Re: Random Number Seed

["Robert G. Brown" <rgb at phy dot duke dot edu>]

On Mon, 21 Feb 2005, Olaf Lenz wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Hello!
> 
> A few months ago, you suggested the following code snippet for seeding
> the RNG from /dev/random:
> 
> - -------------------------------------------------
> #include <stdio.h>
> #include <sys/time.h>
> 
> unsigned long int random_seed()
> {
> 
> ~ unsigned int seed;
> ~ struct timeval tv;
> ~ FILE *devrandom;
> 
> ~ if ((devrandom = fopen("/dev/random","r")) == NULL) {
> ~   gettimeofday(&tv,0);
> ~   seed = tv.tv_sec + tv.tv_usec;
> ~   if(verbose == D_SEED) printf("Got seed %u from gettimeofday()\n",seed);
> ~ } else {
> ~   fread(&seed,sizeof(seed),1,devrandom);
> ~   if(verbose == D_SEED) printf("Got seed %u from /dev/random\n",seed);
> ~   fclose(devrandom);
> ~ }
> 
> ~ return(seed);
> }
> - -------------------------------------------------
> 
> I've used the code for quite a while now and only today I noticed a big
> problem with it. The code tests, if /dev/random can be opened, but it
> does NOT test if the fread has actually read any number.
> 
> In my case, this resulted in the fact that the seed was not seeded at
> all and all processes used the same seed.... P-(
> 
> So to all who have been using the code, I would recommend to check their
> results. For the future, I would recommend to use the following code:

My bad, sorry.  I was assuming that the system itself wasn't broken
here, so that /dev/random might be absent but wouldn't be broken.  In
the man page (and to my experience):

       When read, the /dev/random device will only return random bytes
       within the estimated number of bits of noise in the entropy pool.
       /dev/random should be suitable for uses that need very high
       quality randomness such as one-time pad or key generation.  When
       the entropy pool is empty, reads from /dev/random will block
       until additional environmental noise is gathered.

So if you can open and try to read /dev/random the device blocks the
read until it successfully returns an integer (which can actually take a
long time if you are asking for a lot of rands as it has to generate
sufficient entropy).

So either this behavior is somehow implementation dependent on systems
with /dev/random, or your system has a broken /dev/random in the kernel,
or the system has a regular null file called "/dev/random" in the /dev
directory instead of the kernel-linked device.  In this latter case
(which I did not anticipate but yes, which is possible) you'd get the
behavior you describe, but I would consider the system "broken" and I'd
think that you want to know it instead of continuing.  Especially since
putting a file in place of /dev/random would make a system vulnerable to
various cryptographic attacks, one would think.  So much so that I'd
"suspect" an attack if I found a regular file there on a system properly
installed from a well-supported linux distribution.

So one might want to replace the "else" for fread with an error message
and punt unless your system has a different man page than mine and
doesn't state that /dev/random (unlike /dev/urandom) should block until
it returns.

    rgb

> 
> - -------------------------------------------------
> #include <stdio.h>
> #include <sys/time.h>
> 
> unsigned long int random_seed()
> {
> 
> ~ unsigned int seed;
> ~ struct timeval tv;
> ~ FILE *devrandom;
> 
> ~ if ((devrandom = fopen("/dev/random","r")) == NULL) {
> ~   gettimeofday(&tv,0);
> ~   seed = tv.tv_sec + tv.tv_usec;
> ~   if(verbose == D_SEED) printf("Got seed %u from gettimeofday()\n",seed);
> ~ } else {
> ~   if (fread(&seed,sizeof(seed),1,devrandom) == 1) {
> ~     if(verbose == D_SEED) printf("Got seed %u from /dev/random\n",seed);
> ~     fclose(devrandom);
> ~   } else {
> ~     gettimeofday(&tv,0);
> ~     seed = tv.tv_sec + tv.tv_usec;
> ~     if(verbose == D_SEED) printf("Got seed %u from
> gettimeofday()\n",seed);
> 
> ~   }
> ~ }
> 
> ~ return(seed);
> 
> }
> - -------------------------------------------------
> 
> Cheers
> 	Olaf
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.4 (GNU/Linux)
> Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
> 
> iD8DBQFCGcwjtQ3riQ3oo/oRAsjeAKC3CIz3kxxt/ZJUiuYzemIU1IqVdgCffoYW
> vXr8SEcXH69ulMzTfBwWuHw=
> =2RKb
> -----END PGP SIGNATURE-----
> 

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb@phy.duke.edu