Path: utzoo!attcan!uunet!crdgw1!uakari.primate.wisc.edu!zaphod.mps.ohio-state.edu!caen!uflorida!thedon.cis.ufl.edu!seeger
From: seeger@thedon.cis.ufl.edu (F. L. Charles Seeger III)
Newsgroups: comp.unix.admin
Subject: Re: Is the encrypted password's salt simply random?
Message-ID: <27219@uflorida.cis.ufl.EDU>
Date: 1 Mar 91 00:28:31 GMT
References: <1991Feb27.202424.16444@rigel.econ.uga.edu>
Sender: news@uflorida.cis.ufl.EDU
Organization: UF CIS Dept
Lines: 59

In article <1991Feb27.202424.16444@rigel.econ.uga.edu> glenn@rigel.econ.uga.edu (Glenn F. Leavell) summarizes:
|In article <1991Feb26.201846.22584@rigel.econ.uga.edu> I recently wrote:
|>I'm using a randomly generated two-character salt from the set [a-zA-Z0-9./],
|>and everything seems to be working fine.  Here's my question:  is this
|>the right way to choose the salt - just a random thing?

Included as part of the summary:

| >The old BSD 4.3 "passwd" program uses 
| >    (void)time(&salt);
| >    salt = 9 * getpid();
| >    saltc[0] = salt & 077;
| >    saltc[1] = (salt>>6) & 077;
| >    for (i = 0; i < 2; i++) {
| >        c = saltc[i] + '.';
| >        if (c > '9')
| >            c += 7;
| >        if (c > 'Z')
| >            c += 6;
| >        saltc[i] = c;
| >    }
| >    return(crypt(pwbuf, saltc));
| >which is based on the time of day clock.

Note that the salt generated by this code does not depend on the time.  The
assignment in the second line discards the result of the time() call.  The
fix that I have seen suggest is to change the assignment operator from "="
to "+=".  However, the UCB folks seem to have changed the code more
drastically.

static char sccsid[] = "@(#)passwd.c    4.42 (Berkeley) 6/19/90";
...
char *
getnewpasswd(pw, temp)
        register struct passwd *pw;
        char *temp;
{
        register char *p, *t;
        char buf[_PASSWORD_LEN+1], salt[2], *crypt(), *getpass();
...
        /* grab a random printable character that isn't a colon */
        (void)srandom((int)time((time_t *)NULL));
#ifdef NEWSALT
        salt[0] = '_';
        to64(&salt[1], (long)(29*25), 4);
        to64(&salt[5], (long)random(), 4);
#else
        to64(&salt[0], (long)random(), 2);
#endif
        return(crypt(buf, salt));
}

Personally, I'm a bit suspicious of NEWSALT and the "&salt[5]".

Chuck
-- 
  Charles Seeger    E301 CSE Building             Office: +1 904 392 1508
  CIS Department    University of Florida         Fax:    +1 904 392 1220
  seeger@ufl.edu    Gainesville, FL 32611-2024