random: Make the min-entropy estimate configurable

[markj]

Summary

Right now the cutoff values for the RCT and APT tests are computed with
a fixed min-entropy estimate of 1. In preparation for permitting
alternative estimates for "pure" sources (i.e., hardware noise sources),
extend the code to handle alternative estimates of an integer number of
bits.

For the RCT test, the cutoff is simply the formula from section 4.4.1 of
NIST SP 800-90B. The APT test, I used Excel to compute a lookup table.

[cem]

sys/dev/random/random_harvestq.c

503

The ht_apt_cutoff for min_entropy=1 changed from 330 to 329 -- what's the significance of that?

692

In a later diff, we change this to just

random_healthtest_init(i, 0);

We could just do that here.

[markj]

sys/dev/random/random_harvestq.c

503

I think there must have been an off-by-one in my original calculation. This table was computed with =1+CRITBINOM(511,POWER(2,-ROW()),1-POWER(2,-34)) per the footnote in the 800-90B publication. I'm quite sure I used the same formula before too, but might have included the +1 term twice.

In practice there's not much significance since the difference means we're testing 329 vs. 330 samples out of 511 for equality with the first sample in the window.

[markj]

Get rid of an unneeded variable.

[cem]

sys/dev/random/random_harvestq.c

503

Cool, thanks.