OSDev.org

Posted: **Sun May 28, 2023 7:54 am**

The os has this gdt:

GDTLM:
	.Null: equ $ - GDTLM
		dq 0
	
	.Code: equ $ - GDTLM
		dw 0xffff
		dw 0
		db 0
		db 10011010b
		db 10101111b
		db 0
	
	.Data: equ $ - GDTLM
		dw 0xffff
		dw 0
		db 0
		db 10010010b
		db 11001111b
		db 0

	.UserCode: equ $ - GDTLM
		dw 0xffff
		dw 0
		db 0
		db 11111010b
		db 10101111b
		db 0

	.UserData: equ $ - GDTLM
		dw 0xffff
		dw 0
		db 0
		db 11110010b
		db 10101111b
		db 0
	
	.Pointer:
		dw $ - GDTLM - 1
		dq GDTLM

And this code is used for switching to user mode:

Code: Select all

cli			
	push 0x23
	push rsp
	pushfq
	push 0x1B
	push _user
	iretq

But when I try it in bochs it says this:

Code: Select all

00392068645e[CPU0  ] check_cs(0x5045): not a valid code segment !
00392068648e[CPU0  ] check_cs(0x5534): not a valid code segment !
00392068651e[CPU0  ] check_cs(0x5534): not a valid code segment !
00392068654e[CPU0  ] check_cs(0x5534): not a valid code segment !

This repeats many times and then bochs says:

Code: Select all

00392068918e[CPU0  ] interrupt(long mode): not accessible or not code segment
00392068918e[CPU0  ] interrupt(long mode): not accessible or not code segment

Then it just restarts the whole system.
I've already asked this question on stackoverflow, but no one helped.
What do I do to make it work?

Posted: **Wed May 31, 2023 11:58 am**

MindW1n wrote:But when I try it in bochs it says this:

I don't see anything in the code you've shared so far that could cause that error. Step through your code using the Bochs debugger and see if you can figure out where the error is happening. I suggest placing a magic breakpoint right before the IRETQ instruction.

Posted: **Wed May 31, 2023 4:34 pm**

I'm going to take a stab in the dark and say that based on the issues I keep having, try changing bit 6 of the the flags/20-16 limit byte of your code segment from 0 to 1. If that bit is 0, it means its a 16 bit segment and if that bit is a 1, it means its a 32 bit segment. When that bit is zero, my system which runs on c++ compiled code doesnt boot even on qemu or bochs.

Posted: **Wed May 31, 2023 4:40 pm**

SomeGuyWithAKeyboard wrote:I'm going to take a stab in the dark and say that based on the issues I keep having, try changing bit 6 of the the flags/20-16 limit byte of your code segment from 0 to 1. If that bit is 0, it means its a 16 bit segment and if that bit is a 1, it means its a 32 bit segment.

It's a 64-bit code segment. Bit 6 must be 0 for a 64-bit code segment.

Posted: **Thu Jun 01, 2023 5:42 am**

Solved! It's wasn't the general protection fault. It was the page fault, I only needed to allocate some page for user code and that solved the problem.

Posted: **Thu Jun 01, 2023 11:10 am**

Don't you have a page fault handler? You should figure out why it wasn't called.

Posted: **Thu Jun 01, 2023 11:17 am**

No, I don't have one

Posted: **Thu Jun 01, 2023 11:22 am**

Why don't you have a page fault handler? Exception handlers should be one of the first things you set up, so you don't waste time trying to figure out where exceptions occur.

Posted: **Thu Jun 01, 2023 12:36 pm**

Yes, you're right. I should have done that first!

OSDev.org

General protection fault when try to switch to user mode

General protection fault when try to switch to user mode

Re: General protection fault when try to switch to user mode

Re: General protection fault when try to switch to user mode

Re: General protection fault when try to switch to user mode

Re: General protection fault when try to switch to user mode

Re: General protection fault when try to switch to user mode

Re: General protection fault when try to switch to user mode

Re: General protection fault when try to switch to user mode

Re: General protection fault when try to switch to user mode