This forums is for OS project announcements including project openings, new releases, update notices, test requests, and job openings (both paying and volunteer).
Below is a screenshot of my new long mode kernel. It currently consists of:
IDT & GDT routines
Simple VGA console driver
Kernel Page allocater (stack based)
Kernel Heap allocater (bitmap based)
LAPIC/SMP initalization routines
Boots with standard GRUB (kernel contains a 32bit bootstrap entry point)
The screenie shows it parsing the MP tables and bringing up a couple of APs. The APs enter the kernel, setup a new stack and loop incrementing a byte in video memory (top left). I've tested it successfully on a range of SMP capable PCs and laptops with success.
Nothing experimental or interesting about this project. My main interest with this project is SMP right from the start.
..will bore you with more details once it's more complete
I'm not really into doing releases or using source control for that matter.
But since there is a small amount of interest, the source is attached below. It's just an achive or where I'm currently at.
It won't compile in its current state as I've had to remove the printf implmentation used. I'm not sure about the license of that code. (It's the Lars & Linus vsprintf....) Plug your own in or comment/#define the calls out. Makefile sucks, you'll probably need to fix that too!
No explicit license attached, do what you want with it. I may in the future release the project under a BSD license though. I'd appreciate it if you could me a PM if you do anything interesting with it
You need a GCC Cross-Compiler for x86 64. By the look of the error, the makefile uses absolute paths to the compiler, so if you already have the Cross Compiler and it is not working, it needs to be installed in /usr/cross/x86-64/bin. Alternatively, edit your PATH variable and edit the makefile so that it doesn't supply the full compiler path.
13th wrote:dosfan!
I have trouble compiling the code you posted
Hi. As AJ said the makefile uses absolute paths. I did mention the makefile was pretty useless and I only released the code in its current form because people were asking. Sorry.
The screenshot below shows the current version loading an ELF binary from an EXT2 filesystem with the page faults doing the work. It's the usual "Hello, world!" but there is no syscall interface. It is of course ring3
Hopefully this will be seen as constructive criticism, but I'd look in to that Makefile thing.
You have made an announcement and generated some interest in your kernel. This means you have a great new resource: willing testers! By making the build process as simple as possible, you should (hopefully) get some useful feedback.
I know I'm jumping the gun, because you only originally intended to release a screenie at this stage, but make use of the people who have shown that initial interest. Just my two penn'orth
Hopefully this will be seen as constructive criticism, but I'd look in to that Makefile thing.
You have made an announcement and generated some interest in your kernel. This means you have a great new resource: willing testers! By making the build process as simple as possible, you should (hopefully) get some useful feedback.
I know I'm jumping the gun, because you only originally intended to release a screenie at this stage, but make use of the people who have shown that initial interest. Just my two penn'orth
QFT. I do appreciate the interest shown. Unfortunately, due to RL I don't have the amount of time available to code that I'd like. (I do 90% of my hobby coding on the train on the way to work). This means half a**ed Makefiles and the like.
There was also part of me which suspected the post was a troll attempt because I thought I made it clear it wouldn't even link nevermind a botched Makefile.
I take your comments on board with spirit you intended. My next "release" will be of a higher standard
.
