what does [BITS 32] do in nasm?

[MarkZar]

Hi
I'm new to os development, and now some small questions confused me.

According to the book I am reading, when you are writing codes in protected mode using NASM, you should add [BITS 32] in place, just as follows:

Code: Select all

[SECTION .s32]
ALIGN	32
[BITS 32]

LABEL_PM_START:
	mov ax, SelectorVideo        ;GDT Selector pointed to the video descriptor
	mov gs, ax

	mov ah, 0Fh	
	mov al, 'P'
	mov [gs:((80 * 0 +39) * 2)], ax        ;display 'P'
	jmp $

I'm really confused with that [BITS 32], because I dont know what intheworld NASM does when I add it. I compiled the asm file and then disassembled it using ndisasm, founding the right codes generated as follows:

Code: Select all

00090320  66B81B00          mov ax,0x1b
00090324  8EE8              mov gs,ax
00090326  B40F              mov ah,0xf
00090328  B050              mov al,0x50
0009032A  6566A34E000000    mov [gs:0x4e],ax
00090331  EBFE              jmp short 0x90331

Exactly same with what I write in asm file, but once I changed the [BITS 32] to [BITS 16], I found the disassembled codes is whole different:

Code: Select all

00090320  66B81B008EE8      mov eax,0xe88e001b
00090326  B40F              mov ah,0xf
00090328  B050              mov al,0x50
0009032A  6566A34E00        mov [gs:0x4e],eax
0009032F  0000              add [bx+si],al
00090331  EBFE              jmp short 0x331

Why?

[bluemoon]

I'm really confused with that [BITS 32], because I dont know what intheworld NASM does when I add it.

Let me google that for you
It's well documented on the manual. However if you are having problem on the bit concept for different architectures, it's a long story.

By the way I think this should go to General Programming section.

[Gigasoft]

[Bits 32] generates code that is suitable for a 32 bit code segment. Specifically, the default operand and address sizes are 32 bit and the meaning of 66h/67h prefixes are reversed.

[MarkZar]

I'm really confused with that [BITS 32], because I dont know what intheworld NASM does when I add it.

Let me google that for you
It's well documented on the manual. However if you are having problem on the bit concept for different architectures, it's a long story.

By the way I think this should go to General Programming section.

Thank you, in fact I have looked through the manual, but still a little confused with it.Maybe practice is more than any words

[MarkZar]

[Bits 32] generates code that is suitable for a 32 bit code segment. Specifically, the default operand and address sizes are 32 bit and the meaning of 66h/67h prefixes are reversed.

Thank you