AhnTri VS HackOS

This is my boot.S code and this is my update list on what was updated to what

For the link, visit AhnTriOS Github.

I was working on a PMM, and finally have created one, you can look up here --> https://github.com/AhnJihwan/AhnTri/blob/main/mm/pmm.c

Hmm if you don't have it yet, u can consider mine XD https://github.com/AhnJihwan/AhnTri/blob/main/drivers/romfont.h

Jihwan Ahn, 2021, https://github.com/AhnJihwan/AhnTri

I'm sorry to say this, If you know how, I will really be pleased to contribute a little. My code is updated, https://github.com/AhnJihwan/AhnTri/blob/main/drivers/timer.c , but it still needs some touch to print how much seconds passed. I have lack of knowledge of NASM, since I only know GAS, so I can't go more with the wiki. I've encountered a problem like these, so please help me... Thanks.

https://github.com/AhnJihwan/AhnTri/blob/main/Makefile Why??

(If you know how, please comment or contribute to my OS AhnTri. link -> https://github.com/AhnJihwan/AhnTri)

Sweet, its at https://github.com/Andrispowq/HackOS. I feel like the structure is self-explanatory, but if you need any help deciphering anything feel free to hit me up with a DM and I'll gladly help you!

I could have something useful for you. What I have is basically a two-stage bootloader, where my first two stages are flat binaries, and the kernel is on a FAT32 filesystem in a different partition, but it looks it up at runtime. Look under boot/BIOS/stage_2/src at https://github.com/Andrispowq/HackOS, hope it can help!

I dont recall this rule, you can check out my code, its jn C++ under kernel/src/proc, at https://github.com/Andrispowq/HackOS. My implementation works for me so you should be able to get it to work.

I have a pretty complete UEFI bootloader here: https://github.com/Andrispowq/HackOS/blob/master/boot/UEFI/gnu-efi/bootloader/main.c, feel free to check it out. It may not be the most complete one out there, and the kernel can be booted from BIOS and UEFI, so if you're interested in that kind of stuff you can check that out as well. I followed PonchoOS's tutorial if you want to look it up it could be useful.

To get started take a look at the fatgen103 document by Microsoft, which will be handy when implementing it. My own implementation in the kernel is at https://github.com/Andrispowq/HackOS/tree/kernel/src/fs/fat32. The separate implementation is at https://github.com/Andrispowq/FilesystemDrivers/tree/FAT32/src. Feel free to use the code for inspiration, although I wouldn't copy it altogether because that would not help learning. Good luck!

So anyway, I started out following James Molloy's tutorials, which has a solution for this, but it's protected mode code for 2-level paging, and I heard it's terrible, so I decided to start making up my own code. You can see the most important part under https://github.com/Andrispowq/HackOS/blob/master/kernel/src/arch/x86_64/paging/page_table_manager.cpp. The main problem is that the moment I switch the CR3's value during a task switch it instantly triple faults. I don't know if this is because the PML4 is created on a different virtual address than it's physical address (I'm uploading the PHYSICAL address), or because the tables are invalid for some reason.

So, I've came to the point where I want to read and write from and to the disk. So far, my only conclusion is that INT 13H AL=02H or AL=42H pretty much works on everything, including emulators, and my own computer. However, after it, I'm pretty much clueless. I've got some ATA code from somewhere which works perfectly on QEMU, until I pass the -M q35 flag to use a Q35 machine, where it stops working. (The code is here: HackOS/ata.cpp at master · Andrispowq/HackOS (github.com)). Also, it doesn't work on my machine.

Well, this is a long topic, and everyone does it differently. What I did was to not use the aligned attribute, because I can't define every page table in advance. Instead I request a page from the page frame allocator, which is guaranteed to be 4K aligned. To map virtual to physical and get the corresponding physical to virtual address, is an interesting thing. I can give you more info in half an hour, until then, take a look at my code, which is 64 bit, but the idea is the same: https://github.com/Andrispowq/HackOS. Go under kernel/arch/x86_64/paging to see the code.

One last thing: check out my github repo, you can use it as a reference. If you have questions, text me, here or in private, and I'll gladly help. https://github.com/Andrispowq/HackOS

On this line, you're looking at the wrong part of the VBE Info Block structure. The mode list pointer is at offset 14, not offset 18.