> On Mon, 13 May 1996, Theodore Ts'o wrote:
>
> > The original reason I didn't provide a CONFIG_RANDOM was because I
> > wanted security-oriented applications (i.e., Netscape, PGP, Kerberos,
> > etc.) to be able to assume that if they were on Linux, /dev/linux would
> > always be present. Good, secure numbers are absolutely vital for
>
> Do You really think that they will be aware of such an Linux *SPECIFIC*
> animal?
the random driver is cool and well implemented, and it has a >minimal<
interface. Why not use it? We are not talking about a 10000 functions API.
> > The random driver also isn't all that big, and the overhead of the
> > add_XXX_randomness() calls were designed to be as small as possible.
>
> That's compleatly wrong!! It is now about 16KBytes. More than the floppy
note that good random numbers are/will be crutial for future networking.
Why not a bit of testing. Or do you want to send your credit card number
mangled with standard libc srand() ? =)
btw, libc should use this interface IMHO, in some way, thus applications
using libc *rand* functions could be enhanced.
-- mingo
ps. the Linux entry from the NSA bible:
Linux: the OS with blocking /dev/random
=B-)