On Thu, Dec 01, 2011 at 03:52:37PM +0800, Herbert Xu wrote:

On Thu, Dec 01, 2011 at 04:24:18PM +0900, Simon Horman wrote:

quoted

So while I agree that optimizing the hash is a good idea.  I don't believe
it is a bottle-neck at this point. Though I could be convinced otherwise if
long collision chains could be constructed with relatively few flows.
Something I had not considered until I rad your email just now.

It's not an optimisation issue, but a security one.  If you leave
a hash like this with a constant seed, an attacker would have an
infinite amount of time to find collisions.

Rehashing isn't all that difficult.

Sorry for missing the point. Yes, I agree that rehashing makes a lot of sense.