OSDev.org

Posted: **Fri Nov 09, 2007 9:50 pm**

	sti();
	__asm__ __volatile__("ud2a\n\t");
	*(int*)2 = 3;
	CKernel::syshalt();

the "ud2a" can trap into interrupt,
but "*(int*)2 = 3;" can't work?
why it's in bochs?

Posted: **Sat Nov 10, 2007 3:51 am**

Code: Select all

*(int*)2 = 3;

That would only cause a pagefault if the page at virtual address 0 is not mapped or you do not have enough rights to write that page.

Posted: **Sun Nov 11, 2007 9:24 pm**

bluecode wrote:
Code: Select all
*(int*)2 = 3;
That would only cause a pagefault if the page at virtual address 0 is not mapped or you do not have enough rights to write that page.

hi,
i only mapped 3G~4G virtual address, and alloc 32M for bochs,

Code: Select all

void CTrap::test_trap(UNUSED int n)
{
        *(int *)0x2 = 3;			// failed trap in fault
        *(int *)0x40000020 = 3;		// failed trap in fault

        __asm__ __volatile__("ud2a\n\t");	// success trap in fault
}

thanks

Posted: **Mon Nov 12, 2007 5:59 am**

Code: Select all

void CTrap::test_trap(UNUSED int n)
{
        *(int *)0x2 = 3;			// failed trap in fault
        *(int *)0x40000020 = 3;		// success trap in fault

        __asm__ __volatile__("ud2a\n\t");	// success trap in fault
}

i made a little mistake, address 0x40000020 traps in gear,
all runs ok.
thanks

Posted: **Mon Nov 12, 2007 6:16 am**

did you forget to unmap the first 4M ? (tlb flushes and all)

Posted: **Mon Nov 12, 2007 6:50 pm**

Combuster wrote:did you forget to unmap the first 4M ? (tlb flushes and all)

maybe i forgot, the first 4M memory setup by grub, i just take reference of linux-2.4, and din't see unmap in linux.
kernel only init addresss 3G~3G+memsize , nothing exception

OSDev.org

strange interrupt problem with bochs?

strange interrupt problem with bochs?

Re: strange interrupt problem with bochs?

Re: strange interrupt problem with bochs?

Re: strange interrupt problem with bochs?