linux-kernel - Re: [PATCH v2] Allow hwrng to initialize crng.

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]

Date:   Thu, 13 Dec 2018 22:58:21 -0500
From:   "Theodore Y. Ts'o" <tytso@....edu>
To:     Ard Biesheuvel <ard.biesheuvel@...aro.org>
Cc:     louiscollard@...omium.org,
        "open list:HARDWARE RANDOM NUMBER GENERATOR CORE" 
        <linux-crypto@...r.kernel.org>, mpm@...enic.com,
        Herbert Xu <herbert@...dor.apana.org.au>,
        Arnd Bergmann <arnd@...db.de>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        gary.hook@....com, m@...s.ch,
        PrasannaKumar Muralidharan <prasannatsmkumar@...il.com>,
        "Michael S. Tsirkin" <mst@...hat.com>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        apronin@...omium.org,
        Jarkko Sakkinen <jarkko.sakkinen@...ux.intel.com>,
        linux@...ewoehner.de, david.bild@...tum.com
Subject: Re: [PATCH v2] Allow hwrng to initialize crng.

On Thu, Dec 13, 2018 at 10:48:07AM +0100, Ard Biesheuvel wrote:
> > @@ -64,13 +65,19 @@ static size_t rng_buffer_size(void)
> >  static void add_early_randomness(struct hwrng *rng)
> >  {
> >         int bytes_read;
> > -       size_t size = min_t(size_t, 16, rng_buffer_size());
> > +       /* Read enough to initialize crng. */
> > +       size_t size = min_t(size_t,
> > +                           2*CHACHA20_KEY_SIZE,
> 
> This should be as symbolic constant that retains its meaning even if
> we move away from ChaCha20 or modify the current implementation

Also, rng_buffer_size() could be less than 2*hCHACHA20_KEY_SIZE, at
which point your goal wouldn't be realized.  What I'd recommend is to
keep the line:

	size_t size = min_t(size_t, 16, rng_buffer_size());

But to loop until rng_is_initialized() returns true or bytes_read is
0.  If you want to be paranoid, you could also break out of the loop
it isn't initialized after, say, 8 times through the loop.

Cheers,

					- Ted