[Discuss] Failing WD Disks

[richard.pieri at gmail.com (Rich Pieri)]

On Thu, 18 May 2023 18:42:31 -0400
Bill Ricker <bill.n1vux at gmail.com> wrote:

> My concern was *since* *his* /dev/urandom reportedly got *faster*,
> does that mean someone *improved* it (safely!), or *broke* it (aka
> ^*improved*^ it, badly)?

Fair point. People break secure enough things all the time, but it's
unlikely in the Stable and Longterm kernel branches. You are running
Stable or Longterm kernels in production, right?

> If it were my system, i'd want to be looking at the change notes to
> see if it was intentional, and if so what cryptographer approved the
> patches to urandom that made it faster.

The 4.8 kernel *did* change how /dev/urandom works and it is faster for
it. But it's no less secure because the CSPRNG is still the same CSPRNG
that feeds /dev/random. There just are fewer steps now. This is
detailed in the article I linked. I don't know if this is what happened
on Kent's system or not. There are lots of reasons why I/O performance
can change -- even a different USB port given that we're dealing with
external storage -- and none of us here but Kent can possibly be aware
of them all.

-- 
\m/ (--) \m/