1996-04-07 - RNG_DEVICE Environment Variable?

Header Data

From: rngaugp@alpha.c2.org
To: cypherpunks@toad.com
Message Hash: 2202b2b96ff9f94efb4e75f1d8dddf1dc7b1d973d42d75869ef255395f100630
Message ID: <199604062155.QAA00591@miron.vip.best.com>
Reply To: N/A
UTC Datetime: 1996-04-07 05:51:08 UTC
Raw Date: Sun, 7 Apr 1996 13:51:08 +0800

Raw message

From: rngaugp@alpha.c2.org
Date: Sun, 7 Apr 1996 13:51:08 +0800
To: cypherpunks@toad.com
Subject: RNG_DEVICE Environment Variable?
Message-ID: <199604062155.QAA00591@miron.vip.best.com>
MIME-Version: 1.0
Content-Type: text/plain


>Considering a couple of RNG hardware manufacturers use different 
>names for device interfaces, perhaps it would be 'convenient' to 
>have apps look for the environment variable RNG_DEVICE which gives 
>the name of whatever device is used (rnadom$, random, rand, even 
>lpt2...) or even a special file that is mixed periodically by a cron 
>job (noiz.c?).

>The assumption is that reads from that device would return "truly 
>random" bytes (not from a pseudo-RNG), either from specialized 
>hardware or a system-noise sampler such as noise.sys (DOS) or 
>random.c (Linux, FreeBSD).

>How software would handle not getting enough bytes is another matter, >
>perhaps left configurable to the app.
>
>Rob. 

My hack to PGP to support RNG uses the varriable RNGDRIVER,
but is in config.txt, not the environment.





Thread