OSDev.org

hang on while i boot the desktop up again

pte 8007803c -> 0x00000000 : pwutnadg
pte c000d4c9 -> 0x0010d000 : PwutnAdg

[me=Tim Robinson]tries to remember what all that means...[/me]

Ah, address 0xc000d4c9 is mapped OK: it's present ('P') and mapped just over 1MB, where GRUB loaded the kernel. But address 0x8007803c (which should be part of the floppy driver, fdc.drv) is zeroed.

Tried "The Mobius" again using the latest boot floppy image in VMWare, I get a triple fault after the modules are loaded.

I haven't looked at "The Mobius" source closely, so I don't remember how interrupt are managed, but I have found a problem in VMWare with my little kernel...

In VMWare if an interrupt via TASK GATE occurs when you are in V86 mode it crashes with a triple fault... all goes well if the interrupt is managed via INTERRUPT GATE. On a real hardware both solutions go well.

This is what happens to me and, obviously, I don't know if it's a VMWare bug or not... Maybe this will help Tim or others to debug their os...

Well, I've a dual board with 2x200Mhz MMX. Maybe this could be interesting for someone who wants to code a kernel with multi processing support. Just mail me
[email protected]

I've tested "The Mobius" and "Clicker" in:

- a real hardware P100MHz, 32MB RAM, ATI 4MB video card.
- a real hardware PIII 600MHz 256MB RAM, Guillemot GeForce 3D Prophet video card
- VMWare 3.1 under Windows in a virtual hardware with 128MB RAM, PentiumIII 600MHz processor.

Testing "The Mobius" on the real hardware P100 ->
...crashes after test.exe is loaded!!!
It repeatedly prints text on the screen (can't read what prints it's too fast) and finally fills the screen with strange blinking/coloured characters.
(...there is not a S3 Trio64 video card)

Testing "The Mobius" on the real hardware PIII 600MHz ->
the same.
(...there is not a S3 Trio64 video card)

Testing "The Mobius" on VMWare ->
...VMWare internal monitor error BUG F(104):1147 bugNr=116
even after test.exe is loaded.
(Again...)

Testing "Clicker" on the real hardware P100 ->
Seems to work correctly, but it should turn off the floppy motor when the kernel is loaded... (the floppy led remains on) ...and too debug messages!!!

Testing "Clicker" on the real hardware PIII 600MHz ->
the same.
   
Testing "Clicker" on VMWare 3.1 (for Wiondows) ->
...VMWare internal monitor error BUG F(104):1147 bugNR=116.
It prints "Loading kernel", "Loading image", "raw host" and crashes!!!

tried "Valeo" from DarylD, which ends with a quite strange result ..

testing environment was BOCHS-1.4.1, megs=16 ...

<code>
[KERNEL ] [2] boot module(s) found
[KERNEL] Boot module found at [0x10C000-0x10F95F] (14687 bytes)
[KERNEL] Boot module found at [0x110000-0x115676] (22134 bytes)
[INTERRUPT] Interrupts set up
[MMU] FATAL: No multiboot memory map information availabe - Halting
<cursor>

This is a request for my OS to be tested. It can be found at http://sourceforge.net/projects/therx

therx disk image written to disk

started bochs with it. Got a screen kind of menue.
After a few seconds it craps out and bochs quits with
prefetch: running in bogus memory.

check it out, then we shall see further.

Same happens to me when booting Therx on Bochs.
If someone could please test Flick (http://www.sourceforge.net/projects/flick/) and tell me if it worked fine, I would be quite happy (note you need Linux and GRUB).

I know this happens in Bochs (but not why) However I am yet to find a real PC that it doesn't run on. Any ideas.

Therx

@therx:

your os comes up with this menue bar and a red(?) bar on the bottom of the screen, but it craps out after that. Keyboard becomes inoperable and doesn't react.

tested on: amd athlom xp 1700+,512 mb RAM, asus mobo, direct boot from disk.

i'll have a workable kernel (note the key word: workable) sooner than i expected. so soon i'll need some people to test it on various computers, to see what it does and doesn't support. i'll post again when i need testers.

unknown user wrote: kbuild.bat only runs without modification on the c:\ drive, with /src being c:\src and /tackle being c:\tackle .

Hum. I'm still unsure of the right policy, but i would like to ask enthousiast beginners to wait until they get a disk image to be tested before they post it here. Remember the purpose is to post tests reports, not test requests (though test requests are ok if self-moderated). Here, the fact Hakware is source-distributed with a huge Dos-only toolbox makes me think it still need a bit of polishing before one can actually test it ... maybe i'm wrong).

btw, i would really ask to avoid posting of "hello world" disk image.

Therx (0.1pre-a.img):
Boots fine, keyboard works fine, task 2 runs properly, clock runs properly, console appears to be functioning.

Pype-Clicker (c32-0-8-0 because that was the most recently update):
The debug mode is annoying as hell, but all initilisation stages went ok. There were a few warning about missing symbols in one of the consoles and the consoles themselves had a nasty tendency to overwrite one anothers top lines. Aside from that it seemed to work. Didn't do much but sit there, but it sat there happily .

Mobius (mobius-20020225):
Dunno if Tim is still around, but this was an old image (6+ months old), tested it anyway. It loaded fine (Although there were a huge number of msgs about the kernel module which went by too fast to read), then locked up at a screen with one blue background line at the top, a few red background lines at the bottom and a whole bunch of alphanumeric soup all over the place.

All tests done on k6-2 333hz, 385mb ram, ali5 chipset, geforce 2mx, sb16.

Curufir wrote:Mobius (mobius-20020225):
Dunno if Tim is still around, but this was an old image (6+ months old), tested it anyway. It loaded fine (Although there were a huge number of msgs about the kernel module which went by too fast to read), then locked up at a screen with one blue background line at the top, a few red background lines at the bottom and a whole bunch of alphanumeric soup all over the place.

Hmm, I'm due a new Mobius release soon. Hopefully I've fixed a few fundamental/stupid bugs which should improve things.