From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S965865AbdCXOrs convert rfc822-to-8bit (ORCPT ); Fri, 24 Mar 2017 10:47:48 -0400 Received: from mail.eperm.de ([89.247.134.16]:58332 "EHLO mail.eperm.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S934911AbdCXOr3 (ORCPT ); Fri, 24 Mar 2017 10:47:29 -0400 From: Stephan =?ISO-8859-1?Q?M=FCller?= To: Krzysztof Kozlowski Cc: Kukjin Kim , Javier Martinez Canillas , Matt Mackall , Herbert Xu , "David S. Miller" , linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-samsung-soc@vger.kernel.org, linux-crypto@vger.kernel.org, Arnd Bergmann , Olof Johansson Subject: Re: [PATCH 1/3] crypto: hw_random - Add new Exynos RNG driver Date: Fri, 24 Mar 2017 15:46:44 +0100 Message-ID: <3628768.ze2UxnIp5I@tauon.atsec.com> In-Reply-To: <20170324144348.v4hegak7dffh5awh@kozik-lap> References: <20170324142446.31129-1-krzk@kernel.org> <2503592.nlbQnVP8UR@positron.chronox.de> <20170324144348.v4hegak7dffh5awh@kozik-lap> MIME-Version: 1.0 Content-Transfer-Encoding: 8BIT Content-Type: text/plain; charset="iso-8859-1" Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Am Freitag, 24. März 2017, 15:43:48 CET schrieb Krzysztof Kozlowski: Hi Krzysztof, > On Fri, Mar 24, 2017 at 03:37:59PM +0100, Stephan Müller wrote: > > Am Freitag, 24. März 2017, 15:24:44 CET schrieb Krzysztof Kozlowski: > > > > Hi Krzysztof, > > > > > + > > > +static int exynos_rng_set_seed(struct exynos_rng_dev *rng, > > > + const u8 *seed, unsigned int slen) > > > +{ > > > + int ret, i; > > > + u32 val; > > > + > > > + dev_dbg(rng->dev, "Seeding with %u bytes\n", slen); > > > + > > > + ret = clk_prepare_enable(rng->clk); > > > + if (ret) > > > + return ret; > > > + > > > + if (slen < EXYNOS_RNG_SEED_SIZE) { > > > + dev_warn(rng->dev, "Seed too short (only %u bytes)\n", slen); > > > + ret = -EINVAL; > > > + goto out; > > > + } > > > + > > > + for (i = 0 ; i < EXYNOS_RNG_SEED_REGS ; i++) { > > > + val = seed[i * 4] << 24; > > > + val |= seed[i * 4 + 1] << 16; > > > + val |= seed[i * 4 + 2] << 8; > > > + val |= seed[i * 4 + 3] << 0; > > > + > > > + exynos_rng_writel(rng, val, EXYNOS_RNG_SEED(i)); > > > + } > > > + > > > + val = exynos_rng_readl(rng, EXYNOS_RNG_STATUS); > > > + if (!(val & EXYNOS_RNG_STATUS_SEED_SETTING_DONE)) { > > > + dev_warn(rng->dev, "Seed setting not finished\n"); > > > + ret = -EIO; > > > + goto out; > > > + } > > > + > > > + ret = 0; > > > + /* Save seed for suspend */ > > > + memcpy(rng->seed_save, seed, slen); > > > > Is this really necessary? If you need to save some seed, shouldn't that be > > an output of the DRNG and not the real seed? > > When suspended to RAM device will loose the contents of registers > (including SEED registers) so it has to be initialized with something on > resume. > > The seed registers are write-only so I cannot read them and store > contents just before suspend. > > I understand that real seed should not be stored... but then if I am not > able to re-seed it with same values, I will loose the continuous and > reproducible pseudo-random generation after suspend. Aren't this > expected out of PRNG? An RNG has to be stateless from the perspective of the caller -- this is the core implication of entropy. Then, if you add the initial seed after the RNG lost its state implies that the same sequence of random numbers starts again. I.e. where is the randomness? > > > Besides, how do you know that slen is not larger than > > EXYNOS_RNG_SEED_SIZE? > > Right, there is a overflow here. It should be sizeof(rng->seed_save); shouldn't it be min(rng->seed_save, slen)? > > Thanks for review! > > Best regards, > Krzysztof Ciao Stephan