linux-kernel - Re: [RFC] Should writes to /dev/urandom immediately affect reads?

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20230920204524.GD914@sol.localdomain>
Date:   Wed, 20 Sep 2023 13:45:24 -0700
From:   Eric Biggers <ebiggers@...nel.org>
To:     Linus Torvalds <torvalds@...ux-foundation.org>
Cc:     "Jason A. Donenfeld" <Jason@...c4.com>,
        linux-kernel@...r.kernel.org, linux-crypto@...r.kernel.org,
        Theodore Ts'o <tytso@....edu>,
        Dominik Brodowski <linux@...inikbrodowski.net>,
        Jann Horn <jannh@...gle.com>
Subject: Re: [RFC] Should writes to /dev/urandom immediately affect reads?

On Wed, Sep 20, 2023 at 01:32:29PM -0700, Linus Torvalds wrote:
> On Wed, 20 Sept 2023 at 13:21, Eric Biggers <ebiggers@...nel.org> wrote:
> >
> > It seems that what you're claiming (in addition to the RNG always being
> > initialized quickly on platforms that are "relevant", whatever that means) is
> > that once the RNG is "initialized", there's no need to reseed it anymore.
> 
> No. You are literally putting words in my mouth that I at no point
> even implied. You're making up an argument.
> 
> I *LITERALLY* am asking a very simple question: WHO DO YOU EVEN CARE
> ABOUT THIS "IMMEDIATE" EFFECT.
> 
> Give me a real reason. Give me *any* reason.
> 
> Don't try to turn this into some other discussion. I'm asking WHY DOES
> ANY OF THIS MATTER?
> 
> The immediacy has changed several times, as you yourself lined up. And
> as far as I can tell, none of this matter in the least.
> 
> > The question is, given that, shouldn't the RNG also reseed right
> > away when userspace explicitly adds something to it
> 
> I don't see that there is any "given" at all.
> 
> We do re-seed regularly. I'm not arguing against that.
> 
> I'm literally arguing against applying random changes without giving
> any actual reason for them.
> 
> Which is why I'm asking "why do you care"? Give em a *reason*. Why
> would a user space write matter at all?
> 
> It was why I also asked about entropy. Because *if* you argue that the
> user-space write contains entropy, then that would be a reason.
> 
> You didn't.
> 
> You argue that the current behavior hasn't been the universal behavior. I agree.
> 
> But considering that we've switched behaviors apparently at least
> three times, and at no point did it make any difference, my argument
> is really that without a *REASON*, why would we switch behavior *four*
> times?
> 
> Is it just "four changes is better than three"?

See my first email where I explained the problems with the current behavior.
Especially the third paragraph.

I'll just wait until Jason has a chance to reply.  This discussion clearly isn't
achieving anything with just us two.

- Eric