libxml2 personal notes

notes_libxml2

root@research:~/dev# cat main.c 
#include <stdio.h>
#include <libxml/parser.h>

int main(int argc, char *argv[])
{
  xmlDocPtr doc;
  xmlNodePtr node;

  /* read xmlfile */
  doc = xmlReadFile(argv[1], 0, 0);
  if (!argv[1])
  {
    printf("[!] You input an null data: '%s'\n", argv[1]);
    return -1;
  }

  node = xmlDocGetRootElement(doc);
  if (!node)
  {
    puts("[!] Make sure that is a correct file.\n");
    return -1;
  }
  
  return 0; 
}


-------------------------- %< --------------------------

0400691 <+1>:		mov    0x8(%rsi),%rdi		;align stack for the 3rd parameter, making argv[1]
							;pointer points to '*filename'
0400695 <+5>:		mov    %rsi,%rbx		;copies data from parameter to argument
0400698 <+8>:		xor    %edx,%edx		;initiates the register with value zero
040069a <+10>:	xor    %esi,%esi			;initiates the register with value zero
callq  0x400640 <xmlReadFile@plt>			;exec function: xmlReadFile(argv[1], 0, 0)


4006a1 <+17>:	cmpq   $0x0,0x8(%rbx)			;knowing that $0x0 corresponds to 'null terminator'
                                                        ; that is equals to 'NULL', then it is a check
							;if "*argv[]" points to NULL data.
4006a6 <+22>:	je     0x4006b9 <main+41>		;jump to "ret" (0x4006b9) whether is NULL
																							;is it looks like: if (!argv[1]) { return -1; }


4006a8 <+24>:	mov    %rax,%rdi				;copies the data stored in rdi (*filename) to rax
4006ab <+27>:	callq  0x400650 <xmlDocGetRootElement@plt>	;exec function "xmlDocGetRootElement(rax);"
4006b0 <+32>:	test   %rax,%rax				;set rax to '1'(true) if rax is '0'(false), then exec next current instruction
4006b3 <+35>:	je     0x4006cc <main+60>			; if rax was set to "1" jump to address 0x4006cc
								; reading the libxml2 documentation, note that the function "xmlDocGetRootElement"
								; has 2 behaviors defined --- from xmlsoft: "Returns:	the #xmlNodePtr for the root or NULL"
								; exec this 3 instructions below if rax was set to '0' -- xmlDocGetRootElement's behavior
4006b5 <+37>:	xor    %eax,%eax				; set eax to 0
4006b7 <+39>:	pop    %rbx					; remove rbx from stack
4006b8 <+40>:	retq   						; return eax

4006b9 <+41>:	xor    %esi,%esi				;reset the 32-bit LSB of the RSI register, - WORD size.
4006bb <+43>:	mov    $0x400868,%edi				;input the value stored in "0x40086" into EDI register.
4006c0 <+48>:	xor    %eax,%eax				;set EAX to '0'
4006c2 <+50>:	callq  0x400630 <printf@plt>			;callq belongs to SIMD (single instruction, multiple data)
																						;perform the printf() execution and pushes the current address
																						;of instruction pointer (%rip) onto the stack and jumps to it.

; return after callq
04006c7 <+55>:	or     $0xffffffff,%eax		;set to '1' all bits (32-bits length)
04006ca <+58>:	pop    %rbx			;remove rbx from the stack
04006cb <+59>:	retq   				;return "-1" (0xffffffff)

; calling "puts" function with message stored at "$0x400890" and do the same thing as above and return -1
04006cc <+60>:	mov    $0x400890,%edi
04006d1 <+65>:	callq  0x400660 <puts@plt>
04006d6 <+70>:	or     $0xffffffff,%eax
04006d9 <+73>:	pop    %rbx
04006da <+74>:	retq 

-------------------------- %< --------------------------

Dropping this as public 2 years later.

[Warning] This note has no intention of teaching. They're just random notes and outburst during my studies two years later.