Locks up on boot

Here you can discuss ReactOS related topics.

Moderator: Moderator Team

n4mwd
Posts: 13
Joined: Mon Mar 13, 2006 12:42 am

Locks up on boot

Post by n4mwd »

This latest version of ReactOS crashes and freezes when its booted. I get to a distorted image of a lake with a mouse stuck in the middle. Keyboard is also dead. This is the second version of ReactOS that I have tested and neither worked but at least the older version worked with 16 color VGA mode a little before it locked.

I just don't know what to say other than its the last time I'm going to waste my time with this project. The developers of this "project" are clearly playing some kind of joke on everybody. I am ashamed to say that I once thought this was a legitimate project.

:x
oiaohm
Posts: 1322
Joined: Sun Dec 12, 2004 8:40 am

Post by oiaohm »

Please bear in mind that ReactOS 0.3.4 is still in alpha stage, meaning it is not feature complete and is not recommended for everyday use.

People are still not reading. Ros works in some emulators. Don't yell at us when you are being a complete idiot about alpha status os's.

Alpha status means if it runs on real hardware you are down right lucky. Even running stably in a emulator cannot be expected. But ros does run well.

Ros stability has incressed. Real hardware support has not expanded yet.
Haos
Test Team
Posts: 2954
Joined: Thu Mar 22, 2007 5:42 am
Contact:

Post by Haos »

Ok n4mwd... You have vented your feelings, now THINK for a moment.

Did you actually stopped to read at least our main page and license? Did you bothered to read the information at the first and second installer stage?

If so, did you understand those informations? Please show the same courage as with your original post and reply...
n4mwd
Posts: 13
Joined: Mon Mar 13, 2006 12:42 am

Post by n4mwd »

Haos wrote:Ok n4mwd... You have vented your feelings, now THINK for a moment.

Did you actually stopped to read at least our main page and license? Did you bothered to read the information at the first and second installer stage?

If so, did you understand those informations? Please show the same courage as with your original post and reply...
Nope, I never read any license page because it crashes before it gets that far. So what you guys are saying is that you are running and testing this thing in a fantasy world "virtual machine" and not on real hardware? For what reason do you publish Live CD's for if its not for testing on real hardware?

You guys need to wake up. If this is a real project then you need to start treating it as such! Trash your emulators and start running this thing on real hardware. No other SUCCESSFUL operating system was developed on an emulator. That goes back to CP/M, MS-DOS, UNIX and Windows. Emulators are for school kids working on a class project. This is the real world. If you want your project to be regarded as legitimate, then you need to leave your virtual reality academic fantasy world and come into the real world if you want your project to have any credibility.

If you are serious, then I want you to succeed. I'd love for you to kick Bill's butt. But right now, I'm not convinced that you are serious. Maybe Bill has scared you off. Maybe that's it.
Haos
Test Team
Posts: 2954
Joined: Thu Mar 22, 2007 5:42 am
Contact:

Post by Haos »

Great. We got an experienced OS developer here. I suggest you subscribe to our Ros-Dev mail list and share some constructive opinions on what we do wrong. Btw, saying the project is a joke, project not being legitimate, credible, serious, is not considered a constructive opinion. You need to do better than that.

I also suggest you dont mention any of your opinions on using Virtual Machines in development, as you apparently have no idea how wide and often are VM`s used in software development nowadays. Not as several multinationals are earning quite a lot of money by providing what you call
"school kids working on a class project" or "virtual reality academic fantasy".

So check your knowledge first, criticize afterwards, and make sure you have any credible arguments on your side. So far you have presented none.

EDIT: If you`re totallu dismissing all our work, you could always support your thesis by presenting your experience and successes on OS development. I am curiously looking forward to see this.
Z98
Release Engineer
Posts: 3379
Joined: Tue May 02, 2006 8:16 pm
Contact:

Post by Z98 »

I'm not entirely sure where you get the impression the developers are working exclusively on virtual machines. Several of the key kernel and win32k developers routinely test on real HW. However, because of the vast varieties of hardware configurations, they don't have the collective hardware to track down every oddity.

Next, at least two major components of the kernel itself is still incomplete or absent. Until they are locked down, I personally wouldn't hold my breath on them working.

Also, there are components that need to be tested that are relatively independent of the hardware compatibility. It's far easier to deal with these tests in a virtual environment, since these tests have nothing to do with how well ROS will deal with real hw.

Ultimately it's your choice whether you believe this project has merit, but do not expect much sympathy from the people who have successfully tested on real hardware if you continue with your attitude.
n4mwd
Posts: 13
Joined: Mon Mar 13, 2006 12:42 am

Post by n4mwd »

Z98 wrote:I'm not entirely sure where you get the impression the developers are working exclusively on virtual machines. Several of the key kernel and win32k developers routinely test on real HW. However, because of the vast varieties of hardware configurations, they don't have the collective hardware to track down every oddity.

Next, at least two major components of the kernel itself is still incomplete or absent. Until they are locked down, I personally wouldn't hold my breath on them working.

Also, there are components that need to be tested that are relatively independent of the hardware compatibility. It's far easier to deal with these tests in a virtual environment, since these tests have nothing to do with how well ROS will deal with real hw.

Ultimately it's your choice whether you believe this project has merit, but do not expect much sympathy from the people who have successfully tested on real hardware if you continue with your attitude.
I have a hard time believing that this runs on any real hardware. I did test this on three separate REAL machines with different hardware configurations. ALL three crash at the exact same point. The only differences were that the display of the lake is less distorted on one machine and another machine crashes before the two icons come up. If people are running this successfully on REAL hardware, then name it.

Simply put, you guys seem bright, but you are putting the cart before the horse on this one. You need to get the kernel to be stable across the board on real HW before you start working on all the other stuff. You seem to be using a Linux kernel of some sort, but somehow its locking. Forget about all the skins and themes and screensavers and such for now. The kernel is the most important part.

As for my references, I started coding in the early 80's. I have two CS degrees. I primarily code in C (at least you guys got that right). I worked on OS/2 at IBM in the early 90's when it transitioned from 16 to 32bits. (OS/2 failed for political and not technical reasons). Today, I do mostly embedded and Win32 stuff. Does that satisfy your curiosity?

So tell me, do you at least have a kernel debugger running so that I can log and control the kernel via the COM1 port? It would be nice to be able to see exactly WHY its locking up at boot on every machine I try.

One of you said in the above post:
"Alpha status means if it runs on real hardware you are down right lucky. Even running stably in a emulator cannot be expected. But ros does run well."

That is where I got the part about it only being tested on virtual simulators.
Z98
Release Engineer
Posts: 3379
Joined: Tue May 02, 2006 8:16 pm
Contact:

Post by Z98 »

You're kidding, right? This is NOT Linux. This is a from scratch implementation of the NT kernel. We use almost nothing from Linux, since it's not worth the trouble of porting their things to the NT architecture. You'll find that quite a few developers hold the Linux architecture in disdain as well.

We also have kernel debugging in all the release builds.
Haos
Test Team
Posts: 2954
Joined: Thu Mar 22, 2007 5:42 am
Contact:

Post by Haos »

I did test this on three separate REAL machines with different hardware configurations.
Three?? THAT MUCH? You have covered almost all possible set-ups... Wait... Its not enough even to cover all CPU makers....
ALL three crash at the exact same point. The only differences were that the display of the lake is less distorted on one machine and another machine crashes before the two icons come up. If people are running this successfully on REAL hardware, then name it.
Only if you`ll be able to get an exact copy of their hardware/drivers setup. Either its pointless.
You need to get the kernel to be stable across the board on real HW before you start working on all the other stuff.
Cool... and as we know, NT kernel devs grow on trees, same as NT WDM kernel driver devs and so on...

Btw, our kernel IS quite stable right now...

No, we cant work purely on real hardware. We do not have enough resources and manpower. We practicaly would have to use countless driver/hardware configurations, to discover all possible hw flukes. WE DONT HAVE RESOURCES FOR THIS.
As for my references, I started coding in the early 80's. I have two CS degrees. I primarily code in C (at least you guys got that right). I worked on OS/2 at IBM in the early 90's when it transitioned from 16 to 32bits. (OS/2 failed for political and not technical reasons). Today, I do mostly embedded and Win32 stuff. Does that satisfy your curiosity?
I dont believe you. Any proof? Why i dont believe you?
You seem to be using a Linux kernel of some sort, but somehow its locking.
Either you are a complete ignorant, who didnt bother to check it out before writing here, or you are just a plain liar.

Or maybe only an ignorant...:
So tell me, do you at least have a kernel debugger running so that I can log and control the kernel via the COM1 port? It would be nice to be able to see exactly WHY its locking up at boot on every machine I try.
Yes... its call a debug mode with KDBG - kernel debugger. Yes, you can receive debug output via COM1, using terminal software, like Hyperterminal.

Yes, it would be nice if you`d just report a bug with a debug log, instead of calling us names. It would also make you appear to us as a professional and serious person, not a spoiled, whining brat, who blames kernel instability, before even checking what was the real reason.

If you want to teach us about kernel importance in OS, you could at least have some decency to do some minimal research on our kernel, instead of discrediting yourself publicly, with opinions like one below:

"You seem to be using a Linux kernel of some sort, but somehow its locking." ...
dreams
Posts: 207
Joined: Tue Sep 26, 2006 10:23 am
Location: Holland
Contact:

Post by dreams »

http://www.youtube.com/watch?v=naPVtDaSVhQ

0.3.4 on real hardware. (took a few reboots but it works)

[edit]
i will try to post the exact specs, but i'm away from that pc this week. it's an asus motherboard, nvidia graphics, ATA harddisk, PS2 mouse/kb.
psyco001
Posts: 28
Joined: Mon Sep 17, 2007 9:30 pm

Post by psyco001 »

n4mwd wrote: I have a hard time believing that this runs on any real hardware. I did test this on three separate REAL machines with different hardware configurations. ALL three crash at the exact same point. The only differences were that the display of the lake is less distorted on one machine and another machine crashes before the two icons come up. If people are running this successfully on REAL hardware, then name it.
0.33 works on my old pc and i test to install 0.34 next weekend.
system:
IBM 200Mhz processor
196 MB SD-RAM
10GB IDE
ATI Mach64 pro
n4mwd
Posts: 13
Joined: Mon Mar 13, 2006 12:42 am

Post by n4mwd »

Since there doesn't seem to be any way to attach files here, I hostclipped a few pics. One shows the startup page where it shows the drivers being loaded. The second is the main screen where it locks with distorted video.

http://www.hostclip.com/dl/1bd40035aee9 ... 77e43b7ad7

I know it isn't fun having your hard work criticized like this, but someone needed to slap you. You need this thing to run on real hardware not on an emulator.

I ran it on:

ASUS A7N8X MB - ATI All in Wonder Radeon 7500 - Athlon XP 2400
Elitegroup PM800 - Onboard video - Celeron 2g
Asus P5AB + AMD 500 Mhz

No that isn't every possible hardware combination, but its enough to say that it doesn't work in any reliable way, but does crash in a reliable way.

And yes, I really did work on OS/2 back in the early 94. I worked on OS/2 Warp. Warp was a code name just like Chicago was the code name for Win95. That was the first time they used the code name in the final product. They had all kinds of trek like names. The one before Warp was Borg I think. There was also a Checkov and Sulu but I forget what they did - maybe commercial versions. But the disclaimer is that I was one out of about 3000 that actually worked on that OS. It was written down here in Boca Raton, FL which was also the birthplace of the IBM PC - the box that started it all.

No I don't pretend to be a great OS coder. I worked at IBM a long time ago and forgot most of the stuff. All I remember for sure is that most of Windows was stolen code from OS/2. But that's more political stuff. I don't do any OS stuff today except to whip up a quick RTOS for embedded stuff. But that's a whole lot easier than what you are doing.

Anyhow, you guys need to stop attacking people who report bugs in your program and just fix them. You say you don't have the resources and I say I've got three boxes that have reproducible bugs when your software is loaded. So there are some resources. If you want, I can try to create log files out of the debugger port if that will help. In any event, you need to stop using the virtual emulator. The deck is stacked in your favor when you use that and you will never get anywhere with it as far as finding and fixing bugs.

BTW, the real reason IBM OS/2 went under was because management used their influence to squelch bug reports so things would look better than they actually were to the higher ups. The result was that OS/2 was too buggy for people to tolerate. People went with Windows and stayed with them. That left nobody to buy OS/2 once the bugs were ironed out. DOES THAT SOUND FAMILIAR?

Lastly, have you actually downloaded your Live CD ISO and run it yourself? Have you compared it to the original to make sure nobody has tampered with it? Remember, although I want you to succeed, there are definitely some very powerful people out there that don't want you to succeed. No need for names, we all know who he is.
oiaohm
Posts: 1322
Joined: Sun Dec 12, 2004 8:40 am

Post by oiaohm »

First two unsupported video cards n4mwd.

Yes producing live and boot cds they do make up part of testing even in emulators. Since they can show completely different errors.

Do you have the reactos log got threw a null modem connected to serial port. Most likely not. Its a requirement for people testing on real hardware. So we have a complete debuging log of what went wrong.

Yes we know the current video driver is picky.

Asus P5AB don't know what its using.

Note keyboard defect is being chased down. For some reason it don't happen in emulators at all. As well it does not happen on all motherboards. Particular chipset don't like how ROS does is keyboard/mouse and become unresponsive.

n4mwd OS/2 history and death was more cost than bugs. Windows was always very buggy. Problem is by the time OS/2 came around on price most software was being coded for windows. So it died.

n4mwd you attacked us. So we responded in a equal way. There are far more critical kernel things to fix so we can use the real video card provided drivers.

Ros core developers is less than 20. Basically ros does not have the resources to chase down every bug. Instead resources have to be targeted where it will do the most long term good. Fixing the driver system is a lot higher than fixing the video driver for real hardware. Thinking due to the errors in the driver system the video driver will have to be redone after the driver system is fixed.

Basically fix the video driver now in about 18 months the driver has to be dropped in the trashcan and recoded due to the fixes required in the driver system. Now stop trying to force developers to work on something that is marked to be destroyed. Ie we don't have the recourses to waste on something that pointless. Now if you want to provide a coder to fix those faults we will accept the patches.

Reactos is a alpha no were do we promise that it will run on real hardware. We know it runs on a really limited subset of real hardware. Just because you have not found it that is your problem.

Ros storage location is outside the reach of all possible threats.

Now listing what you tested on was a basic requirement for you first post if you were not wasting our time. Now you really are wondering why we are cutting you head off? You have been a idiot provided us with incomplete information. Then insulted us. Really what in heck did you expect us to do. Be friendly. Heck no. That is not our nature. We are nice to people who are nice. People who are nasty gets picked on.

Now since you are a win32 coder why in heck can you not fix the bugs yourself n4mwd. Or are those degrees you have useless bits of paper because at some point you should have covered OS development and the nightmares in it.
No other SUCCESSFUL operating system was developed on an emulator.
Complete bull that line. You must not class linux as Successful. Early Linux was developed inside boch and on a extreamly limited subset of real hardware note even a smaller subset than what reactos runs on now until it got stable enough for more hardware. Ie lots of people did not have the hardware that matched early Linux either. VM's are a good way of letting more developers in until the OS gets out of that nightmare. Almost all real-time oses were developed in cpu emulators so coders could see exactly what was going wrong.

Basically n4mwd stop proving to everyone that you are a complete idiot. The more you prove it the less we should trust your claims. At least your OS/2 claims sound right for a IBM internally contained person who could not see from the real world point of view of price and conversion costs.
n4mwd
Posts: 13
Joined: Mon Mar 13, 2006 12:42 am

Post by n4mwd »

If you use an emulator, the proper technique is to discover bugs in real hardware and then see if you can duplicate them in the emulator. You guys are relying on the emulator to say that it works in the release. That is what set me off. There's a difference between using it as a tool and using it as a crutch.

As an additional test, I went ahead and tried to install the OS to a spare HDD. As soon as it gets to the "Choose Language" page, it locks the keyboard and mouse. I am using standard PS/2 keyboard/ mice on all machines.

I admit that I haven't looked at your code in any detail, but that part of the install on most systems is usually written in 16bit. Something tells me that you did it in 32. There is nothing wrong with that, but if I'm right, then the preceding page that shows drivers being loaded is in 16 and you lose the keyboard/mouse controller when you make the switch.

I also tested a USB mouse just in case but it ignores that. I didn't expect that to work in an alpha version because that is advanced stuff. I just thought I would test it anyway.

As to my video cards, I know the difficulties in dealing with a million different cards and ten million different ways to access them. I assume that you are using a universal VESA driver until you get the proper driver loaded. Anyhow, the video driver thing is a non-issue at this part of the game. I said before that I actually have a picture, just not all that good, but a good enough of one to see what is going on.

The real problem is in the K/M locking like that. That did not happen with the previous released version. If you knew about that, then you should not have released it in that condition.

When I have time (I am very busy at the moment) I will take a look at your source and see if I can spot the problem with the K/M. If I find something, I will let you know. The same thing goes for the kernel debug. As you know, RS232 ports are hard to find these days, but I can do it, its just going to take a lot of messing on my part.

As far as OS/2, it was in existence long before Windows 1.0. It started as a souped up DOS around 1985 or so. The original PC-DOS was an 8 bit OS designed to run on an 8 bit processor that worked with 16 bit instructions - the 8088. The 8086, the 16 bit version, was never actually used in PC's. I wrote a multitasking system for the 8088, but it never caught on. When the 286 came out, OS/2 came out also and capitalized on the 286's hardware abilities that better supported multitasking. Windows followed soon after. At the time, OS/2 was being jointly developed by both MS and IBM. The two later had a falling out.

It wasn't until the early 90's that OS/2 went 32 bits. Now here is a little tidbit you might could use. Although Borg used a 32 bit kernel, about 90% of the drivers were 16 bits. This way, it was easier to get things going. When I left them, they still had about 50% of the drivers running only in 16 bits. Not very pretty, but it did work.

You are correct that IBM ran out of money. They had more than enough to develop it 10 times over. The problem again was in the foolishness of the upper management. They would literally hire two complete departments, with multiple software engineers and then pit them against each other developing the exact same code. In other words, the department that completed their task first, say a new mouse driver, got to keep their jobs.

Another bit of information that you probably already know is that the Windows function calls don't always work the way it says in the published documentation. That's partly because its actually documentation for OS/2 and OS/2 does things a little differently than Win32. Remember I said MS stole a lot of their code from OS/2. That includes documentation. So when you are making your OS, don't put that much faith in the Win32 documentation. Go by what it does, not by what it say it does.
Haos
Test Team
Posts: 2954
Joined: Thu Mar 22, 2007 5:42 am
Contact:

Post by Haos »

Okay, let us start over.

n4mwd: i would never attack a person that does the bug report. You can check my posts all over the forum. If someone reports that ROS doesnt work on his real hardware, i usually explain how to get a debug output out of it, and to report this bug. I do this even if we know what bug it supposed to be, still it needs to be confirmed.

We are aware of limitations in development based on virtual machines, but we cannot afford using any other methode. We cant set up a farm of thousands of different PC`s to test ROS on them, as we lack resources and manpower not even to set it up, but also only to process such amount of information. Microsoft in the same time is using supposignly, hundreds thousand of different PC sets, to test its OS...

We use mostly (but not only) VM to get a fairly repeatable baseline for developing and testing. To intensify our real hardware testing, we need people commited to the project, that would test ROS regularly, at least once a week. And we need a lot of them.

I`m sorry if you feel offended, but believe me, venting and questioning the project`s credibility is not a best way to get your message across. We know our current limitations and always are welcome for a constructive criticism.

Let me then apologise you and still have hope that you will stay interested and maybe even contribute to this project.
Post Reply

Who is online

Users browsing this forum: No registered users and 33 guests