Re: [Pgubook-readers] movl _start, %eax

[Jonathan Bartlett]

Very wierd.  Looking into it...

The _start symbol is replaced correctly.  Anyway, I'll try to mess w/ it
this week.

Jon

On Sun, 21 Nov 2004, Shuaib wrote:

> I dissasemble the maximum program in chapter 3.
> (objdump maximum -D) and I get this (only a part of
> it):
>
> 08048074 <_start>:
>  8048074:       bf 00 00 00 00          mov
> $0x0,%edi
>  8048079:       8b 04 bd a8 90 04 08    mov
> 0x80490a8(,%edi,4),%eax
>
> 080490a8 <data_items>:
>  80490a8:       07                      pop    %es
>  80490a9:       00 00                   add
> %al,(%eax)
>  80490ab:       00 08                   add
> %cl,(%eax)
>
>
> When I try "movl _data_items, %eax", %eax has 7 in it.
> When I try "movl $_data_items, %eax", %eax has
> 080490a8 in it. Great, everything makes sense!
>
> On the other hand, When I try "movl $_start, %eax",
> %eax has 08048074 in it (good; makes sense since
> $_start is the address). When I try "movl _data_items,
> %eax", %eax has 0xcc in it. The address location
> pointed to by _start contains bf000000, and I was
> expecting bf000000 when I executed that instruction.
> WHY am I getting 0xcc? I suspect other people will get
> other values in the %eax.
>
> Here is what I'd like some elite to clarify: Does the
> OS (or the loader) relocate the _start section
> somewhere else in the memory and 0x08048074 no longer
> points to the the original address that _start pointed
> to? Where does this 0xcc come from? Will I be able to
> see this in the dissasembly section?
>
> I'd very much appreciate an explanation. I do not
> believe the reason is in the book.
>
>
>
> __________________________________
> Do you Yahoo!?
> The all-new My Yahoo! - Get yours free!
> http://my.yahoo.com
>
>
>
>
> _______________________________________________
> Pgubook-readers mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/pgubook-readers
>