OSDev.org

Hi All,

I have finally plucked up courage to requesting a test for Caracal. Note that I am still in a fairly early 'draft' stage of the OS, and am just interested in whether you get the initial page shown (screenshot here).

You should get a (deliberate) PFE at 0x10000000 to demonstrate my BSOD. Currently, other than GRUB the OS does not try to read/write to any disk drivers (in fact, it leaves the floppy spun up at the moment).

The download page is at: https://sourceforge.net/project/showfil ... _id=193516.

Thanks in advance for any comments!

Cheers,
Adam

Works fine under bochs (with lots of warnings in the log)

Bochs wrote:LGDT: not recognized in virtual-8086 mode

VirtualPC has more issues with it (it could very well be the VME bug):

Virtual PC wrote:Caracal Boot Loader Version 0.8.1

Booting Kernel...
Privileged instruction 0x00CD, prefix 0x0000 is not supported by
the v86 mode monitor.
System Halted.

Testing on a real machine (Amalthea) displays three(?) lines of white text before going to a black screen after which there are no noticeable events.

Can't get into VESA mode in Bochs (default Windows binaries):
bochsout.txt:

pcmattman wrote:Can't get into VESA mode in Bochs (default Windows binaries):

Code: Select all

Caracal Boot Loader Version 0.8.1

Booting Kernel...
Error with VESA mode detection!
System Halted.

I get this also, in Bochs and VMWare

On one pc it works fine on second VBE is not present! Maybe try VESA instead of VBE2. I’m looking for a starter for my os. I’ve checked out about 20 os’es. Then I leave 6.(caracal, minirighi,netcore,idylla,aceos and amos) Due to last changes I decided to choose caracal. Great work.

viki wrote:Maybe try VESA instead of VBE2.

Dude, its the same thing

I just tested it on another build of bochs (emulating a 386) and it triplefaults on a move from CR4. All my other builds work fine. (for the record, they're all v2.3, maybe that has something to do with it)

Cheers for all the hard testing work, guys. The VBE thing is the fact that because it is just a stub at the moment, I have no 80*25 text mode driver - the system is therefore halted. That is next on the todo list! The VirtualPC error is indeed the V86 mode extensions error - but I'm glad that it at least gives you a sensible message. When I implement a v86 mode monitor in my kernel, I will add it to my second stage loader too, so I am not dependant on VME.

The LGDT error is a bit of a botch that I am taking out right now . When I originally wrote my v86 section, I got out of virtual mode by doing an LGDT, and thus causing a GPF which is caught by my boot loader!

More worrying is combuster's three lines of white text - I'll have a look at that later because I don't get it on my machine.

Again, thanks for the tests next job is to make it less dependant on specific hardware.

@viki: cheers for the complement - let me know how you get on!

Cheers,
Adam

Combuster wrote: I just tested it on another build of bochs (emulating a 386) and it triplefaults on a move from CR4. All my other builds work fine. (for the record, they're all v2.3, maybe that has something to do with it)

Looks like I have some detective work to do . Out of interest, how many builds of Bochs do you have

AJ wrote:Out of interest, how many builds of Bochs do you have

Eight:
2x bochs with 64-bit support (with and without debugger)
2x bochs with smp support (with and without debugger)
2x stock bochs (with and without debugger)
a 386 bochs
a bochs with debugger and a strange feature set