On Wed, Jun 16, 2021 at 11:16:02PM +1000, Michael Ellerman wrote: > Nicholas Piggin <npig...@gmail.com> writes: > > I would be happier if you didn't change this (or at least put it in its > > own patch explaining why it's not going to slow down other platforms). > > It would essentially be a revert of 01c9348c7620 ("powerpc: Use hardware > RNG for arch_get_random_seed_* not arch_get_random_*") > > Which would be ironic :)
You expect me to remember things I did 6 years ago? :) I'll take this part out. My thinking originally was that since darn on Microwatt is a single-cycle instruction, it would be faster to use darn every time rather than run a software PRNG seeded from darn. It's not critical though. Paul.