[0/2] prandom_u32: make output less predictable

This is the cleanup of the latest series of prandom_u32 experimentations
consisting in using SipHash instead of Tausworthe to produce the randoms
used by the network stack. The changes to the files were kept minimal,
and the controversial commit that used to take noise from the fast_pool
(f227e3ec3b5c) was reverted. Instead, a dedicated "net_rand_noise" per_cpu
variable is fed from various sources of activities (networking, scheduling)
to perturb the SipHash state using fast, non-trivially predictable data,
instead of keeping it fully deterministic. The goal is essentially to make
any occasional memory leakage or brute-force attempt useless.

The resulting code was verified to be very slightly faster on x86_64 than
what is was with the controversial commit above, though this remains barely
above measurement noise. It was only build-tested on arm & arm64.

George Spelvin (1):
  random32: make prandom_u32() output unpredictable

Willy Tarreau (1):
  random32: add noise from network and scheduling activity

 drivers/char/random.c   |   1 -
 include/linux/prandom.h |  55 ++++-
 kernel/time/timer.c     |   9 +-
 lib/random32.c          | 438 ++++++++++++++++++++++++----------------
 net/core/dev.c          |   4 +
 5 files changed, 326 insertions(+), 181 deletions(-)

Cc: George Spelvin <lkml@sdf.org>
Cc: Amit Klein <aksecurity@gmail.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: "Jason A. Donenfeld" <Jason@zx2c4.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: tytso@mit.edu
Cc: Florian Westphal <fw@strlen.de>
Cc: Marc Plumb <lkml.mplumb@gmail.com>
Cc: Sedat Dilek <sedat.dilek@gmail.com>

On Tue, Sep 1, 2020 at 8:43 AM Willy Tarreau <w@1wt.eu> wrote:
>
> This is the cleanup of the latest series of prandom_u32 experimentations
> consisting in using SipHash instead of Tausworthe to produce the randoms
> used by the network stack. The changes to the files were kept minimal,
> and the controversial commit that used to take noise from the fast_pool
> (f227e3ec3b5c) was reverted. Instead, a dedicated "net_rand_noise" per_cpu
> variable is fed from various sources of activities (networking, scheduling)
> to perturb the SipHash state using fast, non-trivially predictable data,
> instead of keeping it fully deterministic. The goal is essentially to make
> any occasional memory leakage or brute-force attempt useless.
>
> The resulting code was verified to be very slightly faster on x86_64 than
> what is was with the controversial commit above, though this remains barely
> above measurement noise. It was only build-tested on arm & arm64.
>
> George Spelvin (1):
>   random32: make prandom_u32() output unpredictable
>
> Willy Tarreau (1):
>   random32: add noise from network and scheduling activity
>
>  drivers/char/random.c   |   1 -
>  include/linux/prandom.h |  55 ++++-
>  kernel/time/timer.c     |   9 +-
>  lib/random32.c          | 438 ++++++++++++++++++++++++----------------
>  net/core/dev.c          |   4 +
>  5 files changed, 326 insertions(+), 181 deletions(-)
>
> Cc: George Spelvin <lkml@sdf.org>
> Cc: Amit Klein <aksecurity@gmail.com>
> Cc: Eric Dumazet <edumazet@google.com>
> Cc: "Jason A. Donenfeld" <Jason@zx2c4.com>
> Cc: Andy Lutomirski <luto@kernel.org>
> Cc: Kees Cook <keescook@chromium.org>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Cc: Peter Zijlstra <peterz@infradead.org>
> Cc: Linus Torvalds <torvalds@linux-foundation.org>
> Cc: tytso@mit.edu
> Cc: Florian Westphal <fw@strlen.de>
> Cc: Marc Plumb <lkml.mplumb@gmail.com>
> Cc: Sedat Dilek <sedat.dilek@gmail.com>
>

I have tested with the patchset from [1].
( Later I saw, you dropped "WIP: tcp: reuse incoming skb hash in
tcp_conn_request()". )

- Sedat -

https://git.kernel.org/pub/scm/linux/kernel/git/wtarreau/prandom.git/log/?h=20200901-siphash-noise


> --
> 2.28.0

On Tue, Sep 01, 2020 at 04:41:13PM +0200, Sedat Dilek wrote:
> I have tested with the patchset from [1].
> ( Later I saw, you dropped "WIP: tcp: reuse incoming skb hash in
> tcp_conn_request()". )

Yes because it's a bit out of the cope of this series and makes sense
even without these patches, thus I assume Eric will take care of it
separately.

Willy

On Tue, Sep 1, 2020 at 4:55 PM Willy Tarreau <w@1wt.eu> wrote:
>
> On Tue, Sep 01, 2020 at 04:41:13PM +0200, Sedat Dilek wrote:
> > I have tested with the patchset from [1].
> > ( Later I saw, you dropped "WIP: tcp: reuse incoming skb hash in
> > tcp_conn_request()". )
>
> Yes because it's a bit out of the cope of this series and makes sense
> even without these patches, thus I assume Eric will take care of it
> separately.

Yes, I am still pondering on this one and really there is no hurry.