where a Unix-like OS is not overkill. Telecom switching equipment and industrial control panels fit into this category. In these segments, it's not unusual at all for your "small" computer to be a m68k, PPC, SPARC, or Pentium running at a respectable speed and having 4, 8, or even 32+ Mb of memory. This is the market space currently occupied by VxWorks ( Wind River Systems), pSOS ( ISI), LynxOS ( Lynx Real-Time Systems), VRTX ( Mentor Graphics), Chorus ( Sun), QNX ( QNX Software Systems), OS-9 ( Microware), and a whole bunch of others that I can't think of off the top of my head. Oh, yeah, and Windows CE.

I even know of folks who are using Solaris or AIX in this type of application. Not to mention the poor fools who are stuck with Windows NT for control and monitoring applications (don't laugh, it happens!).

Now, the majority of these real-time operating systems (RTOS) are POSIX-compliant, at least to some degree. Usually, this means that they implement the POSIX.1 APIs. So, except for the tasking and memory models, they look a heck of a lot like Unix/Linux. Some of them (LynxOS, QNX) are downright Unix clones (like Linux). All of them have TCP/IP networking available, all of them either already have Java support or have promised it Real Soon Now, lots of them even use gcc/gdb as their toolchain.

Given that, why not Linux? Margins tend to be sensitive here, so you'd really like to eliminate those per-cpu royalties that the commercial (I should say "proprietary", since MontaVista is obviously commerical) vendors charge. And (contrary to what some /.'ers have posted) uptime is critical in many of these applications. Long-term supportability and credibility is also critical, as many of these systems may have production lifetimes of years, and field lifetimes of decades. Customers already recognize the value of having access to the source code in these circumstances, and they tend to demand it from their software vendors (it is usually provided under NDA or escrowed).

Now, check out the founders of MontaVista. Jim Ready helped put together VRTX, one of the first commerically-successful RTOS's. Jerry Kirk founded Microtec, which is the dominant vendor of compilers and debuggers for embedded systems (the XRAY debugger is the most-used debugger in the embedded world). So I would guess these two ought to have a good understanding of how to put together a OS + development enviroment that would appeal to the RTOS market.

It looks to me like, between this and eCos, we're about to see The Cathedral and the Bazaar dynamic in yet another OS market.

While Linux is larger than Emacs, at least Linux has the excuse that it needs to be. -- Linus Torvalds

Doesn't seem very convincing to me. If people are still trying to solve QOS issues with overkill capacity then that's seems like little more than a cludge.

I should be able to do video-conferencing no problem on 10Mbit/s Ethernet (the bandwidth is there), but if the image breaks down whenever there's a burst of activity from the department file server that's a pretty fragile solution.

Overkill capacity is only overkill until someone builds a faster file server, or until you are unlucky and someone accesses a large cached file from a fast Linux machine, saturating your 'overkill' net.

Sounds to me like the difference between a real RTOS and a timesharing system. Try asking a developer who uses QNX and a 68k for a real-time app whether they would like to switch to Windows NT and an Alpha 21264-500 with overkill processing power. All the processing power/bandwidth isn't going to help you if one app decides to monoplise it.

By the way, I know that people keep trying to build RT systems out of NT. I can't imagine why. I even worked on a project that did that, and it was painful.

Astronomers should know about it. RSA is typically performed using 512 bit prime numbers. There are approximately 3.778e151 such prime numbers. Using the advanced storage technology available to the NSA, it should be possible to store a 512 bit number in a single hydrogen atom. A typical universe (e.g. ours) contains approximately 1e90 hydrogen atoms. If the NSA has hidden 3.778e61 universes in an inconspicious little building in Maryland, astronomers should notice some deviations in the gravity field in the area.

if someone know QNX, they sell their OS on a CD and there's a standalone floppy disk that boot into Photon (their GUI) and a little program detect everything in your computer and display it as a tree, like in windows/control panel/system/peripheral (it's not their demo disk, it's the boot one for the CD, i don't know if they can make it available to everyone?), i use it sometimes in computerstore to check compatibility, i'm looking for a notebook that can run QNX and BeOS fully. also there's some dos freeware programs that identify a lot of hardware, i remember one written in TP with sources.
--

why write code when program can write it for yourself? i'm using the ApBuilder from QNX/Photon for those who know it, i click 2 or 3 times, write a label, click generate, click make and it works! then i can do a wc -l *.h *.c and wow what's a number! :o)
--

Cisco licensed QNX last May, so I don't think they're going to use Linux or go Open Source. Check out QNX's press release.

The whole point of a server is stability and reliability. You want to setup a server to perform certain services (file service, print service, mail service -- or daemon depending on what world you come from) and that it would just do it's job with very little intervention from the
administrator. Ideally, all the administrator should be doing is to continually customize the services the server is performing to meet the
needs of those receiving service. However, things not being ideal, an administrator has to worry about hardware failure and hence must add
tape backup, RAID, etc. to his server for fault tolerance, insurance against data loss, and to minimize down time. Now concerning the issue
of whether or not to put certain drivers in the kernel, you must look at your application. In a server environment, you want to maximize
stability and reliability. Therefore, the ideal situation is to minimize the number of factors that can directly affect stability. Device drivers are included in that category. With the creation of new hardware comes new drivers. The question is, is the driver updated enough to run the risk of failure? Patches and fixes are great, but you still have the possibility of failure present if such patches and fixes
are needed. The question when looking into the operating system's design, and determining what drivers to include, do the drivers present
a danger? Are they patched frequently? Are they stable? But those are all relative to what we know about the driver and it's history? What
about rare and unknown bugs? Bugs that show up under unusual circumstances. Ever had a machine crash once out of the blue without knowing why? I have. Only God knows why in such situations. Maybe it never happens again. But it still happens. So what does one do? Do we accept that it will happen infrequent enough that we won't have to deal with it again? If you want stability, that is not acceptable. You cannot
just leave such unknowns unanswered. NT and Linux both have such vulnerabilities since they both can have drivers present in the kernel.
Getting back to the server, what do we do? Well, a design which I have been very impressed with in terms of an operating system is QNX. I do not know how many of you are familiar with it. It has a small message
passing kernel and runs everything else as protected and separate entities. I found their microkernel model to be very impressive. I
downloaded their demo and was very impressed with it also. For a server environment, it is a much better design than Linux or NT. I hope that
someday Linux and NT will follow it's example. Perhaps two versions of each operating system. One version with a QNX-like model for servers,
and one for high-performance workstations that don't need that kind of reliability. Another thing of noteworthiness that I wanted to look at
deals with a difference between linux and NT. GUI interfaces are nice and pretty and are sometimes easier to use than a console. But if you think about it, on NT, the GUI is always running wasting precious memory and cpu time. I mean, how often does an administrator tinker with his server. Not often unless he is experiencing a high demand for changes from his users. Linux allows you to run the GUI if you want to, when
you want to, and shut it down when you are done. In my opinion, that makes it a better design. I hope that microsoft will go back to the Win
3.xx/DOS model that is similiar to the X-Windows/Linux Model. It would certainly improve their server product. One of the engineers that
designed VAX VMS worked on NT as you probably know. I believe at the heart of it, the microkernel is pretty good. It is all those nasty
libraries on top, the integrated GUI, and the registry (yuck!!!had to mention that) that mess everything up. Hopefully my logic on this stuff is sound, but if someone finds fault, please reply and set it all straight. The point of discussions like this is not about who wins the argument, but that we find the best answer. One additional note about QNX, it is designed for embedded systems and the currently do not seem interested in the server market. However, I have heard rumors about Cyrix and QNX getting together to build ultra-cheap boxes for web-browsing or something. If you access their search engine, it provides detailed explanations about the QNX microkernel. I hope you all will read up on QNX. Linux and NT could definitely learn some things from it.

strange - my Cyberstorm 604e PPC 233MHz with 68060 50MHz amiga seems pretty real. Think before you write something. Dolt. Look where team amiga are in the RC5-64 decryption tables. Yep. Hovering about sixth or seventh. Not bad, for such a minority eh? PPCs kick x86 butt.
Incidentally, I dual boot AmigaOS 3.1 and LinuxPPC (APUS is the name of the Amiga ppc linux port - it's on sunsite.auc.dk). I have an 8MByte Permedia 2 gfx card, 128Mbyte ram, and ultra-wide scsi 3 4.3 gig HD.

No, it's not a top of the range spec, but I'd say it's not bad.

Makers of the amiga PPC hardware ->
www.phase5.de

Makers of the PPC-side kernel favoured by the majority of PPC amiga users, and the Warp3D Amiga OpenGL API->
www.haage-partner.de

Claim to be developing an amiga G3 card ->
www.escena.de

amiga ppc news->
come.to/amigappc

Amiga ->
www.amiga.com

QNX ->

www.qnx.com

Yes, most decent amiga stuff comes out of the UK and germany, and this is a US site, so some ignorance among the readership of this site is to be expected, I guess.

PS. I suppose my machine isn't a "desktop" - It's in a tower case...

Check out QNX. It's intended for embedded/small memory systems. They fit networking, kernel, a gui, and a browser on a bootable 1.44 disk. I think you can download the disk image.

x86 - 386, i386 EX,
Am386SE/DE,
AMD ÉlanSC400/410, 486,
Cyrix MediaGX, Pentium,
Pentium Pro, Pentium II

PowerPC - 401, 403, 603e,
MPC860, MPC821, MPC823

MIPS - R4000, R5000,
NEC VR4300/4102/4111


http://www.qnx.com/products/os/neutrino.html

I, too wish QNX was open source - It looks like it is where the HURD is headed. Howver, one thing to remember is that it is the kernel that is closed source - device drivers written by third parties need not be, they don't run in kernel space, and the standard dev tools for Neutrino are GNU ( not for older QNX OSes though, they use Watcom...). The OS is POSIX, so porting is pretty easy, too - except that QNX has that Photon MicroGUI windowing system, which has a somewhat different API to x-windows... That's not to say x-windows isn't available for QNX - it is, the QNX people just consider it too big for embedded applications...

http://www.qnx.com
http://www.amiga.com

Chapters bookstores in Canada are using QNX for salesdroid book lookups and POS. They have terminals around the stores for the droids to key things into, with a nice text-based captive user interface.

Plenty of power, plus you get QNX reliability and support.

For my mission-critical system, I'd probably take QNX over Linux too. :)

And of course, anything is better than NT. Well, maybe not Win95. :)

zaha wrote:

until linsux is real-time, its applications in the embedded, mission-critical markets will be limited.

I wouldn't use a standard Linux kernel for a real-time application, that would be silly. The same goes for a Solaris kernel or an NT kernel. None of them are designed for real-time operation. If I was doing a simple real-time system, I would use a simple real-time kernel like QNX RTOS or Caldera's DR-DOS.

For a more complicated system, I would use a hybrid RT system like Real-Time Linux. Check it out, I think you'll be pleasantly surprised.