Newbie Question

Newbie Question

[dAvId KeDrOsKy]

Hello, 

I have recently started learning x86 assembly and have been programming it in 
Linux with NASM (Netwide Assembler). 

I am trying to write a program that takes a string as input, reverses it and 
calculated its length.  I'm using the stack to do so.  

Here's my problem..  NASM only lets me move the string into a 32-bit register.

e.g.  		mov	EBX, the_string		; move memory address of the_string into EBX

At this point I need to push the register on the stack.  I know I can do:

push EBX
call string_length_procedure

BUT I was wondering how it is possible to push BX instead.  Does BX also hold 
the string in this case, considering the size of the string it isn't that 
big?  If not, is there a way to put the contents of EBX (the string) into BX? 
For the time being I would like to just deal with 16-bit registers because it 
is easier to follow along with the book I am using.

P.s. I am using an already coded procedure that gets the string for me and 
puts it into 'string' in this case.

Re: Newbie Question

[Rudolf Marek]

> BUT I was wondering how it is possible to push BX instead.  Does BX also hold 
> the string in this case, considering the size of the string it isn't that 
> big?  If not, is there a way to put the contents of EBX (the string) into BX? 
> For the time being I would like to just deal with 16-bit registers because it 
> is easier to follow along with the book I am using.

Yes but if you are using linux it uses 32-bit memory pointers so you 
should use the 32 bit regs. If you want use some "done" work which is 
16bit you can just put "e" in front of 16 bit registers to ensure 
functionality. (I think)

Regards

Rudolf

Re: Newbie Question

[Brian Raiter]

> Here's my problem..  NASM only lets me move the string into a 32-bit
> register.

Please note that the register does NOT contain the string -- just a
POINTER to the string (i.e., its memory address).

> BUT I was wondering how it is possible to push BX instead.  Does BX
> also hold the string in this case, considering the size of the
> string it isn't that big?

The size of the POINTER has no relationship to the size of the
string. (If I give you a street address, like 1148 23rd Ave. S, you
have no way of telling from that address if the house is big or
small.)

> If not, is there a way to put the contents of EBX (the string) into
> BX?  For the time being I would like to just deal with 16-bit
> registers because it is easier to follow along with the book I am
> using.

Then perhaps you need a different book.

Imagine an airplane pilot climbing into the cockpit of a 747 and
asking how he can use a propeller instead of a jet engine, because he
doesn't need to fly very far, and the flying instructions he was given
was for an airplane with a propeller. What odds would you give that
that pilot is going to be able to get to the runway, much less off of
it?

b

Re: Newbie Question

[Robin Miyagi]

Here is a 32 bit i386 assembler course;

http://www.drpaulcarter.com/pcasm/

It has good linux coverage.  This link can be found on Konstanin's
linux assembly resources page.

On Thursday 24 October 2002 16:43, Brian Raiter wrote:
>
> Then perhaps you need a different book.
>

-- 
Robin Miyagi<penguin-computing@c2c2c.ca>
http://penguin.c2c2c.ca/
The funny thing about brakes, is that when they break, they don't brake.

Newbie Question

[dAvId KeDrOsKy]

The following program gives me a seg fault.  But I don't understand why.

1.	mov	ebx, dword 1234
2.	mov	ax, [ebx]

To me the code would mean 1) move the integer 1234 into ebx and then 2) move 
the value of ebx into ax.  But of course my reasoning is incorrect.

Feedback would be great.

Dave

Re: Newbie Question

[Dragan Stancevic]

On Thursday 27 February 2003 14:57, dAvId KeDrOsKy wrote:
> The following program gives me a seg fault.  But I don't understand why.
>
> 1.	mov	ebx, dword 1234
> 2.	mov	ax, [ebx]
>
> To me the code would mean 1) move the integer 1234 into ebx and then 2)
> move the value of ebx into ax.  But of course my reasoning is incorrect.

You are trying to apply indirection to ebx, failing while trying to read from 
a memory location that is not accessible to your process.

Drop the brackets, and also you probably want eax, right?

-- 
Peace can only come as a natural consequence
of universal enlightenment. -Dr. Nikola Tesla

Re: Newbie Question

[Rudolf Marek]

> 
> The following program gives me a seg fault.  But I don't understand why.
> 
> 1.	mov	ebx, dword 1234
> 2.	mov	ax, [ebx]
> 
> To me the code would mean 1) move the integer 1234 into ebx and then 

its OK


2) move the value of ebx into ax.  But of course my reasoning is incorrect.

BAD 1) you cant use opernads of different sizes.

on one side must be 32 on another 32 tooo 

2) operator [ ] is used for referencing into memory,
so you were accessing a memory and thats why the segfault 
(adress 0x1234 of process is onexistent)

if you want to "shorter" 32 bit value to 16 bit then use

mov eax,0  ;xor eax,eax works too
mov ax,bx  ;copy value from BX to AX


Ah dragan you were faster....

Regards

Rudolf



 


> 
> Feedback would be great.
> 
> Dave
> -
> To unsubscribe from this list: send the line "unsubscribe linux-assembly" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

Re: Newbie Question

[Robert G. Plantz]

On Thursday, February 27, 2003, at 02:57 PM, dAvId KeDrOsKy wrote:

> The following program gives me a seg fault.  But I don't understand  
> why.
>
> 1.	mov	ebx, dword 1234
> 2.	mov	ax, [ebx]
>
> To me the code would mean 1) move the integer 1234 into ebx and then  
> 2) move
> the value of ebx into ax.  But of course my reasoning is incorrect.

Assuming that you want the (16-bit) word at address 1234 to be loaded
into ax, I see two problems:
a. address 1234 probably cannot be accessed by your program, and
b. linux is unhappy if you change ebx. (Your program needs to
    save/restore ebx.)

Bob

------------------------------------------------------------------------ 
---------------
Robert G. Plantz, PhD		| http://www.cs.sonoma.edu/~bob
Sonoma State University		| ph: (707) 664-2806
CS Department			| fax: (707) 664-3012
1801 E. Cotati Avenue		| email: plantz@sonoma.edu
Rohnert Park, CA 94928

Re: Newbie Question

[Slack Traq]

--- dAvId KeDrOsKy <davidkedrosky@rogers.com> wrote:
> The following program gives me a seg fault.  But I
> don't understand why.
> 
> 1.	mov	ebx, dword 1234
> 2.	mov	ax, [ebx]
> 
> To me the code would mean 1) move the integer 1234
> into ebx and then 2) move 
> the value of ebx into ax.  But of course my
> reasoning is incorrect.

First of all, if you need to pass an integer to ebx
(32 bit register) you may wanna to do
1. mov     ebx, 1234
Second, if you put the brackets between a register,
its memory address is passed to ax, in this case.
Third, you can't move a 32 bit register into a 16 bit
one. Use eax, instead.
2. mov     eax, ebx
Advice: you can decompose a 32 bit register, ie eax,
in ax. Then you can decompose ax in ah and al (8 bit
registers).
If hope it will help you.
> 
> Feedback would be great.
> 
> Dave
> -
> To unsubscribe from this list: send the line
> "unsubscribe linux-assembly" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at 
http://vger.kernel.org/majordomo-info.html

Regards,

                                            Slack

__________________________________________________
Do you Yahoo!?
Yahoo! Tax Center - forms, calculators, tips, more
http://taxes.yahoo.com/

Re: Newbie Question

[Dragan Stancevic]

On Thursday 27 February 2003 15:26, Rudolf Marek wrote:
> Ah dragan you were faster....
> Regards

:-))))))

-- 
Peace can only come as a natural consequence
of universal enlightenment. -Dr. Nikola Tesla

Re: Newbie Question

[hpr]

Robert G. Plantz am Donnerstag, 27. Februar 2003 23:42:
> Assuming that you want the (16-bit) word at address 1234 to be loaded
> into ax, I see two problems:
> a. address 1234 probably cannot be accessed by your program, and
> b. linux is unhappy if you change ebx. (Your program needs to
>     save/restore ebx.)

not linux, but (probably) "C"...
i/o related syscalls use ebx for the file descriptor

hp

-- 
Linux,Assembly,Forth: <http://www.lxhp.in-berlin.de/index-lx.shtml>
    pse, reply to << lxhp -at- lxhp -dot- in-berlin -dot- de >>

newbie question

[raphael]

Hello list

I would like to know how i can manage to read the entry
on the keyboard without using C functions like scanf or get.

thank you for the help and forgive my poor english

Raphael

Re: newbie question

[jeff]

On 09/03/2005 10:16 am, raphael wrote:
> I would like to know how i can manage to read the entry
> on the keyboard without using C functions like scanf or get.

You can use kernel calls.  The keyboard operates in two
modes and you need to select the mode first.

In raw mode each key press is provided.  In cooked mode
a whole line of data is provided when <enter> key pressed.

There are lots of examples, but the code looks like this:
 1. select mode, save previous mode
 2. wait for key press
 3. restore mode

Here is some code from asmlib.  It shows handling of multi
byte keys and the mouse.  This is probably does more than
you want, but may be useful.

  [section .text]

 extern raw_set2
 extern raw_unset2
 extern kbuf
 extern key_poll
 extern mouse_check
 extern poll_rtn
 extern kbuf_end
 extern key_flush
     

;   read_stdin - read stdin data to "kbuf"
; INPUTS
;   none
; OUTPUT
;   kbuf - contains keys read with zero termination
; NOTES
;    source file: read_stdin.asm
;
;    read_stdin calls raw_set2 to put keyboard in  raw
;    mode.  It then reads data to "kbuf" and calls
;    "raw_unset2 to restore terminal state.
;
  global read_stdin
read_stdin:
  call	raw_set2			;set raw mode and save previous mode
km_10:
  mov	ecx,kbuf
more_keys:
  mov	eax,3			;kernel read call
  mov	edx,1			;read one byte
  mov	ebx,0			;read stdin
  int	80h
  or	eax,eax
  js	km_15
  add	ecx,eax			;advance buffer ptr
  call	key_poll
  test	byte [poll_rtn],8	;check for error
  jz	km_20
km_15:
  call	key_flush
  jmp	short km_10
km_20:
  cmp	ecx,kbuf_end
  je	k_exit			;exit if buffer full
  test	byte [poll_rtn],1
  jnz	more_keys
k_exit:
  mov	byte [ecx],0		;terminate string
  call	mouse_check		;check if mouse data read
  call	raw_unset2
  ret 

Re: newbie question

[Martin Fflores]

hi people, how i can have noticie that a key was pressed , but reading at 
port 60h, my problem is that i cant find a good documentation, if some one 
have some sourcecode(in ATT Assembly)Documentation or web link that will be 
great.

Take Folks.

_________________________________________________________________
Don’t just search. Find. Check out the new MSN Search! 
http://search.msn.click-url.com/go/onm00200636ave/direct/01/

Reading at port 60h

[Martin Fflores]

hi people, how i can have noticie that a key was pressed , but reading at 
port 60h, my problem is that i cant find a good documentation, if some one 
have some sourcecode(in ATT Assembly)Documentation or web link that will be 
great.

Take Folks.

_________________________________________________________________
FREE pop-up blocking with the new MSN Toolbar - get it now! 
http://toolbar.msn.click-url.com/go/onm00200415ave/direct/01/

Re: Reading at port 60h

[Richard Cooper]

> how i can have noticie that a key was pressed , but reading at port 60h,

You aren't doing that in Linux are you?  Linux reads the keyboard via an  
interrupt, if you're polling through port 0x60/0x64, you're not going to  
get anything because the kernel will get to it first.

You can have the kernel give you the scancodes from the keyboard on your  
stdin by using ioctls.  You can get more information about that by typing  
"man console_ioctl"

Basically, you call KDGKBMODE to get the current mode and save it, then  
call KDSKBMODE to set "RAW" mode, and when your program exits, you set   
the keyboard mode back like it was.  However, before that, you have to  
install signal handlers to catch if your program crashes or someone tries  
to kill it, because if your program doesn't set the keyboard mode back,  
the kernel isn't going to do it.  Also, before calling KDSKBMODE, you need  
to call TCSETSW to change the TTY settings, otherwise scancode 3 will be  
interpreted as contol-c, and scancode 13 will be translated to 10, and 127  
will be translated to 8, and probably some more.  You also have to use  
TCGETS to save those settings, so you can restore them too.  Then you have  
to watch for Alt-Fn and signal the kernel via VT_ACTIVATE to switch  
consoles, because it doesn't watch for Alt-Fn when they keyboard is in RAW  
mode.  On the plus side, you don't have to switch back to non-raw mode  
when you call VT_ACTIVATE, because the kernel keeps each console in it's  
own mode.

It's a mess of work, but it's doable, and will certainly work better than  
direct access.

> my problem is that i cant find a good documentation,

I've got a book, The Indispensable PC Hardware Book, which is a pretty  
good book.  It tells how to do direct hardware access with most things  
that it covers (but not much on VGA, unfortunately), and most of the  
examples are in assembly language.  The keyboard chapter is a pain to  
read, and I never did anything sucessfully with it, but then last time I  
tried was 10 years ago when I was 16, so maybe I just wasn't smart  
enough.  However, I didn't have any trouble with doing direct IDE access  
after reading the chapter about that, so I don't know, maybe there's just  
something wrong with that chapter.

Looking it over, it is quite confusing (e.g., the "output buffer" is the  
port that you read scancodes from, and the "input port" isn't a port at  
all), but it seems like the information is all there.

Looks like the keyboard I/O is definately confusing, though.  You've got  
the two ports you communicate to the controller with, then the controller  
has two ports that it communicates to the keyboard with, as well as  
setting things like the A20 gate, and then the keyboard is what does  
stuff.  Along the way there's an "output port" (of which the A20 gate is a  
part of) which the controller can both write to and read from, but to  
access it, you have to tell the controller to access it, because it's the  
controller's port, not yours.

> web link that will be great.

Normally the best documentation you can find on the web is the datasheet  
for the component itself:

http://www.jameco.com/webapp/wcs/stores/servlet/ProductDisplay?langId=-1&storeId=10001&catalogId=10001&productId=53014

However, the keyboard controller is just a generic programmable  
computer-in-a-chip.  So the datasheet doesn't help much.

Book link:
http://www.amazon.com/exec/obidos/tg/detail/-/0201596164/qid=1125801989/sr=8-1/ref=pd_bbs_1/102-3152860-8785751?v=glance&s=books&n=507846