Messages in this thread | | | Date | Fri, 15 Aug 2003 18:34:02 -0400 | From | Theodore Ts'o <> | Subject | Re: [RFC][PATCH] Make cryptoapi non-optional? |
| |
On Thu, Aug 14, 2003 at 06:17:13PM -0600, Val Henson wrote: > There is no way in which folding is better than > halving, and folding is demonstrably worse if SHA-1's output is > correlated across the two halves in any way (which is almost certainly > true).
If there any kind of correlation between any two individual bits, such that XOR'ing them together resulted in a pattern which was not statistically random, it would be a significant weakness in the SHA algorithm.
Consider that a requirement of a secure cryptographic hash is one where for any (x,y) pair where y=H(x), finding another x' where y=H(x') is no easier than brute force search. If two bits in the output of the hash have any kind of statistically detectable correlation, then it effectively reduces the size of the output space of the hash, and so it would take less effort than a brute force search of (on average) 2**80 random inputs before finding a hash collision.
- Ted - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |