On Mon, Oct 17, 2016 at 8:21 AM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
Heikki Linnakangas <hlinnaka@iki.fi> writes: > On 10/17/2016 05:50 PM, Tom Lane wrote: >> The real issue here is whether we are willing to say that >> Postgres simply does not work anymore on machines without standard entropy >> sources. Doesn't matter whether the user cares about the strength of >> cancel keys, we're just blowing them off. That seems a bit extreme >> from here. I think we should be willing to fall back to the old code >> if we can't find a real entropy source.
> I'm scared of having pg_strong_random() that is willing to fall back to > not-so-strong values. We can rename it, of course, but it seems > dangerous to use a weak random-number generator for authentication > purposes (query cancel, MD5 salts, SCRAM nonces).
I think that it's probably moot on all modern platforms, and even on platforms as old as pademelon, the answer for people who care about strong security is "--with-openssl". What I'm on about here is whether we should make people who don't care about that jump through hoops. Not caring is a perfectly reasonable stance for non-exposed postmasters; otherwise we wouldn't have the "trust" auth method.
I would be satisfied with making it a non-default build option, eg add this to pg_strong_random:
+1 for that approach. I really wouldn't want to see it fall back completely transparently in case something stops working. But if it's a non-default build option, that's not a problem, and it should make it possible to make it work on older platforms.
