page fault handler

[mr. x]

What shall I do to prevent my page fault handler to not start recursing? (When setting the present bit)

[mr. x]

I'm getting page faults when I try to read from memory, even when I haven't enabled paging.
The address I'm trying to read is 0x9C0000.
When I try to read 0x200000 it crashes without a page fault.

[Pype.Clicker]

well, one possible thing is to have something like

Code: Select all

pf_handler() {
    static int nested_pagefault=0;
    if (nested_pagefault) panic("Cannot handle page fault ...");
    nested_pagefault++;

    do_the_job();

    nested_pagefault--;
}

if a 'static' nested_pagefault doesn't suit your needs, you may try a thread_local one (but that supposes your OS has thread_local storage already)

if the PG bit of CR0 is cleared, then you shouldn't be receiving page faults (exception 0E). Are you sure your interrupt system is reliable ?

[mr. x]

Well, floppy and keyboard are using interrupts and both are working correctly.

[mr. x]

btw, here's my linking script:

Code: Select all

OUTPUT_FORMAT("binary")
ENTRY(_start)
OUTPUT("kernel.bin")

INPUT(kinit.o)
INPUT(kints.o)
INPUT(ints.o)
INPUT(driver_keyboard.o)
INPUT(driver_floppy.o)
INPUT(kernel.o)
INPUT(app_console.o)
INPUT(io.o)
INPUT(pic.o)
INPUT(idt.o)
INPUT(panic.o)
INPUT(doprintf.o)
INPUT(queue.o)
INPUT(string.o)
INPUT(memory.o)
INPUT(math.o)
INPUT(dma.o)
INPUT(driver_pefs.o)
INPUT(timer.o)
INPUT(mem_manager.o)

SECTIONS
{
  .text  0xFF800000 : {
    *(.text)
  }
  .data  : {
    *(.data)
  }
  .bss  :
  {                
    *(.bss)
  }
}

Code: Select all

[Pager]

Is it the 0xFFC'th entry or the 0xFFF'th entry which is used to point to the PageDirectory itself when we map the PageTables into the current address space? and can anyone clearly explain how this mapping the PageTables into the current address space works (in simple english).

[Pype.Clicker]

<timbot mode engaged>
Tim already did in one of its memory management tutorials on Bona Fide, iirc
<timbot mode disabled>