Latest version: 0.04
You are not logged in.
Hi all,
I am uploading an image of my boot disk. It is a compressed WinImage file, so you're going to need WinImage in order to place it on a floppy or create a regular image file for Bochs.
I have gotten clock code working, with an update IDt call in the OS API. My next step is considering how to load a driver into memory and run it. I am going to use a modified DOS driver setup...so it may take some time.
BTW, the OS should display everything it is doing on the screen. Then at the bottom of the screen a display of the boot time will occur, and approximately 60 seconds later a message should appear at the bottom of the screen, then the clock should resume being displayed. I have set the PIT to 1000 Hz, and am checking the timer for 60,000 tics. The odd thing is it appears to be a little fast compared to the system clock (off by one second).
Anyway, enjoy!
Offline
Hi,
I tested your OS on my 486 laptop and on my p2 333 mhz.
P2 specs:
- intel p2 333 mhz
- 2 HD's: Maxtor 4 gb and an 850 mb quantum hd.
- 64 mb PC66
- VESA 2.0 compatible video card with 4 MB of memory (it's an ATI 3D Rage Pro AGP 1x)
- a 28x speed cdrom player
- a 1,44" fd
- a 3c90x NIC
and 486:
- 486 cpu dx2
- 8 mb
- a non working HD, have to figure out one day to get that working...
- normal VGA card
I gave it a run on both computers. I got on both computer @ "A20 gate: " and then it somehow hangs. (waited at least for a minute)
On BOCHS however within a second it ran.
Maybe I should download/test it again. I think it's rather weird. 
HTH,
DennisCGc.
Offline
Humm!
Perhaps I need to look at my A20 code more closely. I wrote that nearly a year ago, and I wasn't as good a programmer then in ASM as I am now. I have four machines I can test on, none are lower than Pentium compatible. The code I've been using for A20 initiation probes the ports directly and there may be some compatibility issues. Also, did the CPU get identified correctly? It should have seen the 486 and the 486DX2 as different types but compatible with one another. I am eager now to see if PCI32 works on those machines...
Do you know if the BIOS calls to initialize the A20 gate works on your machines? I may put a patch in place in order to see if I can get that working...
BTW, this is my first bug, thanks!
Offline
BTW, On the second machine, it should recognize that there isn't enough RAM and ask you for a key press to reboot.
I'm calling the first machine DennisCGc_1 and likewise the second DennisCGc_2.
Offline
Hi,
mm, about the CPU thingy: DennisCGc_1 reports "Pentium or equalivent found" (spelling 
)
But it reports the same @ DennisCGc_2.
About the BIOS thingy, no idea, maybe it does. What I found weird is that my a20 works at my computer (I think)
Nothing is reported about the memory thing.
DennisCGc.
Last edited by DennisCGc (2005-06-11 19:28:40)
Offline
Cool, I'm gonna test is asap, but I don't have WinImage right now.. 
Offline
Here is a new one to try A20 gate.
Wow, I'm concerned it didn't understand your CPU. I need to check that again as well. I will have to proof all my words too, thanks!
Offline
Great job smiddy, just test it and it worked great, found every thing fine.
(I will post full spec on test PC later.)
Photo shot here : http://falconrybells.co.uk/SmiddyOS.jpg
PS: I have the book "Dissecting DOS" if you want me to look any thing up let me know, the OS that comes with it is here : http://world.std.com/~mikep/rxdos-page.html
PS: I should of reset the screen over a bit for the photo .
Offline
Thanks Dex4u...I'll have to give some screen shots of my three machines. As you may well guess, if there are too many devices, like my wife's laptop, the display is too full to show everything. This leads me to developing a logfile, that is switchable from the config.sys so that I can capture everything due to the over shoot.
I looked at you screen shot prior to reading your whole text and notice this too. It is using 700 x 400 to get 90 x 60 characters, actually using Chris Giese's code, where he uses the VGA registers to program it. Unfortunately, I haven't figured out how to program it myself yet. I tried installing a different font and I get what looks like a 16 line character, overlaid with an 8 line character. It looks funny, but doesn't do what I wish, yet.
I downloaded the source and I'll take a look when I'm able. We're packing things today and throwing things away...I'd rather be coding, but alas, the trials of adult-hood. Thanks again for you testing!
Offline
Here's an image from my wife's laptop (boy, I must be really shakey):
The second one is much clearer...but I can not get everything in focus...ARGH!
Last edited by smiddy (2005-06-12 21:06:37)
HPIM4023-1.JPG, Size: 55,460 bytes, Downloads: 936 HPIM4033-1.JPG, Size: 156,262 bytes, Downloads: 932Offline
@smiddy, it looks like a scene from the film "Hackers" .
PS: I see you have a AC97 on board sound chip, i written a basic AC 97 sound card driver for "Dex4u" if you want a demo of it, to try let me know (it may work with yours ?), it comes with a CLI CDplayer.
Dex4u.
Offline
Figured out that the WinImage file could be renamed to .zip
Your OS worked fine, and did even report the devices on BOCHS, that seem
to use Compaq hardware.
Tested BCOS ( brendan's? OS from mega-tokyo ) at the same time, your OS look
alot like his but does a better job with PCI.
/ Christoffer
Offline
Hi,
It works now. Still there remains an error @ my 486 laptop. It still reports it has found a pentium processor. And it hangs when "Entering Unreal Mode: " is displayed. On my P2 it works great. Nicely done!
DennisCGc.
Offline
Hi DennisCGc,
Thanks for testing...and the complement!
[b]UNREAL MODE[/b]
I'll put together a DEBUG version to display information along the way. Essentially it should directly after Unreal Mode: go into probing memory.
[b]CPU Identification[/b]
Do you know the specifics of the processor? What speed is it operating at? I assume 100 MHz, but could be pumped up higher. Are you able to identify it correctly with your code? Is it an Intel, or some other processor?
Hey bubach,
Thanks for testing my OS. As for the Compaq devices, you can change that to Intel by turning on the 440FX (I think, I'm at work, which makes me brain dead) chipset in the config/setup file for Bochs.
I hadn't had the opportunity to look at BCOS yet. I will try to this evening but doubt it. It may be this comming weekend.
Offline
BTW, I don't know if anyone noticed, but on my wife's laptop the screen writes off to the right side. This is interesting because a normal buffer would go on to the next line. But on this screen it just doesn't display. Since video is a linear buffer that is displayed as an array, I wonder what is going on in this case.
Does anyone have an idea as to why? I don't know enough about programming the VGA registers yet in order to figure it out.
Offline
[b]CPU Identification Revisited[/b]
I looked at my code again and think I know what the issue is. Your CPU has CPUID, which I attributed to Pentium and above. This is a wrong assumption on my part so I need to rework that code to consider the multitude of possibilities.
I'll post a my revized image probably this evening (your very early morning) for you DennisCGc...
Offline
bubach wrote:
Figured out that the WinImage file could be renamed to .zip
Interesting, inside is a IMA file! I used this knowlege with the Dex4u floppy maker. Made the zip, tried to extract the contents.... Got an error Maybe that is why you can't use it in emulators :shifty_eyes:
I still have to test soe of these OSs, the only problem is that my laptop game me errors about the disk I'll go buy a pack soon.
Jas2o
Offline
smiddy wrote:
[b]CPU Identification Revisited[/b]
I looked at my code again and think I know what the issue is. Your CPU has CPUID, which I attributed to Pentium and above. This is a wrong assumption on my part so I need to rework that code to consider the multitude of possibilities.
I'll post a my revized image probably this evening (your very early morning) for you DennisCGc...
MM, you're right. my OS does a CPUID as well on that machine (and therefore it crashes, because I immediatly assumed when if the CPUID command is supported, RDTSC is supported too, which is not correct).
Good luck with your OS 
DennisCGc.
Offline
DennisCGc wrote:
MM, you're right. my OS does a CPUID as well on that machine (and therefore it crashes, because I immediatly assumed when if the CPUID command is supported, RDTSC is supported too, which is not correct).
Yupper, that'll do it. There is a huge assumption out there (the cloud of the net) that RDTSC is support if CPUID is supported. Which pretty much kills CPUSPEED testing using RDTSC. I was working on a routine a while back to do CPUSPEED based on the CPU and clocks/opcode, but there are so many different processors out there now that it has become difficult to approach it from the clocks/opcode angle. But perhaps I'll reopen that [b][i]after[/i][/b] I get my device manager done.
Offline
smiddy wrote:
Yupper, that'll do it. There is a huge assumption out there (the cloud of the net) that RDTSC is support if CPUID is supported.
It wasn't the assumption, it was the laziness I had. (hence the code @ the osfaq )
Offline
