Re: [Nasm-users] 2 Problems: (Strange Linking Problem) && (Strange Syntatx Error)

[Frank K. <fbk...@co...>]

Matthias-Christian Ott wrote:
> 
> Hi!
> I've 2 Problems with nasm:
> 
> 1: *Strage Linking Problem*
> 
> I have this code:
> 
> SECTION .DATA
>         hello:     db 'Hello world!',10
>         helloLen:  equ $-hello
> 
> SECTION .TEXT
>         GLOBAL main
> 
> main:
>         ; Write 'Hello world!' to the screen
>         mov eax,4            ; 'write' system call
>         mov ebx,1            ; file descriptor 1 = screen
>         mov ecx,hello        ; string to write
>         mov edx,helloLen     ; length of string to write
>         int 80h              ; call the kernel
> 
>         ; Terminate program
>         mov eax,1            ; 'exit' system call
>         mov ebx,0            ; exit with error code 0
>         int 80h              ; call the kernel
> 
> If try to assemble, link (by using ld) and run it, I only get a
> segementation fault (I have to rename main to _start):

"_start" is the default entrypoint known to ld (should be
declared "global" so ld can "see" it). You can tell ld to
use a different symbol (or an address, I guess) as the
entrypoint with "--entry=main".

> nasm -f elf hello.asm
> ld -s -o hello hello.o

I'm surprised you don't get a complaint from ld here!

> ./hello
> segementation fault
> 
> If I try to link it with the gcc (_start -> main), everything works fine.

C knows "main" as the entry into *your* part of the code,
and it's "call"ed, so you can end with "ret". In this case
the "_start" label occurs in the "C startup code", which
diddles the stack, among other things, and then calls
"main".

> What's wrong? Why doesn't it work?

Just that you didn't use "_start" or tell ld another
entrypoint.

> 2: *Strange Syntax Error*
> 
> I have this code (I know that it will only work in the priviledged mode,
> but it should assemble cleanly (?)):
> 
> global sys_core_in
> 
> sys_core_in:
>         push    ebp
>         push    dx

This will work, but you probably want to push/pop edx here.
It'll generate shorter code, and leave your stack aligned
better.

>         mov     ebp,esp
>         in      edx,[ebp+4]

<http://nasm.sourceforge.net/doc/html/nasmdocb.html#section-B.4.119>

>         mov     eax,dx

I don't think you'll need to do this, but...

<http://nasm.sourceforge.net/doc/html/nasmdocb.html#section-B.4.156>

... or use "movzx"... but I don't think you need it...

>         pop     dx
>         pop     ebp
>         ret
> 
> global sys_core_out
> 
> sys_core_out:
>         push    ebp
>         mov     ebp,esp
>         out     [ebp+4],[ebp+8]

<http://nasm.sourceforge.net/doc/html/nasmdocb.html#section-B.4.194>

>         pop     ebp
>         ret
> 
> If I try to assemble it I get this:
> 
> io.s:7: error: invalid combination of opcode and operands
> io.s:8: error: invalid combination of opcode and operands
> io.s:18: error: invalid combination of opcode and operands
> 
> What's wrong with the syntax?

Just like what the error message says - the operands you're
using aren't valid for the opcodes.

I don't know anything about "privileged mode", but you can
get "permission" to in/out ports with "sys_ioperm"... might
be useful for testing(?).

Best,
Frank

P.S. I've removed "nasm-devel" from the reply... not really
a "development issue"...