#!/usr/bin/perl -w

# ip2host - Resolve IPs to hostnames in web server logs
# Maurice Aubrey <maurice.aubrey+ip2host@gmail.com>
#
# Usage: ip2host [OPTIONS] [cache_file] < infile > outfile
#
# For a usage summary, run: ip2host --usage
# For documentation, run: perldoc ip2host

# $Id: ip2host 24 2007-01-27 05:33:06Z mla $

our $CHILDREN = 40;        # Number of processes to fork
our $TIMEOUT  = 20;        # DNS timeout
our $BUFFER   = 50000;     # Maximum number of log lines to keep in memory
our $FLUSH    = 500;       # Flush output buffer every $FLUSH lines
our $CACHE    = '';        # Optional disk cache file to use
our $TTL      = 86400 * 7; # Seconds until disk cached ips are expired
our $DEBUG    = 0;
# Regular expression for matching an IP address
# $1 should be the IP
# If IP is always in first column then this would also work:
# q/^([\d.]+)/s;
our $REGEX    = '\b (\d{1,3}\.\d{1,3}\.\d{1,3}\.\d{1,3}) \b';

use strict;
use vars qw( %Opt %Buffer %Pending %Cache $Output_Line $Input_Line );
use Socket;
use IO::Handle;
use IO::Select;
use Getopt::Long;

our $VERSION = '1.10';

BEGIN {
  package IP_Cache;

  use strict;

  our $TTL;
  our @ISA = qw/ DB_File /;

  sub TIEHASH {
    my $class = shift;
    $TTL      = shift;

    require DB_File;
    $class->SUPER::TIEHASH(@_);
  }

  # In order to implement EXISTS, we'd need to parse
  # the value to see if the IP has expired, which is just
  # as expensive as FETCH.  So we'll just make sure we
  # never use it.
  sub EXISTS { die 'exists not implemented!' }

  sub _DB_FETCH {
    my $self = shift;
    my $ip   = shift;

    my $val = $self->SUPER::FETCH($ip) or return;

    my($utc, $host) = split /:/, $val, 2;
    time - $utc < $TTL or return;

    return $host;
  }

  {
    my %cache; 

    sub FETCH {
      my $self = shift;
      my $ip   = shift;

      return $cache{ $ip } if exists $cache{ $ip };
      return $cache{ $ip } = $self->_DB_FETCH($ip);
    }

    sub STORE {
      my $self = shift;
      my($ip, $host) = @_;

      $cache{ $ip } = $host;
      $self->SUPER::STORE( $ip => (time . ':' . $host) );
    }
  }
}

sub usage {
  my $exit = shift || 0;

  print STDERR <<EOF;
$0 version $VERSION

Usage: $0 [OPTIONS] [cache_file] < input_log > output_log

See the POD for more details: perldoc $0

Copyright 1999-2007, Maurice Aubrey <maurice\@hevanet.com>

This module is free software; you may redistribute it and/or
modify it under the same terms as Perl itself.      
EOF

  exit $exit;
}

sub resolve_ips($) {
  my $parent_fh = shift;

  $SIG{'ALRM'} = sub { die 'alarmed' };

  while(my $ip = <$parent_fh>) { # Get IP to resolve
    chomp($ip);
    my $host = undef;
    eval { # Try to resolve, but give up after $TIMEOUT seconds
      alarm($Opt{timeout});
      if (my $ip_struct = inet_aton $ip) {
	  $host = gethostbyaddr $ip_struct, AF_INET;
      }
      alarm(0);
    };
    # XXX Debug
    if ($Opt{debug} and $@ =~ /alarmed/) {
      $host ||= 'TIMEOUT';
      # warn "Alarming ($ip)...\n";
    }
    $host ||= $ip;
    print $parent_fh "$ip $host\n";
  }
}

sub fork_children($) {
  my $num_children = shift;

  my $read_set  = IO::Select->new;
  my $write_set = IO::Select->new;

  for(my $child = 1; $child <= $num_children; $child++) {
    my($child_fh, $parent_fh) = (IO::Handle->new, IO::Handle->new);
    socketpair($child_fh, $parent_fh, AF_UNIX, SOCK_STREAM, PF_UNSPEC)
      or die "socketpair failed: $!";
    $child_fh->autoflush;
    $parent_fh->autoflush;

    if (my $pid = fork) { # parent
      close $parent_fh;
      $write_set->add($child_fh); # Start out writing to children 
    } else { # child
      defined $pid or die "fork failed: $!";
      close $child_fh; close STDIN; close STDOUT;
      resolve_ips($parent_fh);
      exit 0; 
    } 
  }

  return ($read_set, $write_set);
}

# Write as many lines as we can until we come across one
# that's missing (that means it's still pending DNS).
sub flush_output {
  for (; exists $Buffer{ $Output_Line }; $Output_Line++) {
    print delete $Buffer{ $Output_Line };
  }
}

%Opt = (
  children => $CHILDREN,
  timeout  => $TIMEOUT,
  buffer   => $BUFFER,
  flush    => $FLUSH,
  cache    => $CACHE,
  ttl      => $TTL,
  debug    => $DEBUG,
  regex    => $REGEX,
);
GetOptions(\%Opt,
  "children|kids=i",
  "timeout=i",
  "buffer=i",
  "flush=i",
  "ttl=i",
  "cache=s",
  "usage|help|version",
  "debug",
  "regex=s",
);
usage(0) if $Opt{usage};
usage(1) if @ARGV > 1;
$Opt{cache} = shift if @ARGV;
$Opt{regex} = qr/$Opt{regex}/sx;

my($read_set, $write_set) = fork_children($Opt{children}); 

if ($Opt{cache}) { # Cache results to disk if asked
  tie %Cache, 'IP_Cache', $Opt{ttl}, $Opt{cache}
    or die "unable to tie '$Opt{cache}': $!";
}

$Output_Line = 1;
$Input_Line = 0;
while (1) {
  my $buffer_full = $Input_Line - $Output_Line >= $Opt{buffer};

  my($readable, $writable) = IO::Select->select(
    $read_set,
    $buffer_full ? undef : $write_set, # Throttle if buffer is full
    undef
  );

  while (@$writable) { # One or more children ready for IP
    my $line = <STDIN>;
    $Input_Line++;
    unless (defined $line) {
      undef $write_set;
      last;
    }
    my($ip) = ($line =~ /$Opt{regex}/);
    my($before, $after) = ($`, $');
    if (not defined $ip) { # No IP seen, pass it through unmolested
      $Buffer{ $Input_Line } = $line;
    } elsif (my $host = $Cache{ $ip }) { # We found this answer already
      $Buffer{ $Input_Line } = "$before$host$after";
    } elsif (exists $Pending{ $ip }) { # We're still looking
      push @{ $Pending{ $ip } }, [ $Input_Line, $before, $after ];
    } else { # Send IP to child
      my $fh = shift @$writable;
      print $fh "$ip\n";
      $Pending{ $ip } = [ [ $Input_Line, $before, $after ] ];
      $write_set->remove($fh);
      $read_set->add($fh); # Move to read set to wait for answer
    }

    flush_output if $Input_Line % $Opt{flush} == 0;
  }

  while (@$readable) {  # One or more children have an answer
    my $fh = shift @$readable;
    chomp(my $str = <$fh>);
    my($ip, $host) = split / /, $str, 2;
    $Cache{ $ip } = $host; 
    # Take all the lines that were pending for this IP and
    # toss them into the output buffer
    foreach my $pending (@{ $Pending{ $ip } }) {
      $Buffer{ $pending->[0] } = "$pending->[1]$host$pending->[2]";
    }
    delete $Pending{ $ip };
    $read_set->remove($fh);
    $write_set->add($fh) if defined $write_set; # Ready for new question
  }

  last if not defined $write_set and not keys %Pending;
  flush_output if $buffer_full;
}

flush_output;
exit 0;

=pod

=head1 NAME

  ip2host - Resolves IPs to hostnames in web server logs

=head1 SYNOPSIS

  ip2host [OPTIONS] [cache_file] < infile > outfile

  infile  - Web server log file.

  outfile - Same as input file, but with IPs resolved to hostnames.        

  Options:

  --children=...  Number of child processes to spawn (default: 40)
  --timeout=...   Seconds to wait on DNS response (default: 20)
  --buffer=...    Maximum number of log lines to keep in
                  memory (default: 50000)
  --flush=...     Number of lines to process before flushing
                  output buffer (default: 500)
  --cache=...     Filename to use as disk cache (default: none)
  --ttl=...       Number of seconds before IPs cached on disk are expired
                  (default: 604800 - One week)

=head1 DESCRIPTION

This is a faster, drop-in replacement for the logresolve
utility distributed with the Apache web server.

It's been reported to work under Linux, FreeBSD, Solaris,
Tru64, and IRIX.  

=head1 AUTHOR 

Maurice Aubrey E<lt>maurice.aubrey+ip2host@gmail.comE<gt>

Based on the logresolve.pl script by Rob Hartill.

=head1 COPYRIGHT

Copyright 1999-2007, Maurice Aubrey E<lt>maurice.aubrey+ip2host@gmail.comE<gt>.

This module is free software; you may redistribute it and/or
modify it under the same terms as Perl itself.

=head1 README

Resolves IPs to hostnames in web server logs.
This is a faster, drop-in replacement for the logresolve utility
distributed with the Apache web server.

=head1 SCRIPT CATEGORIES

Web        

=cut