OSDev.org

0x401000, I think I've seen that before...

Would it be any harder to redesign from a kernel based at 0x100000(It's correct this time, I'm going with your suggestion, 1mb. ) to higher half, than to redesign from 0xc... to 0xb...?

Fwiw, I'm building my kernel to 0x100000 and not relocating it at all until there's a very concrete reason to do so.

Would it be any harder to redesign from a kernel based at 0x100000(It's correct this time, I'm going with your suggestion, 1mb. ) to higher half, than to redesign from 0xc... to 0xb...?

It's not moving the kernel that's a problem. Because you will have the source code of your kernel. It's moving the apps, (you know the one's written by 3rd parties that are not open source and only available in binary form) that is an issue.

In all honesty though, I just cannot imagine a world where someone writes software for my O/S.

Candy wrote:Fwiw, I'm building my kernel to 0x100000 and not relocating it at all until there's a very concrete reason to do so.

Good to hear an established opinion that agrees. I was beginning to think I was noob-ish to even consider having a lower half kernel.

As to having to move compiled apps, I'll just have a set entry point. When my kernel fails to fit between 0x100000 and entrypoint, I'll just switch to a higher half kernel. The reason for not starting with a higher half kernel is that, I don't expect it to ever get bigger than a few mb.