QNX 6.3 Released

Lufi2 writes "QNX 6.3 was released on 3 Jun. New features include accelerated 3D, the Voyager 2 browser which supports HTML 4.01 and XHTML 1.1 now, SCTP (stream control transmission protocol) and packet filtering with NAT! GCC 3.3.1 is also included. If it's not a typo, the Professional version costs $8695/user o_O Usual QNX NC (non-commercial = free beer) LiveCD is not available on the download area yet (As of 9 Jun)... But it sounds very promising"

You don't have the slightest idea...

[ReKleSS]

QNX is older than linux. It's a microkernel. It's realtime. Linux is neither. There was a very appropriate quote from somewhere - "Linux is what you trust your server to. QNX is what you trust your nuclear reactor to." They each have their place. -ReK

Re:You don't have the slightest idea...

[Anonymous Coward]

"QNX is what you trust your nuclear reactor to." Which would be rather stupid. From the license:

B3.2 High Risk. Unless QSS has provided its express written consent for each Runtime Component in the Runtime Configuration, the Software may not be, and OEM will ensure that it is not, used in any application in which the failure of the Software could lead to death, personal injury or severe physical or property damage (collectively, "High-Risk Applications"), including but not limited to the operation of nucle

Erm...

[ReKleSS]

I never said the quote was *accurate*, just *appropriate*. That said... what the heck do they run nuclear reactors on, anyway?
-ReK

Re:Erm...

[turgid]

That said... what the heck do they run nuclear reactors on, anyway?

Well, in the UK they all have analogue multiply-redundant and diverse saftey ciruits that resond in miliseconds or less.

For routine temperature monitoring and data display they use everything from archaic 1970s minicomputers to RISC workstations and even the dreaded Windows NT.

Many of our reactors are designed such that they can be run safely (albeit in steady sate i.e. constant temperature and power output) on a few passive analogue guages.

Re:Erm...

[sporktoast]

If you are Ohio's FirstEnergy [ohio.com], you run your hole-in-the-head Davis Besse nuke [ohiocitizen.org] with Windows NT [security-focus.com].

Re:Erm...

[Short Circuit]

A navy Nuke could tell you, but then he'd have to kill you. :)

Re:You don't have the slightest idea...

[turgid]

Nuclear sites, in the UK, do not rely on manufacturers warranties or pledges to maintain nuclear safety. Rather, they have in place their own systems of design, resting, certification and legal regulation. You can bet they use products whose manufacturers disclaim liability. What manufacturer wouldn't? However, in the UK, the government underwrites nuclear liabilities. In other words, it takes on the responsibility itself. It does not leave it to others i.e. to chance.

For example, a friend of mine once worked at a UK nuclear power station wher all the machines in the control room ran Solaris on SPARC hardware. It has a similar license clause to QNX. At my old powerstation we used all kids of stuff, from QNX to DOS, Windows 3.11, NT4, Solaris on SPARC, VMS on Microvax... you name it. Once again, it was horses for courses, and the liability was all underwritten by HM Government.

Re:You don't have the slightest idea...

[happyDave]

Disclaimer--I wouldn't even say "standard," because they allow a loophole--"unless QSS has provided its express written consent." That seems much more willing than others would be to let their software run in "High-Risk Applications."

Re:You don't have the slightest idea...

[ameoba]

The point is that you -can- get "express written consent" to put the systems in a nuclear facility. Of course the licence, by default, is not going to allow such activity; doing so opens you up to a metric assload of liability. It's going to cost more to insure a company with contracts like that floating around, so the customers that need it are going to pay more.

Um

[bsd4me]

"QNX is what you trust your nuclear reactor to." Which would be rather stupid. From the license: ...

I seem to recall print ads from the early to mid 90's that showed QNX being used in a nuclear facility. I think the ad told about how they were able to update the system without taking it down because of the microkernel architecture.

Re:You don't have the slightest idea...

[Anonymous Coward]

You deserve your Informative moderation (I was wondering what QNX was too!), but in general you should try to avoid feeding trolls like that. That he was trolling, rather than merely ignorant, is obvious from the way he describes BSD as a "Linux clone".

Re:You don't have the slightest idea...

[Samrobb]

QNX is older than linux. It's a microkernel. It's realtime. Linux is neither.

Stock Linux is neither. I personally know of at least one company [timesys.com] that offers a hard real-time [timesys.com] version of Linux.

ObDisclaimer: yes, I work for TimeSys.

Re:You don't have the slightest idea...

[ArmpitMan]

The webpage doesn't have much information about how this is actually achieved -- can you actually make kernel calls from a hard-realtime process, or is it just another hack to allow tasks run at a higher priority than the really-not-particularly-realtime kernel?

ObDisclaimer: I just finished spending the last 5 months of my undergraduate career fighting tooth and nail with RTAI [polimi.it] for my thesis. Before that, I'd spent four years as a QNX consultant. Man, what a difference.

Re:You don't have the slightest idea...

[Short Circuit]

The kernel gets interrupted by events, and calls the process associated with the event.

Any process that process sends a message to gets immediately elevated to the priority level of the process sending the message...i.e. it gets run immediately.

The kernel also can get its temporary priority elevated in this way.

Re:You don't have the slightest idea...

[ArmpitMan]

Okay, but have you somehow made the kernel hard realtime, or just "low latency"? I can pre-empt it, but I don't suppose I can re-enter it? Say the kernel is in a busy loop waiting on the floppy disk drive, processing a request from some lowly soft-realtime task. Do I have some guarantee that it's going to come out in time? Or should I just never touch the kernel in a hard-realtime task?

This is why a microkernel and user-space drivers make so much more sense for real-time. Your priorities are much fin

Re:You don't have the slightest idea...

[ArmpitMan]

Timesys's website is infuriating. I click on "Technical details" and get "Here's a list of drivers that it comes with." Well, are they semi-standard Linux kernel drivers somehow recompiled to run in user space? Are they custom drivers developed by TimeSys? When you claim you've turned a non-realtime OS into a hard-realtime one, there should be some mention of *how*, somewhere.

I *think* the drivers are custom-written. At which point, why are you bothering to call the thing Linux? What's the advantage

Re:I don't get the attraction

[variable]

QNX isn't a Linux clone. QNX is a POSIX based OS that has a history that goes back a decade before Linux was even a glimmer in Linus's eye.

There is nothing given away for "free" either, there is an evaluation that is free - much the same as vmware has a 30 day eval of their product. After that 30 days, Neutrino will fall back into a more limited mode but will remain operational (like it was with NC before 6.3.0).

www.qnx.com [qnx.com]

www.qnxzone.com [qnxzone.com]

www.openqnx.com [openqnx.com]

Some reading to bring you up to speed on things.

Re:I don't get the attraction

[JustinXB]

Get a clue.

First, BSD is UNIX. It's not a clone of UNIX, it's not a clone of Linux. It has ties to the original UNIX.

Second, Linux is a clone of UNIX. A bad one at that.

Third, QNX is a clone of neither. It's a real-time microkernel. It supports things BSD nor Linux support, such as userland file systems and namespaces. It's an OS you install places where you need quality, something Linux knows nothing about.

Get a damn clue.

Um, what?

[gotr00t]

Well, it seems that justification of claims is obviously something that nobody takes into consideration on this thread.

Not to be overzealous, but if Linux is such a horribly done clone of UNIX, then why is it the most widespread and most used *NIX-esque operating system around, even more than BSD? No, seriously, please tell me.

Re:Um, what?

[Anonymous Coward]

Not to be overzealous, but if Linux is such a horribly done clone of UNIX, then why is it the most widespread and most used *NIX-esque operating system around, even more than BSD? No, seriously, please tell me.

Be careful about arguments like that -- it leads to others such as "If MS Windows sucks so much, why's it installed on 90% of home computers?"

Licensing Cost

[nuxx]

Just FYI, that licensing cost is most likely a developer per-seat cost with some redistribution allowed.

As far as per-user stuff, it's likely that most people use QNX in one form or another every day without knowing it. From cable boxes to ATMs, traffic lights, etc.

QNX is put in places where failure cannot happen. At all.

Real Time

[niker]

QNX is put in places where failure cannot happen. At all.

Not quite! QNX is a Soft Real Time Operating System - situations that need to fulfill "hard deadlines", for instance a medical monitoring device, will use a Hard Real Time Operating System.

Clicky [msoe.edu]- QNX is Soft Real Time

Clicky [real-time.org]- Different Real Time concepts

Re:Real Time

[variable]

Actually, QNX is a hard realtime system. Upper bounds are fixed and QNX is in fact used in many medical devices because of this behavior. Linux is a soft realtime system, since it can provide good response times but the upper bound is unlimited so it cannot assure those times when placed under load. The only real solution to this today is RTLinux.

The paper in the first link is very weak on details, not surprising for a 300 level undergraduate course.

Re:Real Time

[LeftOfCentre]

Is it sufficient for a system to be referred to as "hard real time" when the interrupt latency is provided? Some say that upper bounds on system calls need to be provided as well. This is not the case for QNX, is it?

Re:QNX is the bad touch

[AKAImBatman]

Most normal stuff you expect to find on a linux install is either broken in some way or missing entirely. It doesn't even have bash!!

Why are you trying to use QNX as a desktop UNIX workstation?

Re:QNX is the bad touch

[tepples]

Why are you trying to use QNX as a desktop UNIX workstation?

Could it be because the AC has found the popular GNU/Linux distributions too bloated for some specific hardware?

Re:QNX is the bad touch

[AKAImBatman]

Could it be because the AC has found the popular GNU/Linux distributions too bloated for some specific hardware?

Oooo... that's gonna leave a mark. Perhaps Mr. AC would like to comment? ;-)

Re:QNX is the bad touch

[AKAImBatman]

We needed a real time OS (or at least close to it) so thats why its running on QNX. The QNX box is communicating with embedded microcontrollers over an external bus.
[snip]
My preference would have been an old Sparc box running SunOS.


The words, "non-realtime OS" keep coming to mind...

They deviate from standards and make you have to dig to find where that deviation is. Is having man pages too much to ask??

Fair enough. But it *is* an embedded system. Footprint is everything. My guess is that you're sim

Re:QNX is the bad touch

[bsd4me]

What I want is for the environment on which I code to be less convoluted.

A big part of being an embedded developer is being able to adapt to different development environments and being able to develop and debug a system with minimal tools. Sometimes you have a VT100, sometimes, JTAG, other times remote debug via Ethernet. I have worked on systems where I have had to debug software with a logic analyzer because that is all that was available.

Re:QNX is the bad touch

[Anonymous Coward]

Tru tru, An instructor i once had said to the class "embedded tools are crap"

That still doesn't make me hate it any less.

Re:QNX is the bad touch

[AKAImBatman]

My point is not that it CAN'T be used as a Desktop Workstation, but rather that one shouldn't expect an embedded system to work exactly like one. For example, the original poster complained that QNX wasn't bundled with BASH. And why should it be? It's not intended for desktop use! If he wants BASH, he can download the source or binaries himself. He's working on an embedded OS, not a Solaris or Linux desktop. He should set his expectations to match.

Re:QNX is the bad touch

[dotz]

QNX is not intended for the desktop use? Man, so why they bundled that Photon graphic user environment?

Re:QNX is the bad touch

[AKAImBatman]

Man, so why they bundled that Photon graphic user environment?

For building user interfaces for devices? IIRC, the iOpener built its GUI around the full capabilities of the Photon GUI. Other uses might include Web Browser kiosks and touchscreen user interfaces.

Re:QNX is the bad touch

[ArmpitMan]

Man, you're a troll. QNX2 came on 5.25" floppies in 1987 [levenez.com]. Here are some examples of some other things that happened in 1987:

• IBM introduced the fantastic new "VGA" graphics standard, blowing everyone away with its 256 colours of fury.
• Windows 2.0 [aci.com.pl] came out.

Surely you'll concede that QNX 2 is superior to Windows 2?

Re:QNX is the bad touch

[AKAImBatman]

What exactly are you talking about? I think you've replied to the wrong person.

The original post (yours?) stated, and I quote:

"I'm developing software on QNX 2.x right now and it BLOWS! "

Shall we assume that the "2.x" was a typo, and that QNX 4.x (or Neutrino 6.x) was the intended version number?

Re:QNX is the bad touch

[variable]

QNX 2.x? There was no GNU toolchain for QNX 2.x, so I strongly doubt you are using it.

QNX isn't Linux, our default /bin/sh is ksh not bash, although you can just fire up qnxinstall and go to the QNX Online repository and install bash. All my x86 workstation machines have bash installed with no trouble at all.

It would have been more informative if you had actually said what was broken with g++ rather then saying it was just broken. The only thing that is different from our 2.95.3 g++ vs. say Linux or BS

Re:QNX is the bad touch

[variable]

Well, I am using a pretty old version of Debian.

#include <iostream>
#include <memory.h>
using namespace std;
int main(void) {
int c;
cin >> c;
char array[c];
memset(array,0,sizeof(array));
cout << "sizeof(array) is " << sizeof(array) << endl;
}

That compiles and runs on Linux and QNX and cygwin without any warnings or errors. Care to post another example of where g++ is "busted" on QNX? Please be sure to test it on other platforms first .

NC Download Changes

[variable]

With 6.3.0 the download version of QNX will actually be our full commerical product, with downloads avaliable for both Windows, Solaris, Linux and Neutrino. After 30 days, these PE (pro edition) versions will turn into what was once the NC edition with the pro features disabled.

I suspect the downloads will be up and ready very soon. You can find more details here:

http://www.qnxzone.com/

Re:NC Download Changes

[Lufi2]

I see!
I'd be happy to use it :-)
But I dont understand one thing: QNX is an OS, how do you (and QNX) mean that downloads will be available for Linux, Windows, Solaris, ...?

Re:NC Download Changes

[variable]

Not everyone develops for an OS under that OS. So we have cross development tools for targetting hardware running QNX that run under Solaris, Windows and Linux.

Dev boards

[jsweval]

I am looking into some uses for embedded devices and was wondering if there was an affordable dev board that I could use to get my feet wet. All the boards I have seen are very expensive ($700+) which is just too much for me. Ebay hasn't been much help either.

Re:Dev boards

[variable]

You can get a used iPaq (36xx/37xx/38xx) on eBay very cheap, and you can have a good time hacking on it with QNX, Linux, Plan9, etc. It's a good and cheap ARM development platform.