Page 2 of 2
Re: Can anyone see any problems with this C++ code?
Posted: Tue Jan 08, 2013 1:33 am
by iansjack
I linked to the SimNow home page earlier in this thread. Unfortunately that page only contains a link to the Linux version although it says there is a Windows version. If I were you I would consider running Linux for OS development, either as a dual-boot or in a VirtualBox VM. It is far easier to set up and use the GNU tools in such an environment than in Windows IMO.
Re: Can anyone see any problems with this C++ code?
Posted: Tue Jan 08, 2013 2:43 am
by Combuster
Looks like SimNow was never updated past Windows XP.
The best alternative is to compile Bochs with debugger enabled.
Re: Can anyone see any problems with this C++ code?
Posted: Tue Jan 08, 2013 3:18 am
by iansjack
You can even use the console in qemu to do limited debugging, although it's not as versatile as other tools.
TBH, I think that careful inspection of the generated assembler code should reveal the problem in this particular case, and would be a very instructive exercise; but that becomes exponentially more difficult as the complexity of the source code increases. Some sort of debugging tools is an essential IMO.
Re: Can anyone see any problems with this C++ code?
Posted: Tue Jan 08, 2013 10:20 am
by Nessphoro
The problem is not in the code.
I would say a missing rodata section or something else wrong with the linker.
Or the bootloader
Re: Can anyone see any problems with this C++ code?
Posted: Fri Jan 18, 2013 3:44 am
by BMW
Ok guys.
I have moved to ubuntu for OS dev. I am already finding it much better.
I think the problem here was that I didn't use a loader (assembly code which calls the kernel entry point), resulting in the kernel being executed incorrectly.
Solved