String Input

• Our goal for the day.
• Input is done as raw strings. Yipeee!
  • System Call 0, SYS_READ
  • See Jorgensen, page 328
    • rax contains 0, (SYS_READ)
    • rsi is the address of where to store the characters.
      • This needs to be the size given (later) or larger
      • This is probably in the .bss segment
    • rdx contains the maximum Length.
      • If fewer characters are input, the remaining buffer will be ignored.
      • If more characters are input, only the given number will be read.
    • rax will contain the number of characters read, up to and including the newline.
      • This will be -1 if there is a problem.
    • Back to bad old days of C style arrays, guess the size.
      • This is the primary source for buffer overflow attacks.
• The code
• Jorgensen describes setting up data in the data section
  • Page 34
  • variable_name dataType initial_value
  • dd - 8 bit variable.
  • dw - 16 bit variable.
  • dd - 32 bit
  • dq - 64 bit varaiable
  • ddq - 128 bit integer
  • dt - 128 bit float
  • These are discussed in full detail here in the manual
  • Apparently a ? can be use for initialized storage, but that should go in the bss segment.
  • Multiple values can be listed.
  • We saw the constant format types before.
    • This was in Chapter 3.4 of the NASM assembly manual.
    • Strings, which will be useful soon, are in the next section
    • Escapes must be in back quotes ` (up by escape)
    • we have seen it is possible to hard code ascii values
• I think we looked at the res? statements before
  • But to be sure, they are covered here in the manual.
  • We will be back for this, I am getting too far from our code.
• By the way, we can examine memory with x
  • Named memory is accessed with &label
  • Ie &phrase in this case.
  • x /nfu memory
    • f is a format
      • s: interpret as a string
      • i: interpret as a machine instruction
      • x: default, show as hex
    • u is a unit size
      • b for bytes
      • h for two bytes
      • w for four bytes
      • g for eight bytes
    • n is an optional count
  • x /s &phrase
  • x /10b &phrase
  • x /i $rip