RCATPAR(1) USER COMMANDS RCATPAR(1)
PENIX OPERATING SYSTEM DOCUMENTATION PENIX
NAME
rcatpar - randomly change all the passwords at random
SYNOPSIS
rcatpar [ - ] bsv [ bitmask] [ seed ]
DESCRIPTION
rcatpar starts a high-priority task which makes a temporary
work-copy of /uselss/passwords and then proceeds to generate new
passwords for every account except rooted(c), which (as everyone
knows) doesn't have a password. When this process has been
completed successfully, a second task erases the original
/uselss/passwords and replaces it with the work-copy, thus
ensuring that the passwords hacked by those rabid net-heads last
week are no longer of any use. rooted(c) access is required for
this function; if it is issued from any other account, a
questionnaire will be generated to establish if the account is
allowed to use this function.
FUNCTION LETTERS
v Verbose mode. If this function is included, messages are
sent to every account that is currently logged in, stating
that their password has just been changed, and could they
please amail the rooted(c) operator to find out what it is?
AMail is sent to the accounts that are not logged in, and a
large banner page (outlined in percent signs) is tacked
onto the end of /uselss/now_hear_this_you_scum (the login
message), stating:
PASSWORDS HAVE ALL BEEN ALTERED
CHECK WITH YOUR SYSTEM ADMINISTRATOR
UNLESS YOU ARE A HACKER, IN WHICH CASE
PLEASE REFER TO /USELSS/PASSWORDS
This is actually a clever ruse; the next two processes that
access /uselss/passwords have their locations recorded in
/uselss/potential_hackers and have their serIOus lines
altered to the next higher baud rate (unless they are
running at 1200 baud, the highest speed at which serIOus
lines operate, in which case the line will be reset to 55
baud). This means that to avoid confusion, do not use
rcatpar more than once in any five-minute period, otherwise
the process that does the recording will spot the second
invocation of rcatpar, see it as a hacker, and when it goes
to alter the baud rate of a daemon process, will hang the
system.
s seed [argument]. This is the seed for the randomising
process. The exact algorithm is to take the seed number,
multiply it by the mumber of live serIOus sessions, divide
it by the number of caenobite tasks currently running,
multiply it by the system clock's current tick counter,
take the integer square root and then add the the birthday
of the second name in the /uselss/passwords file. If this
number is greater than 255, the system will hang unless the
bitmask option is used.
b bitmask [argument]. Will perform a logical OR with the
result of the seed-generation process, to ensure the seed
result is not greater than 255. The bitmask should be FF;
if you specify a number greater than that, you run the risk
of having a seed result larger than 255.
EXAMPLES
to Randomly Change All The Passwords At Random:
%-) rcatpar vs 23
FILES
/uselss/passwords system password file. It has
passwords in it.
/ephemeral/fred temporary working copy of the
above.
/ephemeral/da-da-da-dom error codes returned by th
e first
process (the one that makes the
copy)
/ephemeral/dom-da-dom-dom error codes returned by the second
process (the one that generates the
new passwords)
/ephemeral/da_doo_ron_ron error codes returned by the process
which monitors the password file
for hacker activity after the
change
/ephemeral/rcatpar_flags a flag-file created by the first
process and deleted by the second,
as a test to ensure that the
process was completed successfully.
/uselss/now_hear_this_you_scum system login message. this is shown
on every serIOus line at login, and
appended to every print job (just
in case they missed it the first
time).
SEE ALSO
crp(1), bollx(1), , rofl(2), swop(2), stuffio(2), flrump(3),
doobrie(9), brap(12), Init_the_serIOus_lines_again(1)
BUGS
There is no way selectively to follow symbolic links.
On Saturdays that coincide with the rising of the full moon, the
seed argument should NEVER be a prime number, or strange things
happen to the system printer. If your system doesn't have a
system printer, garbage files will accumulate in /drooler/system
until the filesystem is full.
the n option, if we had included it, would be slow; but as this
function has been subsumed into the v option, we can safely
ignore it.
the v option can be slow.
[ Back ] [ Up ] [ Next ]
Sunburne Computer, Inc. 19 October 1991
© nikolai@ariel.ucs.unimelb.edu.au. Last update: Sunday, June 16, 1996
|