[SOLVED] AMD athlon 64 bit assembler

[gerard4143]

Hi how do you read the argc and argv arguments in 64 bit assembler(AMD athlon 64 bit)?...I know how to do it in 32 bit intel/amd assembler i.e.

.section .data

.section .bss

.section .text
	.global _start

_start:
			nop
			popl	%ebx
			popl	%ebx
			popl	%ebx

			movl	$8, %eax
			movl	$0644, %ecx
			int	$0x80

			movl	%eax, %ebx
			movl	$1, %eax
			int	$0x80

compile and then run the program
./programname filename
and you will create a file called filename

but when I try the same in AMD athlon 64 bit i.e.

.section .data

.section .bss

.section .text
	.global _start
_start:
			nop
			popq	%rbx
			popq	%rbx
			popq	%rbx

			movq	$8, %rax
			movq	$0644, %rcx
			int	$0x80

			movq	%rax, %rbx
			movq	$1, %rax
			int	$0x80

compile and then run the program
./programname filename
and its does not create a file called filename

The gnu debugger gdb (in the 64 bit example)does show a char pointer in %rbx with the value "filename" but the program fails to create a file
any help will be appreciated...Thanks Gerard
Note: its not a permissions problem because I create and delete files in this folder all the time in C/C++.

[gerard4143]

AMD Athlon 64 bit assembler

Not a solution(at least not a pretty one) but it does copy the data from argv[1] when compiled and run

./programname one two three four

it will write "one" to the terminal. All I did was copy the pointer value of avgv[1] off the stack and dereferenced it byte by byte and copied into one, two, three and then printed it with the system call write(system call 4). Like I said not pretty...but if I use the copied pointer value directly in the system call write nothing happens...
could it be that the kernel has problems with the larger pointer values ie.

the pointer value in %rdi = 0x7fff5a703342 and I can dereference it properly in the assembler code but as soon as I try to pass it in a system call it fails...any thoughts on this would be appreciated...Thanks Gerard4143

.section .data
	one: .byte 'a' 
	two: .byte 'a'
	three: .byte 'a'

.section .bss

.section .text
	.global _start
_start:
			nop
			movq	16(%rsp), %rdi
			
			movb	(%rdi), %cl 
			movb	%cl, one
			movb	1(%rdi), %cl 
			movb	%cl, two
			movb	2(%rdi), %cl 
			movb	%cl, three

			movq	$4, %rax 
			movq	$1, %rbx
			movq	$one, %rcx
			movq	$3, %rdx
			int	$0x80

			movq	$1, %rax
			movq	$0, %rbx
			int	$0x80

Here's the original problem with file creation system call...it works but not pretty

.section .data
	one: .byte 'a' 
	two: .byte 'a'
	three: .byte 'a'

.section .bss

.section .text
	.global _start
_start:
			nop
			movq	16(%rsp), %rdi

			movb	(%rdi), %cl
			movb	%cl, one
			movb	1(%rdi), %cl
			movb	%cl, two
			movb	2(%rdi), %cl
			movb	%cl, three

			movq	$8, %rax
			movq	$one, %rbx
			movq	$0644, %rcx
			int	$0x80

			movq	$1, %rax
			movq	$0, %rbx
			int	$0x80

[gerard4143]

This have really changed for 64 bits AMD machines...this is the solution

and here's the web page that I got the info:

http://ubuntuforums.org/showthread.php?t=728995

.section .data
	
.section .bss

.section .text
	.global _start
_start:
			nop
			movq	24(%rsp), %rsi

			movq	$1, %rax
			movq	$1, %rdi
			movq	$3, %rdx
			syscall

			movq	$60, %rax
			movq	$0, %rdi
			syscall