#!/usr/local/bin/perl -w
#Password file checker written by: Benjamin A. McFarland
#
#Purpose: This program is meant to munge through N passwd files and
# check for logins with multiple UIDs, UIDs with multiple
# logins, and logins with a UID of zero (0) that are not root.
# Logins with multiple UIDs are listed in the output file:
# "multi_uids.dat" as are users with an UID of zero (0). UIDs
# owned by more than one login are listed in "multi_logins.dat
+".
# All other users who pass these criteria are listed in the
# file: "goodusr.dat"
$length = scalar(@ARGV);
$SYSLIMIT = 100;
#We set this constant to allow for system logins. The constant allows
+for
#SAs to alter it here if they decide to allow system logins to have UI
+Ds
#higher or lower.
if($length == 0)
{
print STDOUT "passchk requires a source password file! USAGE:\n\tus
+er_name>passchk /path/to/your/passwd\n";
exit 0;
}
#The above conditional confirms that the script is run with at least o
+ne
#source file for data, else it exits with failure.
open(LOGPUT, "> "."multi_uids.dat");
open(UIDPUT, "> "."multi_logins.dat");
#Then the log files for the script are opened for writing. Old files a
+re
#clobbered since we're looking for the most recent data, and I didn't
+feel
#that the old data was pertinent.
#So, now we cycle through the number of source files given in the ARG
+V
#array.
for($i = 0; $i <= $length; $i++)
{
open(CURFILE, "$ARGV[$i]");
while($input = <CURFILE>)
{
#As long as we're getting data from the Current datafile, read it into
#the variable $input.
@dataf = split(/:/,$input);
#we break the data at the colons and place it into an array.
if(($dataf[2] == 0)&&($dataf[0] ne "root"))
{
print LOGPUT "**\n*WARNING* User: $dataf[0] has a UID of $dataf
+[2]!\n**\n";
}
#The above loop checks right away for any multiple logins with a UID o
+f
#zero.
if((!defined $loghash{$dataf[0]})&&($dataf[2] > $SYSLIMIT))
{
$loghash{$dataf[0]} = $dataf[2];
$badhash{$dataf[0]} = 0;
if(!defined $srcfile{$dataf[0]})
{
$srcfile{$dataf[0]} = $ARGV[$i];
}
}
#Now we check for values that haven't been entered into our logging
#arrays and that have a UID of larger than the constant set at the
#beginning of the script. If so, we log them, and set the source array
+.
#This is done in the conditionals above and below.
if((!defined $uidhash{$dataf[2]})&&($dataf[2] > $SYSLIMIT))
{
$uidhash{$dataf[2]} = $dataf[0];
$badhash{$dataf[2]} = 0;
if(!defined $srcfile{$dataf[2]})
{
$srcfile{$dataf[2]} = $ARGV[$i];
}
}
#Now we check to see if the UID doesn't match the one previously set
#with this login and that the UID is above our system constant. If not
+,
#we log it into the appropriate file and set the bad flag.
if(($loghash{$dataf[0]} ne $dataf[2])&&($dataf[2] > $SYSLIMIT))
{
print LOGPUT "$dataf[0] had multiple UIDs $loghash{$dataf[0]}
+ from file: $srcfile{$dataf[0]}\n\tand $dataf[2] from file:$ARGV[$i]\
+n";
$badlist[$dataf[0]} = 1;
}
#Same conditional as above, but this time we're looking to see that th
+e
#login is the same one set to this UID, the opposite condition of the
#previous conditional. If not, then we set the bad flag.
if(($uidhash{$dataf[2]} ne $dataf[0])&&($dataf[2] > $SYSLIMIT))
{
print UIDPUT "UID $dataf[2] is being used by both:\n$uidhas
+h{$dataf[2]} from file: $srcfile{$dataf[2]} and\n$dataf[0] from file:
+ $ARGV[$i]\n";
$badlist{dataf[2]} = 1;
}
}
close(CURFILE);
#We're done with the current file, close it before we open a new one.
}
#At this point, we've completed all of our source file munging, and
#should have compiled all the valid and invalid logins and UIDs. So,
#we open up our good data file, pull up the keys, sort them alphabetic
+ally,
#check that they haven't ended up in our bad list, and print them out.
open(GOODFILE, "> "."goodusr.dat");
foreach $key (sort keys(%loghash))
{
if(($badlist{$key} != 1) && ($badlist{$loghash{$key}} != 1))
{
print GOODFILE "$key is using only valid UID $loghash{$key},
+ from:\n$srcfile{$key}\n";
}
}
close(GOODFILE);
close(LOGPUT);
close(UIDPUT);
#We're done with this bad boy, so we shut all our files like a good
#monkey.
|
