prologue ripper

[Michael Chastain <mec dot gnu at mindspring dot com>]

Mark Kettenis <kettenis@chello.nl> wrote:
mark> Thanks!  This really is the info I need.  Could you post (or mail me)
mark> the perl fu?

Yeah, I'll just stick it at the end of this message, just in case anyone
besides me and thee ever cares.  :)

mark> On systems that use the -freg-struct-return by default (FreeBSD,
mark> OpenBSD, Cygwin and a few others) I guess there are a bit more
mark> possibilities.

I don't have a Cygwin, but I have been using FreeBSD and OpenBSD
in the HP Test Drive cluster.  I'll just hop on, do some builds,
and prologue-rip them.

Michael C

===

#! /usr/bin/perl
#
# Copyright Abandoned 2004, Michael Chastain, <mec.gnu@mindspring.com>
# This script is public domain.
#
# Process the prologues from an executable file.
#
# Usage:
#
#   objdump -d cc1plus > cc1plus.odd
#   perl prologue.pl cc1plus.odd > cc1plus.pro
#   sort cc1plus.pro | uniq -c > cc1plus.sort
#
# The stuff at the end is x86 specific.
# You can hack it for other arches.
# Have fun!

use integer;
use strict;
use warnings;
$| = 1;

while (<>)
{
  # scan for beginning of function
  next unless $_ =~ m/^0[^<]*<([^>]*)>.*$/;
  my $name = $1;
  my $line = "";

  # scan the first few instructions
  my @insn = ();
  while (<>)
  {
    last if $_ =~ m/^Disassembly of /;
    last if $_ =~ m/^$/;
    chomp $_;
    # standardize the whitespace a bit
    $_ =~ s/^\s*[0-9A-Fa-f]+:\s*([0-9A-Fa-f]+\s+)*//;
    $_ =~ s/,/, /g;
    $_ =~ s/\s+/ /g;
    $_ =~ s/^\s+//g;
    push @insn, $_;
    last if scalar(@insn) >= 7;
  }

  # the following is x86 specific

  # print short lines for the simple folks
  if ( $insn[0] eq 'push %ebp'
  and  $insn[1] eq 'mov %esp, %ebp')
    { print "\@0\@ $insn[0] | $insn[1]\n"; next; }

  # print lines for the 1-insn-inserted folks
  if ( $insn[0] eq 'push %ebp'
  and  $insn[2] eq 'mov %esp, %ebp')
  {
    my $i1 = $insn[1];
    $i1 =~ s/\$0x[0-9A-Fa-f]+/\$0xIMMEDIATE/;
    $i1 =~ s/0x[0-9A-Fa-f]+/0xADDRESS/;
    print "\@1\@ $insn[0] | $i1 | $insn[2]\n";
    next;
  }

  # print lines for the 2-insn-inserted folks
  if ( $insn[0] eq 'push %ebp'
  and  $insn[3] eq 'mov %esp, %ebp')
  {
    my $i1 = $insn[1];
    $i1 =~ s/\$0x[0-9A-Fa-f]+/\$0xIMMEDIATE/;
    $i1 =~ s/0x[0-9A-Fa-f]+/0xADDRESS/;
    my $i2 = $insn[2];
    $i2 =~ s/\$0x[0-9A-Fa-f]+/\$0xIMMEDIATE/;
    $i2 =~ s/0x[0-9A-Fa-f]+/0xADDRESS/;
    print "\@2\@ $name : $insn[0] | $i1 | $i2 | $insn[3]\n";
    next;
  }

  # print the whole schmeer for the complex folks
  print "\@Z\@ $name : ", join (' | ', @insn), "\n";
}