random: make mixing interface byte-oriented
authorMatt Mackall <mpm@selenic.com>
Tue, 29 Apr 2008 08:03:05 +0000 (01:03 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 29 Apr 2008 15:06:25 +0000 (08:06 -0700)
commit3acd35037b8711d801db6a0c2735e337b07096aa
treef74a1d9585295463f166fdcffcc8e28527728b1a
parent40a7eceb473f0004c3905f06b99480b980120ce9
random: make mixing interface byte-oriented

Switch add_entropy_words to a byte-oriented interface, eliminating numerous
casts and byte/word size rounding issues.  This also reduces the overall
bit/byte/word confusion in this code.

We now mix a byte at a time into the word-based pool.  This takes four times
as many iterations, but should be negligible compared to hashing overhead.
This also increases our pool churn, which adds some depth against some
theoretical failure modes.

The function name is changed to emphasize pool mixing and deemphasize entropy
(the samples mixed in may not contain any).  extract is added to the core
function to make it clear that it extracts from the pool.

Signed-off-by: Matt Mackall <mpm@selenic.com>
Cc: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
drivers/char/random.c